Use AI securely in your data room with Datasite MCP

Doug Cullen (00:00)

Good morning, good afternoon, and good evening to those joining us from all around the world. Thank you all for coming and attending the first in an incredible series of webinars that we have. This is particularly focused on the Datasite MCP for Claude. It's the first in our series, so please join us for the remaining parts over the next couple of weeks. My name is Doug Cullen. I'm Chief Strategy Officer here at Datasite. I have been in the dealmaking ecosystem for a couple of decades, and I think right now is the most exciting time.

I also oversee corporate development for us and joined through the acquisition of Blueflame about a year ago, where Alice works for us. Before we get too far into the conversation, I just wanted to cover a few quick housekeeping items. First of all, we want to hear from you. There is a question button there, so please ask questions throughout the webinar. We will leave some time at the end of the presentation to cover these questions, and we really want to hear from the dealmakers about what you want to know about MCP, specifically MCP for Claude. Tell us what you think via our survey.

As I said, this is one of the new series. This is something of interest to dealmakers around the world, and we want to do a lot more of these types of series. There are additional resources located in the console, including a couple of different documents about MCP, Datasite, and maybe Blueflame as well. Please check those out. And of course, just so you know, the session will be recorded and available on demand. If your colleagues didn't get a chance to join us live, please forward it around and encourage people to take advantage of these new capabilities.

Any webinar would not be the same without a quick legal disclaimer. As a reminder, all opinions expressed are our own and do not necessarily reflect those of, or those endorsed by, Datasite or Blueflame. So with that, let's get to the program. Alice, welcome. Alice Esmerian is a product strategist and is currently with Blueflame AI. I would love to get a little bit of background about you, Alice. How long have you been at Blueflame? What were you doing before you joined Blueflame?

Alice Esmerian (02:25)

I've been with Blueflame for six months now, so it has definitely been a very exciting time to be part of this AI ecosystem. Before that, I was in investment banking and private equity for eight years, so it's great to have the opportunity to bring my past-life experiences into this AI ecosystem.

Doug Cullen (02:45)

Yeah. That is super important. We're trying to bring our practitioners to the forefront, which is very important to how we think about dealmaking. We do think about it both as a set of technologists and as people who have been there and done that throughout our careers.

And so we're going to be trying to bring some of that to life to inspire you as dealmakers on how to take advantage of this next set of technologies. But with that, let's take a step back, and then we will get into the demonstrations. Let's think a little bit about how you have been reflecting on your journey at Blueflame. How have some of the technologies really started to change the way you think about how dealmaking gets done today and into the future?

Alice Esmerian (03:36)

It has been really exciting to picture our customers experimenting with AI and bringing more and more technology into almost every workflow in the life cycle.

For us, seeing people want to make AI very useful, practical, and easy to use, while also maintaining total compliance and governance within their systems, has been such an interesting question. Really, the future of AI within dealmaking is going to be about how you can prompt-drive a lot of what you do while maintaining absolute safety and governance within your systems.

Doug Cullen (04:11)

Yeah. It is really a fascinating time for us. We actually launched our first MCP partners to Datasite. For those of you who are familiar with Datasite, we have been a bit of a closed ecosystem for a long time, and we felt like we had a massive opportunity at DealMAX to launch our first MCP server, connecting out to various different LLMs.

Within that, we did some videos with Raj, the founder of Blueflame, and he introduced this concept of a prompt-first dealmaker. How do you think dealmaking has evolved, and what does a prompt-first dealmaker mean to you?

Alice Esmerian (04:55)

Well, that's really interesting. Honestly, this MCP availability for Datasite is really powering all our users to interact with their data rooms and bring together the content and the preparation of their deals from a single prompt. We'll be looking into it today within more practical workflows. There are a lot of use cases that will apply to this MCP.

We'll cover a few of them. We've built a lot of skills and gathered context for our users to use from the easiest prompt possible. But yes, it is really changing things, and a lot of tasks that were very manual and time-consuming are going to be accelerated significantly, leaving a lot more time for the really interesting parts of the job. Overall, it's very exciting.

Doug Cullen (05:45)

Yeah. I think one of the ways we talk about it is that MCP gives us this opportunity to introduce a layer of governance. I know not everyone gets excited about talking about governance, but one of the things we think is paramount to effective dealmaking is the opportunity to leverage and build on the trust that a brand like Datasite has in the market.

We want to make sure that we're maintaining permissions throughout, building on that framework of confidentiality, and extending it to the point of creating an audit trail and, ultimately, confidence for dealmakers, because people certainly entrust us with some of the largest deals around the world. So we wanted to make sure that people around the world knew that, as we extend our Datasite ecosystem to providers via MCP, we think it is super important to do so in a governed and risk-compliant way.

Alice Esmerian (06:49)

Sorry. And that's exactly where Blueflame AI really brings this value.

We know that Datasite customers choose Datasite every single day because they want their data in a very secure environment. Having Blueflame query this content and this context is the optimal way to maintain all of these security layers, traceability, and auditability. We'll get into that in our demo, but that is the best way to bridge between an LLM provider and Datasite content by bringing Blueflame into the equation.

Doug Cullen (07:19)

Yeah. And that's where Blueflame, for us, acts as this layer, as Alice is saying, between the Datasite data room content and the LLMs.

This is important because one of the things that we loved about Blueflame is the primary focus on dealmaking. We really believe that is a secret ingredient to great prompt-first dealmaking moving forward. But with that, in the spirit of showing rather than just talking, why don't we get right into some of the demonstrations?

Alice Esmerian (07:54)

Sure.

Doug Cullen (07:57)

And of course, we're doing this live, so hopefully everything will go well. Why don't we pull up some of this?

You're going to open up Claude. It has MCP access to Datasite and to connected systems. One of the things I think is really critical is when you get this huge volume of content, no matter where you are, and you're thinking about initiating and setting up a data room from the very beginning. Can you walk us through some of the power that Claude brings with that Blueflame layer and ultimate connectivity to Datasite?

Alice Esmerian (08:32)

Totally. Here, we're already connected, and now we're just going to show you four workflows.

There are obviously a lot more examples of use cases you could use by leveraging Blueflame into your data room through your MCP connectors. Feel free to ask questions in the Q&A if there are any workflows that you would like us to discuss. We'll review all of that, and it's very helpful for us. I connected my Datasite here through MCP connectors, which you can see at the bottom from the plus button. I also enabled Blueflame, which is going to be able to enter the content of all my Datasite files and really absorb and reason around the content of these documents, which is where the real value stands.

The first workflow is when you set up your data room. You're already in a smart environment where the agent automatically asks you a few questions to understand the context of your transaction. It is going to ask you what type of deal it is, what industry the deal covers, the size, and where the data is hosted in terms of geography. That automatically prepares your deal in the Datasite environment, and the model will come back with a suggested index. This is something that we have shipped within eight skills that cover eight different steps of the workflows, and we've really tailored them to our audience because we understand the pain of some Datasite and data room interactions.

This really accelerates that. Here, for example, I have a suggested index that covers 12 sections. Once I've reviewed that, I can provide simple feedback. In this particular case, I wanted to remove a section and a subsection. The model understands my feedback and, once I've approved it, will push that into my Datasite environment. Here it is: my Project Angel is live, and 155 subfolders have been created. It's crazy to think that this was just done in a few seconds. Before that, when I was a banker, I would have had to create all these folders manually and make sure there were no typos and that all the names were consistent across the 12 sections.

This is now completely accelerated. You just need to come back with your comments and your reorganization, and this is all set for you.

Doug Cullen (10:55)

Yeah. So just to take a slight step back, and we've accelerated this a bit for the purposes of the webinar, we are listed as a connector within the Claude ecosystem. If you have a Datasite login and if you have the availability on your Claude Desktop or Claude application, you can connect right into Datasite. One thing to note is that we are always respecting the permissions that you have in terms of accessing your files.

As you're seeing, we're opening up the ability to create a Datasite directly from your Claude environment. The other thing of note is that we will be talking about things at a slightly different level, because this data room that we have, or that Datasite has, has really been created with AI-empowered content that is highly processed and available via the Blueflame AI Assistant. Some of the things you're seeing here may or may not be available on your project, but we're trying to paint a picture of availability. I think this is so powerful because, as Alice was saying, if you had to do this previously, maybe you had the folders or maybe you had an Excel file with the folders, but now you're really able to do this.

We're able to do this specific to the type of deal, transaction type, and/or industry to power some pretty amazing things. I think the next thing you're going to do is work with these folders up there. We've created the index. What are some of the other things that you may have been doing as a banker over long weekends, or perhaps late at night, in preparation for getting this Datasite or data room set up?

Alice Esmerian (12:38)

So the next step in your workflow is typically that you have your data room set up and you're going to start wanting to incorporate documents.

Most of them you receive from your client. And the client, let's just say, is not going to spend a ton of time making sure that naming conventions are respected, that any scanned document doesn't read as Scan 002, and so on. Having a very clean pass at renaming all your files is something that is cumbersome yet extremely necessary. Obviously, it requires you to open every file and make sure the name ties to the right year and the right quarter. That's complicated. Now, by enabling this AI into your data room, the system is going to be able to first get a first pass at making sure that all the years and dates are being applied consistently.

When you're referring to the company name, it is always referring to it with the same spelling and wording. You're going to have the ability to flag any duplicate names, etc. Here you can see how the model went through and, just by reading the file names, suggested changes to each of them. However, as you pointed out, this is not leveraging any of the content within the files. This is just cleaning up existing file names. What you really want me to do, to replicate what you would have done as an analyst or as someone working on the deal, is to look into the content of these files so that it can intelligently provide guidance regarding the renaming.

I put here a use case that I think speaks quite well to the capabilities. I wanted all my contracts to have the same structure: the name of the party that signed the contract and the year in which the contract was signed, which is not readable from just a simple document name. Blueflame AI is enabled. It will safely go into your Datasite, read all the information, with no document or information ever leaving your Datasite ecosystem, and then reason with this data about where your contracts are.

What are they? Are they employee contracts, MSAs, or vendor agreements? And how do I rename these so that they're consistently labeled? In just a few seconds and with just a few prompts, you can see here that the model batched them by contract type and renamed everything, and I just need to agree with that and it's done.

Doug Cullen (15:03)

I mean, this is not to be overstated in terms of the power that something like Blueflame and Claude bring together in this great marriage, because I've been in many a data room throughout my life.

I'm sure we've all encountered data rooms that have poorly named folders. And then once you get inside the folders, you get even more poorly named folders. Then maybe you get subsets of documents that aren't properly named either. How many "IB Book 08.xls" files have I had to click into and look at in order to figure out what year this is from a financial operating perspective? Here, you're demonstrating such a valuable light touch in terms of saying, let's make sure there is consistency in how this document is named. But we're going one layer deeper in value and presentation, looking into the content itself, extracting key components like the type of legal document and ultimately the year.

I think that both accelerates the setup of the data room, and I can also imagine, since you've been on the buy-side a lot, what some of your experiences have been in trying to sift through massive amounts of content in order to find the information that you need to get going.

Alice Esmerian (16:26)

Yeah, absolutely. Really, having this ability to enter a space as a buyer and reviewer and automatically see the level of care that has been put in place, and the work that has been done by the banker and their team to create the best buying experience, is already setting you up for a successful process.

I feel that this is value-added for the person who is going to have more time to think about deal prep and the content of the data room rather than being stuck in the nitty-gritty. But also, as you pointed out, from the buyer's standpoint, entering a space that feels organized and clean, where you can easily identify the data you want, the year you want, and the name of a particular customer or vendor that is clearly labeled in a contract, is already setting you up for success.

Doug Cullen (17:16)

So that's a great value, and I always think that the quality of the organization, and the quality of the data room, is a reflection of the quality of the process and ultimately the underlying asset.

So I think, from the preparation side, on the sell-side, it's super important to understand the documents. LLMs are very good at helping us do that, and Blueflame is even better at organizing the content in a very digestible way, allowing you to really get to what you're looking to accomplish in a deal, which is typically looking to understand the underlying asset. Is this something that is consistent with your investment thesis? Is this something that is additive to your corporate portfolio? You're really trying to get through that initial phase of diligence in order to figure out whether this is something you want to move forward with.

But on that side, I think the next thing we're going to talk a little bit about is Q&A, one of the most painstaking parts of the process. What are some of the things you've been able to leverage via MCP and Blueflame to make Q&A more seamless, effective, and efficient?

Alice Esmerian (18:32)

I wouldn't say a little bit more; I would say immensely more. With Blueflame enabled, having access to all this content being aggregated, reasoned upon, and queried in an unlimited way is super helpful.

As a banker or someone preparing a data room, you're going to be able to track the information requests and whether the information has been provided or not. You're also going to be able to populate questions, whether they're from your vendor diligence providers before launching your process or further down the road as you start receiving unstructured questions from different buyers and all their advisors. It's all flowing through email. It's hard to really track. For each of these, you have someone from the sell-side team prepopulating draft responses, then going to the management team for sign-off. That's a long process, and here you can submit your question to the LLM.

Because Blueflame is layered in with all these skills that we're bringing together, it knows how to respect a banker tone, knows where to locate information, and knows how to provide the right source and level of confidence to the reviewer. Obviously, everything we're showing here is not replacing any human work. It's more about enabling and empowering our users to have this layer of human review and human overarching view that is just accelerated and simplified. Here, the example we're looking at is uploading a Q&A tracker, but that could also be a data request list, or DRL.

Everyone has different references, but you feed in an unstructured set of requests and the model understands what they are and what it is supposed to do. It organizes them here and sets them up by workflows: legal, tax, technology, etc. It found the right information within all the content of your VDR and then prepopulated responses and sources. One thing I'm very adamant about is zero hallucination. The skills really prompt the model to be highly transparent about its ability to respond to a certain question. You can see here that of my 72 questions, the model clearly tells me that it is confident about 51 of these.

For 20 of these, it's partial, and for one of them, it couldn't answer. You then have the option to review that, whether it's in an Excel tracker or a dashboard. Here, when I open my dashboard, I have all my questions with their level of confidence from the LLM. As a banker, I can jump into this where the lowest level of confidence is. When I open it, I have the draft response leveraging, in this case, an ARR bridge. It leverages the customer revenue cube and the financial data, and all my sources with the VDR index. It's an immense time-saver.

You can just cross-check the information, check that the document reference is right, and you have the first pass of responses for all your Q&A. So that's a massive time-saver.

Doug Cullen (21:36)

And I don't have to tell you this, but this is probably one of the most crucial parts of the process because we've got these questions, whether you're on the buy-side seeking answers from the sell-side banker or a different advisor sitting in the middle. You're doing your best to involve all the subject matter experts around the organization in order to answer this. With this capability, you're able to query the content itself, the content gets cited, and you're putting together a confidence score based on that.

So you're really taking potentially 500 or 600 prospective questions and narrowing in on the ones that probably need a little bit of extra attention. I've also heard bankers really doing this preliminarily to get familiar with the content, almost running this as a simulation: what customer questions do we anticipate will come? They're getting familiar with that content and almost preemptively looking at what queries may be coming up to get even more prepared.

Alice Esmerian (22:37)

Yeah, which is a great transition to our last workflow that we'll be showing you today. We did the prelaunch readiness.

This is something that we felt, as old practitioners, would have been really helpful when you're about to push live on your data room. You always have this moment of stress and anxiety, thinking, "Is there anything that actually doesn't tie out? Is there a document that is corrupted? Is there any information that shouldn't be there?" So we packaged this skill that we call prelaunch readiness, which will cover three main elements. The first one is a gap analysis. For each section, it is going to have a sort of reference and know if you might be a bit light on some topic.

That will prompt buyers to immediately react with information and additional requests, so that's being looked at. The second one is document quality. It will look for any corrupted document, any blank scan, any unsigned contract, and more importantly, any PII information that has not been redacted and is available in your data room. The last section is a risk review. We put together a list of potential risks for tax, finance, tech, IP, etc. The model and Blueflame are going to look at your content and cross-check any of these risks and assets.

Is there anything that is going to prompt the buyer to think about the potential for the assets you're presenting? All of that is treated and then available in a Word document where you can see the overall context of your project, the model's recommendation, and each section where it sees the most need for attention. I think that, as an old practitioner, this would have given me an extra layer of comfort. It's like having an extra person in the room, an extra pair of eyes that gives you confidence in the last mile before pressing publish.

So I think this is another great way of leveraging Blueflame within your data room.

Doug Cullen (24:38)

Yeah, very powerful. We've always talked about Datasite broadly as an extension of your deal team, and I don't think that has ever been more true than it is right now. Real quickly, and then we'll go into the Q&A, we've got some great questions here, so I'll be able to get to hopefully a few of them. Can you just quickly cover, Alice, the difference between maybe a connector and a skill?

We've also taken the liberty of publishing specific dealmaking skills into the Claude environment. Can you quickly cover the difference between connectors and skills, and where can people find the skills if they're looking for them?

Alice Esmerian (25:15)

Absolutely. The way this works is that whichever LLM you're connecting your MCP into, you will rely on it to have access to this indexing and names and then do the reasoning. When you enable Blueflame within your VDR, Blueflame is going to do the work of reading the content and analyzing all of this content.

When you add the skills on top of that, you're really providing guidance as to how each workflow should be processed. You're removing any risk that your LLM will not read the information the way it should, and you're providing guidelines and guidance for the model to do it the way your firm, or you yourself, would expect it to be done. How we built these skills is really by thinking about buyers, the pitfalls you want to avoid, and the information you want to make sure is surfaced and triangulated together. That really adds additional value to your workflows.

Doug Cullen (26:11)

Yeah, awesome. The superpower is the connectivity between the data, the data room, and the MCP environment via Blueflame, as well as these skills. You can find these skills in our GitHub repository by searching for them. You can add those to your Claude environment, and that will give you the benefit of our dealmakers having prebuilt several of these based on some of our experience. Okay, we have a lot of questions and not as much time. The best question I had was: is there going to be availability of a recording from Jesse?

Yes, this is being recorded. It will be available. Please access it and pass it along. Let's propagate and help build more prompt-first dealmakers around the world. Okay, question: do you anticipate Datasite MCP for Claude being used outside of IBs, such as consulting companies being able to query a Datasite project for specific pieces of information?

Alice Esmerian (27:17)

Totally. We are fully cognizant that we don't want this to be just for our banker clients here.

This is going to be available to the reviewer side, so the buy-side and all of their advisors. As a banker, as you onboard more advisors or even your client, you'll also have the ability to give anyone who has access to a Datasite account and an LLM access to that.

Doug Cullen (27:41)

Yeah. So we are hyper-focused here on more of a sell-side workflow, but we will probably do some future programming around the buy-side. There are a lot of questions about what is available. There is a whole set of capabilities within your Claude connections environment that are available to you if you are on the buy-side or the sell-side, in other words, an administrator or a reviewer.

So you can see that again. One important thing to understand is that the project itself must have the Blueflame AI Assistant enabled, which is something that the sell-side would set up in order to extend a lot of the powerful search capabilities and be able to look into the documents there. We don't have a ton of time to talk about that, so I encourage you to reach out to Datasite via any of our channels and we can walk you through it in great detail. Or maybe I'll do some follow-on content around that. But this is sell-side focused. Just to avoid any doubt here, 88% of the viewers of Datasite are reviewers.

So we have a lot of familiarity with these workflows across the entire dealmaking community. This webinar happens to be sell-side focused, but let me see if I can blast through one more question. Let's see. Let's just go back. Does batch-uploading question lists by function or other segmentation yield better responses by managing context?

Alice Esmerian (29:21)

I haven't tested that, but I don't expect it would. In any case, whether you batch-submit your questions, or just the tech ones, the LLM and Blueflame will always make sure that there isn't any other document or information available in a section that is not specifically labeled technology.

So it is going to do the work of almost scrubbing every single piece of information for any question. I don't anticipate it would necessarily get you better outcomes, although it might accelerate the timing.

Doug Cullen (29:53)

Yeah. And I would just offer as well that one of the powerful things here is that these are probably the worst these models will ever be in terms of our ability to interact with and connect to Datasite. So I anticipate a lot of additional functionality and capabilities developing. We think context is super relevant, and we also feel that Datasite is at the center of really understanding these workflows as well as anyone out there.

So with that, we're at time. Thank you for joining us. This is our first in a series of webinars, and I think we've got a couple of other programs coming. Check out the recording. Make sure that you become a prompt-first dealmaker. We certainly aspire to be prompt-first dealmakers. We've been utilizing these capabilities ourselves, and I'm amazed every day by how much better things get. Make sure you understand that this is a true leapfrog. We think the models are very accessible, and the conduit within Datasite makes sure it is secure, compliant, and remains auditable.

Hopefully, we've earned your trust through the years, and we certainly want to maintain that moving forward. Alice, I want to thank you for joining us and bringing to light some of the capabilities. Of course, thank you to those around the world who joined us for this webinar. We've got a little bit of an announcement at the bottom around upcoming webinars. I think the next one will focus on a different frontier model, but a lot of the content will be somewhat similar. We'll try to learn and hopefully make sure we're giving you what you want out there.

Thank you for joining us in the next level of dealmaking. Thank you, Alice, and we look forward to seeing you on the next webinar. Thanks, everyone.

Doug Cullen (00:00)

Good morning, good afternoon, and good evening to those joining us around the world. My name is Doug Cullen. I'm Chief Strategy Officer here at Datasite, and it's my pleasure to welcome you to the Datasite MCP for ChatGPT webinar. This is the second in our series, where we walk through the opportunities now available because we've enabled and connected Datasite, the world's leading VDR, to various frontier model platforms via MCP. Today we're going to talk about ChatGPT, which is a fantastic platform.

Before we get going, I just want to remind everyone of a few things. Okay, first of all, we want to hear from you. Please ask questions in the Q&A panel throughout the session. We have reserved a little bit of time toward the end to get to those questions and answers, so please ask them. We also have a survey, and we'd love to know what you think. There are additional resources around the console as well, including some PDFs, documentation, and frequently asked questions.

Please check there for the resources. This session will be made available on demand, so if you have colleagues who weren't able to attend live, or if you want to replay it to understand some of the things that we demonstrated, please feel free to do so. And the last one is: who doesn't love a good legal disclaimer? I will read this one out here, just so you know: all opinions expressed are our own and not those of Datasite or Blueflame AI. So with that, Alice, thank you so much for joining. I'd like to introduce everyone to Alice Esmerian.

She's a product strategist at Blueflame AI, and I know you've been here for a few months. If you don't mind, please give the audience a little background on what you're doing at Blueflame AI and what you did prior to joining us.

Alice Esmerian (01:53)

Sure. Thanks for having me today. Yes, I joined Blueflame six months ago on the product strategy team. My background is in investment banking and private equity, where I spent almost eight years. It's great for me to have the opportunity to work on products that really enable investment bankers and private equity professionals to leverage AI in their workflows. I'm super excited to be here today and talk about a few workflows around the MCP product release.

Doug Cullen (02:25)

Awesome. We want to get into this; it is definitely show, not tell. There is a real step-change moment happening in dealmaking. I've been in dealmaking for almost 20 years now and have been watching a lot of the transformation across how deals get done. I run corporate development, and we've done nine acquisitions in the last couple of years. I think this is the greatest time to be a dealmaker, when you think about what we can take advantage of with some of these new technologies.

But before that, one of the things we wanted to do is talk about how we launched MCP at DealMAX. This was a big deal for us here at Datasite, allowing the first-ever connectivity from a data room, our Datasite Diligence platform, and our Acquire platform to MCP. We'll be demonstrating enterprise ChatGPT, one of our favorite platforms, here today. It's a new age, as we're talking about. Raj introduced the concept of a prompt-first dealmaker, which we think is amazing.

One of the things we also focus a lot on here is making sure this is done in a proper way. This is a serious business, right? We want to make sure that it's offered in a compliant, resilient, and effective way because that matters a lot in deals.

Alice Esmerian (03:45)

Yeah, totally. We're seeing it a lot today with our clients at Blueflame. When I joined six months ago, I could see that people were still experimenting a bit with AI. They were trying to understand where it sat in their organization, and which workflows were applicable or not.

Today, things have changed. Every organization is really trying to deploy AI consistently across the firm and across workflows. It is really important that we enable them to power AI while staying totally safe and maintaining all compliance and governance that are essential to executing a deal safely. The data they are dealing with is so important and valuable that we really empower them to do that. It is very exciting to leverage this MCP connector to bring this safe AI connectivity into Datasite for the first time.

Doug Cullen (04:36)

Yeah, it's an incredible time. Let's transition here quickly. We want to think about how AI is actually used. We want to demonstrate some of that and talk about it securely inside the data room. This is one of the most important things. We launched the Blueflame AI Assistant within Diligence a few weeks ago, and that is the vehicle we use to connect MCP. We just wanted to ground some of the things we're talking about for the audience because this came up earlier this week.

I wanted to make sure there are a couple of core concepts. One is MCP. That's what we're talking about here. This is the connection. We've built an MCP server and published it via ChatGPT through our great partners there, which makes an available connection for any enterprise ChatGPT user to connect to and enable their data room. This is always going to respect the role you have, whether you are an admin or a reviewer, and certain things are going to be enabled via that MCP connection.

The other thing that really makes the MCP connection even more valuable is that the content within the data room has been AI-enabled. You're seeing here the Blueflame AI Assistant. If you have not seen it yet, within Diligence, our core platform, you have your standard Diligence platform with a small right panel and an icon in the upper right. You click that, it slides out, and it gives you the full agentic experience of Blueflame within the trusted environment of Datasite. So you've got the Datasite project, you've got the Blueflame AI Assistant running within the Datasite project, and then that project itself can be connected to via MCP in enterprise ChatGPT.

I just wanted to anchor those concepts. We'll be showcasing the MCP connection today because that is the purpose and the new innovation. But to reinforce and underscore it, that is powered by the Diligence project that has the Blueflame AI Assistant enabled. Then there is Blueflame, the enterprise company, which is used by private equity firms, investment banks, and corporate development executives that have a separate Blueflame subscription. These are the three dimensions we want to talk through. This came up on our first webinar, and we wanted to give a little more grounding in terms of what we're going into.

With that, we can probably shift into the show, not tell, aspect of this and highlight and talk through some of the powerful use cases that a platform like ChatGPT brings to bear. One of the most challenging things when you get going is that you have all this content and are trying to figure out a way to set up the data room. You've done it many times in your life. What we're going to showcase here is probably the most profoundly efficient and effective way, as an administrator, to create and set up a VDR.

So why don't you go through a little bit of what we can do with ChatGPT?

Alice Esmerian (08:11)

There are plenty of things we can do today. We're going to run through four workflows that are essential to the admin work within the data room. There are plenty more workflows that are applicable. If you have any questions, ideas, or suggestions, feel free to add them to the Q&A section. We'll look at them afterward. Today we'll look at four different workflows. On top of delivering this MCP connector and Blueflame enablement, we've built eight skills, which are essentially, for those not too familiar with this concept, almost like eight chef's recipes that help the model be even more empowered with additional guidance on how to understand queries around data room-specific workflows.

These are going to be available at the GitHub link that will be shared with you in the resources for this webinar. I just wanted to preface that because these skills are going to power some of the workflows we're going to look at today. The first one here is a simple one. When we go back to this dealmaking prompt, you can see that we're not starting with anything sophisticated. We're essentially asking the data room, "Can you start this project for me?" and providing it with a bit of context.

Models tend to operate better when they get a bit of context around the task you're asking them to carry out. Here, we're just setting the stage with a bit of information around the company, its industry, and its size. Immediately after asking for that, the model is going to proactively share an index suggestion, which is powered by our indexing skill. Here, I'm getting a detailed index with subfolders as well, and I can immediately provide some feedback around it. You can see how a task that would have taken a lot of time, whether putting together a spreadsheet index or just reorganizing folders and naming them while making sure everything was clean and had zero typos, is done from a simple prompt in a few minutes.

Doug Cullen (10:16)

Yeah. This is something where dealmakers are very often baptized by fire. They may or may not have ever set up an index. What does a good index look like? That is so paramount to the success of a deal because one of the things that I believe really reflects the quality of the asset and the quality of the advisors that people have hired is the way the information is reflected about the underlying company. It's that adage about a tidy house: you really want to make sure everything is tidy, easy to find, easy to locate, and follows a very logical setup.

This is just a super powerful capability that AI can help us with, and ChatGPT in this instance can just give us a great, clean index.

Alice Esmerian (10:59)

Yeah, absolutely. You enter, and you automatically have a great first impression of what you're going to experience within this data room. The index is set up. Here, I made some suggestions for editing given the particular context of that company. The model automatically understands it, and I can then just ask it to push it. The model comes back in 53 seconds.

You can see on the screen that my 14 folders and all the subfolders have been pushed to Datasite. That is a massive time save. That then gives me a lot more time to do everything that is more valuable for me as an admin creating a data room.

Doug Cullen (11:40)

Yeah, I mean, it looks kind of magical. It is kind of magical to get that content and put it in that very logical place with a logical order. Now I think we're going to go into maybe some file naming conventions.

Alice Esmerian (11:55)

Yeah. Typically, when you start your data room, you have your index set up, you can push your documents, and then you just want to make sure that every document is clearly labeled. There are two ways you're going to do that. One is looking at the existing name and cleaning it up, which is totally doable with just the MCP before you enable Blueflame, because the MCP connector has access to every single file name. For example, if the company we're working on here is called Voxel Matter, and it sees capital letters, no capital letters, VM instead of Voxel Matter, or just clunky names, it will be able to clean that up, clean the typos, etc.

What it's not going to be able to do until you enable Blueflame AI is actually access the content of the document and leverage that content to inform how to rename a document. A good example here is that I wanted all my contracts to have the year in which those contracts had been signed in the name so that I could clearly orchestrate the content. To do that, you obviously need to enter a document, read it, and pull the year into the name, which the model was able to do because this data room is empowered with Blueflame AI.

Doug Cullen (13:10)

Yeah. A couple of steps here, just to rewind the tape. One is to get the name of the document consistent, which in and of itself doesn't always happen. In many a data room I've gone into, I've seen the famed Book 007.xls. What the heck is that? Who knows? Is it a financial model? Is it part of a customer queue? We have no idea. So getting the name into a logical place is super important.

As I'm clicking and doing the diligence, either in preparation before opening it up to the buy side or ultimately for my buyers, prospective buyers, lenders, or investors, I'm able to see what that document is before looking at it. Then we take it up a notch because we're actually able to look at the content of the file itself and read that content. In your great example, it's, okay, I know this is a contract. Okay, great. It is a contract. Well, what type of contract is it?

We're using the ability to look at when the contract itself was executed and apply that as part of the naming convention.

Alice Esmerian (14:23)

Again, as someone working in a data room, it's very time-consuming to do this, and it's not highly rewarding because you're really just searching for information and spending a lot of time checking for typos or inconsistencies. This is something you can automate so that you can spend more time on more valuable and informative tasks to get your process ready and prepared in the best form possible.

One thing we potentially wanted to highlight for people watching is that when we say Blueflame enables you to access information within the document, there is absolutely zero information retention.

This is why Blueflame is particularly powerful and safe in a deal context, because highly confidential information will never be retained by Blueflame.

Doug Cullen (15:12)

Yeah, no retention and no training, right? In the early days, people were really worried about this highly sensitive content being put into LLMs to allow the LLMs to be trained. There is nothing like that going on here. We are a financial services powerhouse, and we understand that the content is highly confidential. So there is no training on the content and absolutely zero retention of any content that is made available via this platform.

Alice Esmerian (15:49)

Yeah. Moving on to another workflow that honestly cost me a few hours and nights back in the day is a Q&A process. I think this is a great addition and a great way to leverage access to the content of your documents. In my example here, I attached a list of Q&A questions. This could happen early in your process, when your VDD providers start asking questions and making information requests around the asset. Later on, when you start receiving plenty of questions from different buyers, their advisors, and everything comes to you in an unstructured manner.

You can then just drop in the questions. Here, you can see I attached an Excel file with unstructured data. The questions are not organized either by theme or by advisor or buyer. Then I'm just asking the model to prepopulate draft answers to all of these questions. This is also powered by one of the skills that we're providing. Something I'm very adamant about in everything we do with Blueflame for our clients is making sure we have zero hallucination. This skill prompts your agent to really look at the context and the content and make sure it provides an honest opinion about its ability to answer the question, as well as a very detailed VDR link and the section of the document that it used to prepopulate the draft answer.

To me, again, this is a very good way of trusting AI safely and having a partner that works with you hand in hand, helping you save time and direct yourself in the VDR while having all the guardrails that enable you to trust and feel safe while using it.

Doug Cullen (17:35)

I actually had the benefit of showcasing this to a managing director here in New York for a very large investment bank, and he was reflecting, "This used to crush my weekends." The availability of taking questions, essentially uploading a set of questions and allowing them to be answered by the content of the data room, is powerful.

Questions are being asked, they're being uploaded, and they're being answered with cited content. We are literally putting the citation to the document that supported it, as well as, just to reiterate what Alice said, a bit of a confidence aspect as well: a confidence score. For some questions, the model is going to be highly confident in its response based on the document and the framing of the content. For others, it will not be as confident. We're able to discern these two categories to narrow in, from a preliminary standpoint, before the asset goes live. You want to understand what questions people are likely to ask and make sure you have the appropriate information in the data room in order to answer them.

And then, of course, when you're live, the orchestration of Q&A is probably one of the most complicated tasks we have when doing a deal. You've got various questions, various stakeholders, and each person has their own series of experts: HR, legal, compliance. We need to orchestrate that question from the buy side to the sell side, to the expert, and back.

Alice Esmerian (19:16)

Totally. You can see here on the screen that the model took all the questions and then batched them by theme. Here on the screen, I'm showing you the commercial section. For each question and topic, it can provide a status, whether complete or partial.

For every draft response, you have the document or documents that it used to address it, as well as the page reference. This is then all aggregated by theme into a single spreadsheet that you will see here on the screen in a few seconds. It is organized nicely by workflow with the number of questions and their level of completeness. Then there is the total Q&A tracker, which, for each single question, includes a draft response, a status, the source reference, a citation, and a follow-up question required or additional data needed.

To me, as a banker, this is a great way to really start the Q&A process and accelerate it while keeping it valuable and high quality.

Doug Cullen (20:17)

Yeah, this frames the problem. It is super valuable for the person helping from the advisory perspective, whether that's a law firm, a bank, or another financial advisor, but it's also super valuable for the corporate. Having been involved in selling the company a couple of different times and selling different assets, as a senior executive, you're super curious about how the process is going and what types of questions are coming up. Being on top of it is the sign of a great advisor.

Being able to give this type of report to someone like me or Rusty, our CEO, is something I really can't overstate in terms of how valuable that is to the underlying corporate. That's the person behind the deal.

Alice Esmerian (21:02)

Yeah. Accelerating reporting and communication with your client, as you mentioned, is something that is now possible with this MCP connection because you can extract activity from the data room. One more thing I wanted to show today that ties to that is something I think a lot of people here will have some sensitivity to.

Publishing a data room for the first time, when you add buyers in, always comes with this little moment of anxiety when you wonder, "I really hope I didn't leave anything contradictory that doesn't tie, or just a document that won't open properly. Are there any risks? What are the risks potential buyers are going to flag immediately?" That's why we put together this skill called the Prelaunch Readiness Report. It's going to focus on confirming three main areas of your data room. The first is that it's going to run a gap analysis.

It will look at all of your sections, keep in mind the context it has around your deal, industry, and size, and provide feedback on how complete each section is. Are there any areas where you might be lacking substance or context? It will provide feedback around that. It will then also look at the quality of the document. It will run through whether any document is blank, any contract is unsigned, or any PII information has not been redacted. That could create a lot of risk around your data room.

Finally, it's going to run through all the potential risks that we provided in this skill around every single area of tax, finance, operations, IT, and IP. It will cross-check all of those and give you feedback on where your largest risks are and where buyers are going to pay the most attention. As an admin, that really gives you time to anticipate what you need to prepare in your defensibility approach and how you communicate this asset to potential buyers. All of that is then consolidated into a single Word document that you can download and share with your client and team to make sure everything is ready.

Doug Cullen (23:17)

I remember hearing a story from a managing director who described to me that every time before she launched a deal, she had to literally go through every single document. That would take hours because ultimately she was the one representing the asset and taking it out. Think about the value from a managing director perspective, a senior banker, or maybe even your analyst or associate who has to do this. I'm sure Alice's pre-readiness report was fantastic back in the day, but this is something that is just available out of the gates, both within the data room from the Blueflame AI perspective and in ChatGPT.

I mean, what a powerful tool for both, really. Everyone across the board, whether you're the corporate, a senior advisor, or someone on the deal team itself, can benefit. It is fantastic.

Alice Esmerian (24:08)

You have this extra pair of eyes, as if someone were really working with you while you're doing something else and reviewing the content for you. That's very exciting, and I really believe that everyone working on sell-side data rooms will find great value from using this.

Doug Cullen (24:25)

Yeah. I think this is changing a lot.

I was talking to someone just yesterday, a very senior banker, and he joked, "I actually may log into a data room again for the first time in 10 or 15 years." This has been a ritual among advisors where the junior teams tend to interact with this. But now, with natural language querying and the availability of these capabilities within the data room, you can log in, ask a few questions, and get that answer in your preferred platform. You can log into the data room or the Diligence project. You can log directly into ChatGPT, and you get that same immediate availability of key information and context. It’s incredible.

Alice Esmerian (25:06)

It is great. Although I think everyone is starting to get used to the speed of AI and ChatGPT, everyone is so concerned about accuracy and getting confidence around the content. This is really a great way to get in there.

Doug Cullen (25:22)

Yeah, this is an incredible capability. We're having a lot of fun. We're in the early days with this. Just to reinforce, this capability launched about three or four weeks ago. The feedback has been tremendous.

We're looking forward to a few questions here, which I think are great. One way to think about this, too, from a broader perspective, is that this is the worst the models will ever be at answering these types of questions.

Alice Esmerian (25:47)

Yeah, it's changing every day. We are keeping up with all the innovation here and making sure that our prompt-enabled dealmakers are getting the best of the tools.

Doug Cullen (25:57)

Yeah, you were saying earlier that you were going to leave today and test one of the latest models from ChatGPT. OpenAI has been great with us in terms of sharing its latest models, and we usually get our hands on those a bit early in a preview session. This is like your Christmas Day, right? You get to open up new presents.

Alice Esmerian (26:19)

Now that happens more than once a year, so it's great.

Doug Cullen (26:23)

Hey, let's take a look at some of the questions. We have a few here that I'm going to try to fly through.

I saw one earlier: what kind of audit trail is available on these actions? The answer is a full audit trail. We're able to capture everything that's been done, all the queries, and all the information, and that will become part of the reporting package that is available in Datasite. Let's see. Another one: how strong do you need to be with LLM prompting to use something like ChatGPT and Datasite?

Alice Esmerian (27:04)

It's going to depend, but honestly, the way we've built it is, as you've seen from all my prompts on the slide, fairly straightforward. As long as you're able to convey it the same way you would ask someone on your team to do a task, you can really just use that, and the model is going to do it for you.

One piece of advice I would give is that iteration is very powerful when it comes to handling AI products. Try something. If it doesn't quite work, refine your prompt. Nothing has been built in this MCP connection that requires any particular prompting knowledge. Any investor with no prior AI knowledge is able to leverage the most of it.

Doug Cullen (27:45)

Yeah, I mean, I think like everything with these LLMs, you just have to jump off the curb and try it. Your framing is really good: as you would speak with a colleague, as you would seek advice from someone else, seek advice and see what types of responses you get.

Then you have the ability to do things with follow-up questions, which are super important. Let's see. Is it possible to give buyer advisors, for example, limited access to ChatGPT output if you do not want to share full access? These are dynamics that we're thinking through right now. To reinforce, a lot of the capabilities you're seeing right now are really for administrators of the projects only. Do we intend to bring some of these capabilities to reviewers and buyers? The answer is yes. We've launched this, as I said, about three or four weeks ago.

We're still in the early days. These are administrative capabilities that we're demonstrating right here. There is a list of capabilities that you have no matter what your role is, such as asking questions like, "How many projects do I have?" They are all very well defined and articulated. Honestly, you can just ask ChatGPT, "What am I able to do with Project Dragon? What am I able to do with Project X-ray?" and it gives you an amazing response directly in there. Okay, last one here, and this is certainly worth reinforcing. My question came after reviewing the material fact sheets and FAQs you shared. It seems that the client's data does not leave the four walls of the Datasite environment. Exactly.

We took meticulous care to make the client content stay within that. I would describe this as the secure boundary and barrier that is Datasite, and all of this is still within that secure operating environment. The extension out to MCP also is governed and needs to operate by those same sacred rules. That is how it works. Anything else to add from that perspective?

Alice Esmerian (30:11)

No. Again, it's a semantic search that goes into Datasite without pulling anything from it.

This is how you operate in the safest way while leveraging AI in your data room with Datasite.

Doug Cullen (30:23)

Yeah. We've gone through all the security reviews, etc., as you might imagine, and we've taken meticulous care in bringing this capability into the incredibly secure operating environment that is Datasite. You trust us. We want to make sure we're handling this in the appropriate way, and we'd like to think that we are. With that, we're at time. Alice, thank you so much. It was amazing.

Thank you, ChatGPT, and our great partners at OpenAI. We do have a few other upcoming webinars. We're taking a little bit of a break as we approach Memorial Day here in the U.S. and probably bank holidays all around Europe and the U.K., but we'll be back with our next series. Pay attention to those. A reminder: this will be made available. Thank you for your Q&A. We did not get a chance to answer all the questions, but we will wrap those into some post-webinar follow-ups in the form of blogs as well as FAQs.

Look, it's never been a more exciting time to be a dealmaker. The capabilities available via large language models and, of course, via incredibly focused, almost vertical-specific capabilities from someone like Blueflame, brought into the trusted environment of Diligence, are really changing and transforming the way deals get done. We're getting to prompt-first dealmaking, and ultimately we want to make sure that we're delivering these results in a way that people can trust and depend upon. It is AI, so you have to double-check things before making them available.

Alice Esmerian (32:06)

It goes without saying: human in the loop, always. But it's really a way to accelerate your dealmaking.

Doug Cullen (32:13)

Thank you once again, Alice. Thank you to the hundreds of people around the world who attended this webinar series. We're so pleased to be able to bring not only the innovative technology, but also this practitioner perspective to help you explore the world of AI. With that, I will thank you and wish you a wonderful day. Thanks, everyone.

Doug Cullen (00:00)

Good morning, good afternoon, and good evening to those joining us from all around the world today. Welcome to the Datasite MCP for Microsoft Copilot webinar. This is the third in our series. All the other sessions are available on demand, so please go ahead and check those out if you are interested in either Claude as a surface for interrogation or ChatGPT as a surface for interrogation. Today, we're here to talk about Copilot, Microsoft Copilot to be specific. My name is Doug Cullen. I'm Chief Strategy Officer here at Datasite. I also oversee corporate development, and I've had my hands in most of the acquisitions that we've done over my last decade-plus here with Datasite.

I'm here to introduce today's topic, which is really about framing how deal teams can work with AI tools like Copilot while keeping sensitive deal content protected inside governed environments. Before I kick things off and introduce my co-presenter in today's webinar, I just wanted to remind people around the world about a few housekeeping items. First of all, we want to hear from you. There is an ability to ask a question on the panel, so please ask your questions. We hope to have about five minutes toward the end of the discussion to address them.

We want to get to your questions, whatever you have, from all around the world. The other thing is that we're going to send a survey. We'd love to know what you think about this program and past programs, and to hear areas for improvement. We are always in a continuous improvement environment here at Datasite. We want to make sure that we are giving you what is valuable. Speaking of what we hope is valuable, there are a bunch of additional resources in the console, including some FAQs and guides along the way to help you on your AI journey specific to MCP for Copilot.

Just like the previous sessions, this will be available on demand, so please forward it to your colleagues and ask for additional access. We'd love the world to understand what we're doing here at Datasite, and specifically how we're powering up Microsoft Copilot via our MCP connection. And the quick legal disclaimer: all opinions expressed are our own and not those of Datasite. So with those behind us, I would like to introduce my friend and colleague, Caitlin Murdy. Caitlin, you're VP of Sales Engineering here at Datasite. How long have you been with us? What do you tend to focus on? Help our customers understand.

Caitlin Murdy (02:57)

Yeah. Hello, everyone. Like Doug said, I'm Caitlin Murdy. I run our sales engineering team. I have been with Datasite for six years now. I've seen Datasite over the transition of the past six years, bringing new products to market, working closely with our product team and our sales team to bring those products to market, receive client feedback, bringing it back in, and continuing to evolve what we offer.

Doug Cullen (03:22)

Yeah, so Caitlin and her team, we've got a global team that does this as well, sort of sit as a linchpin between our customers, our go-to-market teams, our service teams, and our product teams.

You've been a great partner helping us build what we think are pretty amazing products over the year. Thank you for that. I think we want to transition a little bit, and it should be stated, just to give people context, that we are currently in process with Microsoft to get this capability approved. So we have the capability, and we're going to be able to showcase that capability. I'd love to be able to tell you that you can go and find this capability immediately, and sadly that's not totally the case. We're going to get a little bit of a foreshadowing of what will come.

We're really more in a production environment that's available for internal folks here at Datasite to at least be able to get a sense of directionally where we're heading with this capability. Don't drive yourself crazy if you continue to look for this as we're going through the final stages of approval by our dear friends at Microsoft. With that as a backdrop, I just wanted to talk a little bit about the overall dynamics that are happening across dealmaking, as we covered in the other sessions. AI is omnipresent, and it's really changing some of the ways that we tend to think about how dealmaking is getting done today, how it's going to get done tomorrow, and what tools people have at their disposal.

So I guess from your perspective, Caitlin, before we dive into the nitty-gritty of some of the demonstrations, what are some of the industry trends you're seeing? You're out on sales calls, you're showcasing some of these capabilities to customers all over the place, across corporate development, private equity, and investment banking. What are some of the backdrops or points that you're hearing over and over again about industry dynamics?

Caitlin Murdy (05:23)

Yeah, I think over the course of the past year, there has been this big transition from people being very scared of AI to people feeling that they need to adopt it.

They have to figure out what tools to use, but they are still getting more and more comfortable. Every month that goes by, I think we see more and more teams adopting tools. But there has definitely been an underlying tone of which tool to use, as you have people who really want to use Claude and then all of a sudden ChatGPT gets ahead with something else. So there is this ever-changing dynamic going on, and it's tough for teams to decide which tools to use and where to commit.

Doug Cullen (06:02)

Yeah. Just as a refresh, a lot of these capabilities that we're showcasing today are very new to market.

We actually launched the MCP connections at DealMAX earlier this month and brought out the first ever kind of Blueflame-enabled AI Assistant within Diligence and Acquire. Those are some foundational elements. I think one of the things that I find fascinating has been some conversations that we've had along the way, and really this notion of prompt-first dealmaking. As you're saying, AI is certainly omnipresent. I think people are in various stages of their journey embracing AI, like many people around the world. Microsoft is omnipresent for a lot of us dealmakers as part of the critical software suite that we have at our disposal.

So I think Copilot is certainly an interesting addition to our set, but one of the things that we've been thinking about a lot is ensuring this level of governance and compliance. Can you give us some perspective from your side in terms of how we've been thinking about AI, how clients tend to think about AI, and how we approached it from a strategic standpoint in terms of activating within our various surfaces?

Caitlin Murdy (07:25)

Yeah, definitely. I think first and foremost, we are a security-first company.

We want to make sure we're doing right by our clients and protecting their data. Everything that we're rolling out is with that mindset. As we think about our MCP connectors into the data room, it's really to allow you to protect that content but leverage AI against it. Through my conversations with clients, I think that's, one, a gap in what they have available today because they don't have tools to leverage AI throughout the diligence process. And two, it's always one of the first questions we get around protecting their data. So I think that's super important.

I'll also say, because you mentioned Copilot and that's the theme of today, that Microsoft is ever-present for so many people. We get a lot of excitement around Copilot and what we're doing here because I think for a lot of clients we work with, Copilot is kind of their first foray into the AI world.

Doug Cullen (08:26)

Yeah, totally makes sense. I think what we're really trying to get to here is showcasing what AI can actually do securely within the data room. Today we're going to showcase the MCP integration with Copilot.

Then we'll talk a little bit about our Blueflame Assistant. This is active again; we've got it in ChatGPT, so you can find those as connectors. We're in the process of getting approval for Microsoft Copilot. With that, I thought it would make sense to level set across a couple of different areas that we're going to frame today. One of these is that you're going to hear me talk a little bit about the Blueflame AI Assistant. This is something that is available today for both our Diligence application and our Acquire application.

As a person running that process, whether as an advisor, the underlying corporate, or even a private equity company dealing with a potential divestiture or portfolio sale, you can elect to have the Blueflame AI Assistant enabled in your project. That's actually a side panel that's available today, and it will allow you to have a full agentic experience within the Diligence project. We'll be covering some of this on future programming, particularly around Blueflame and the abilities that we have via that MCP connection as well. But then you're also layering in this additional layer.

This is MCP, right? Think of MCP as a different surface, something that you can use. In this instance, we'll be talking about Copilot. We've got this surface that is actually connecting based upon your credentials. What you are authorized to do within the platform, what projects you have access to, and what you are able to do within that project will determine how this connector gives you this beautiful surface of Copilot, which gives you another way to act on the data room content. We'll showcase what those things are able to do. So we've got the Blueflame AI Assistant within the project experience.

You can go there and activate it. Then we have some customers who are seeking to do things outside of the traditional Datasite experience, and they want to be able to interact with content via MCP, which again we're going to showcase. You can do it within Datasite, and you can do it within MCP or Copilot. I just wanted to level set those two things. If you have any questions, please reach out to anyone at Datasite and we can talk with you about that. With that, let's get to the demonstration portion of it and talk a little bit about some practical examples.

Again, these are going to be more sell-side admin workflows. Caitlin is going to pull up her screen here, and we're just going to talk about this. Okay, let's scenario-plan here a little bit. I now want to set up a project. Can I actually do that from a surface like Copilot?

Caitlin Murdy (11:38)

You can. It's pretty incredible. A lot of the things that you used to have to log into the data room and do through our UI, we're now offering via the connector. Without further ado, I'll pull it up so you all can see it, and I'll call out as we go through.

Doug mentioned using the Blueflame AI Assistant to enable AI-powered search, understanding, and interaction with the content. I'll call that out in this first example, where you're just creating a new project that can be done via the MCP connector. I said, create a project title. It's Snowflower, Sunflower, I don't know what a snowflower is. What you can do through this is you'll see Copilot come back, and there is a level of detail that we need in order to create the project. Copilot knows that. It asks, it confirms, I give a quick free-text response with what is needed, and then my project is created as easily as that.

If I want to go in, it provides me with the link so I can quickly access it. It actually prompts me to say, do you want to create a folder structure? How do you want to build it out? Based on the data I've already provided, it then goes ahead and builds out my folder structure. I can obviously edit this later and alter it to be more specific to my process, but it does a first pass at that and builds out the in-depth folder structure that I'll need to leverage. Then I'm thinking about it: Doug is actually supporting me on this deal, so I want to make sure he's added to the project as well.

All I need to do is say, "Doug Cullen," and he's now added as a project admin. I specified his role. If I hadn't, it would prompt me with the different roles that are available and then ask me which one I want to provide to him. It's pretty cool and powerful what it can do when you think about having to log into the site and do those things through the tool. Now, if you're already working in Copilot, you can just continue working there and do exactly what you need.

Doug Cullen (13:47)

Yeah, an amazing thing for us. We have literally hundreds of thousands of users who access Datasite every single month across 28,000 companies.

A bunch of people are super sophisticated and experienced. They've done this a bunch of different times. Others may be doing it for the first time. One of the things that you're able to benefit from and draw some power from, if you will, is our collective experience having supported literally hundreds of thousands of transactions across the world. Here, you're basically able to create this folder structure. You can get these as templates within our core platform, but it's really going to create this dynamic folder structure. So you're already a step ahead. Maybe you had to ask a former colleague, "Hey, do you have a previous index?" Maybe you call our Project Pro Team.

They can do this. But right now you're able to do so much just out of this surface of Copilot that it's already, I think, game-changing.

Caitlin Murdy (14:43)

Definitely. And I think what's really cool too is the flexibility. If you know exactly the deal you're working on, you can describe the asset that you're selling and it'll get even more in-depth and particular to your deal. Or if you want to keep it vague, you can do that, but it gives you the flexibility to run it how you need.

Doug Cullen (15:01)

Yeah. Then you know what typically ends up happening before this. I get my folder structure, I start to organize, and I add the appropriate documents. If I'm the corporate, I'm doing that directly. Maybe I'm working with this as an investment banker, a lawyer, or, depending on where you are, maybe an advisor of some other sort. Then I think a really powerful thing is that usually you want to get a sense of how this is looking. Is the content logical? Is it connected to the right things?

Are there potential gaps? Is AI there for us to be able to get that type of insight?

Caitlin Murdy (15:37)

It is. It's pretty incredible. I think what you're describing, that kind of launch-ready checklist, can be one of the most tedious, time-consuming, nerve-wracking parts of the process. You have to make sure everything is ready to go, so the right buyers are getting access to the right things. Now you can leverage AI as that first pass.

In this case, it's leveraging Blueflame AI to analyze the content, understand what's in there, and determine where you have gaps. But it's also leveraging the regular MCP connector to look at things like whether things are published and who has access to what. In this case, I've asked it to run my prelaunch readiness scan. I love the visual traffic-light status it provides, so I can really easily see that my first folder looks ready to go. My other folder is not fully published, and I might want to look into that, or maybe it's intentional.

Then we dive deeper into Round 2, and I can see that, based on the AI scan, a lot of these folders have the right content in there. There aren't a lot of gaps, but where there are gaps or where things are missing, that's flagged for me to dig deeper into. Then it even goes ahead and gives the top five issues or fixes that I should make immediately. What I love about it, and with a lot of these prompts you'll see this, is that it really drives me to take action and points me toward what a viable next action is. I need to figure out my gaps in what's published and unpublished, or I need to rename specific files as it went through certain things in here.

They were not named with the broader naming structure. Or maybe there are some draft documents or things that look unfinished that I need to look into. The final thing I asked it to do as part of the readiness scan is start to anticipate what buyers might ask me. I want to get in front of what their questions are going to be. I want to understand if there are going to be any gotcha moments before I go live. It drafts all of that within here, linking obviously back to the data room, which I think is crucial too, and then provides a quick launch checklist.

Like I said, it's going to drive you to take action, which I think is super important in this situation.

Doug Cullen (17:54)

Yeah. So let's maybe take those a little bit one at a time and unpack them. First of all, we've got the Blueflame-enabled content. We're able to have the folder structure that we've established. We're then able to look into the content of the documents themselves and do a little bit of a compare and contrast against the content of the document.

Is that being represented by the name? I've supported quite a few deals in my life, but I can tell you that adding something like Scan 4000 is not very helpful, nor is Book 009.xls. If I'm on the buy side doing a review, it's very difficult to understand that. So I think this is a crucial element for those in that preparation phase. As a sell-side team, you want to make sure your house is tidy and in order. You've got a logical index, you've got the right documents in the right locations, and those documents are named properly in order to facilitate smooth diligence, right?

I think this ultimately is reflective of a high-quality asset. Most people, when they're looking at a platform like Datasite, are really dealing with the best of the best, and we're trying to layer in the best technology to aid you on the way. You mentioned that this is probably one of the most stressful times that a dealmaker has, that moment before launch. It is a proverbial measure-twice-and-cut-once moment. I've heard massive stories around managing directors literally reviewing every file in every folder before launch because they were ultimately representing this on behalf of the customer.

I've heard of general counsels at corporates literally going through every file. I've seen controllers going through every single financial disclosure in order to tick and tie this. I'm not saying that you don't want to look at that, but this is where AI can be super powerful in terms of helping us connect those dots.

Caitlin Murdy (19:59)

Exactly. That's what it is. It's almost an extra layer of eyes on what you're doing so you can feel more confident in what you're doing and what you're executing. To your point on file renaming, as we went through that checklist, we saw that some of the files were not to the consistent naming standard.

As you said, on deals, when you're looking through and trying to find content and it's named incorrectly, that's going to slow down your deal. On one hand, AI is helping the admin craft and do all of this quicker, but on the other hand, it's helping the buy-side because you're finding a better-formatted data room, so you can go through, find the content you need quicker, and keep the deal progressing.

Doug Cullen (20:45)

Yeah, no. I was going to say that the next thing that we just covered, because I think this is a super important part, and maybe just bring up the screen again if you don't mind, is this notion of preparing for potential Q&A. One of the most critical things within a deal is really trying to think about, if I'm selling an asset, what are the likely questions that I'm going to get? If I'm an advisor, I'm trying to anticipate that. If I'm an underlying corporate, I'm trying to deal with that. I'm working with advisors on doing a deal in order to understand that, not only to make sure that we're prepared for any prospective questions.

Do I have the right documents? Do I have the right command of the overall content? Some of these projects involve massive amounts of content. If you showcase that again, can you walk a little bit through anticipating potential questions and how you're leveraging Copilot in this instance to maybe give you some preliminary responses to that?

Caitlin Murdy (21:51)

Yeah. What we're seeing here is that it's going through my data room. Again, it's reading through the content to understand what's there and then start to think about what buyers might ask, what gaps we might have, or what gotcha moments might come up.

If there's something that maybe you didn't realize you were sharing or how you were sharing it, it even flags where it thinks you're going to get a lot of questions. I was talking to a group yesterday of bankers who mentioned that a lot of times what they're doing now is they want to make sure their data room is set up and ready to go in alignment with the story they're trying to tell from the CIM. To be able to use a tool like Copilot to say, this is the story we're selling, by uploading the CIM and what they've crafted there, does the data room align to that?

That's exactly what this is doing here. It's helping you get a sense of the story you're telling with your data room by the questions people might ask.

Doug Cullen (22:50)

Yeah. It is a crucial component of how deals get done. On the advisory side, you're doing a lot of this hard work upfront, right, to try to anticipate, be prepared, and optimize the time of the people who are involved in the deal. Across the deal team, of course, you've got the deal team extension, which we like to think of as your Datasite team, Project Pros, and of course our platform and our experience.

But this is really showcasing some new capabilities that platforms like Copilot, really leveraging the capabilities that Blueflame AI brings to the table, are taking up a notch.

Caitlin Murdy (23:34)

One hundred percent. And I think to that point, another use case that we want to show is file renaming. We flagged it as we went through our readiness checklist. You might identify that you have inconsistent naming standards, or you might be working where you're receiving all this content, like you said, scanned documents where the title is Scan #4 XYZ. It's really tough for teams to understand what that is.

In working with admins of data rooms, I often get asked, how can we use AI to rename files? This is one of my favorite use cases to show. Again, this is an example of where it's leveraging both the Blueflame AI Assistant and MCP. Sure, with just MCP, you could push out new file names. But by having Blueflame AI enable the content to provide the LLM with the understanding of what's in the document, the titles that it provides are going to be a lot more accurate and realistic to what the content is.

So if you're ever a buyer in the data room, you know what you're looking at when you open a file. In this case, I'm saying, review the standard file naming convention and then, in this folder, update the file names to make sure they incorporate that. What I love about it is that, obviously, naming the files in a data room is a pretty risky task. You don't want it to go wrong. Copilot is going to double-confirm with me. It's going to first lay out what naming convention it sees across the data room, where it sees the gaps, and then put together what its plan is to proceed.

Once I go in and confirm that, then it executes, and now my file names are all consistent throughout.

Doug Cullen (25:27)

I mean, it's kind of magic. If you think about how this needed to be done even in our world-class platform in Diligence, there are a couple of different ways that we had this. Before this, you could do it in an Excel spreadsheet, you could load that, or you could tap into our Project Pros. They spend plenty of time renaming files on behalf of customers. I mean, literally hundreds of hours of work we're happy to do on your behalf.

But again, these are human-centered processes. Now we can have that consistency at the content layer, at the folder level, and at the file-name level. Again, that tidiness goes a long way around the quality of the asset and ultimately the quality of the diligence. What we're looking for is effective and efficient diligence when you're selling an asset.

Caitlin Murdy (26:15)

And efficient is such a key word there. It's efficient in how the admins are doing it because now they can rename the files, or whatever the task is, way quicker than before.

But it also adds efficiency to the buy side because they can find what they need quicker.

Doug Cullen (26:33)

Yeah. And just to clarify as well, because these have come up a couple of different ways, these are capabilities right now that are enabled on the project via the Blueflame AI Assistant, and what we're showcasing right now are tools and capabilities that are available to admins only. These are people who have full rights to rename, move content, etc. Our plan is to extend some of these capabilities to reviewers, or the classic sell-side capabilities. Administrators are sell side; reviewers would be buy side. That's something that we are currently working on.

You can stay tuned for some of those capabilities. We have a few minutes, so maybe we'll go into some of the Q&A. Okay. This question is, does this mean that sellers' preparations can largely take place in the data room rather than beforehand in their own file storage system?

Caitlin Murdy (27:40)

Yeah. They can definitely do things in the data room, or if the content is moving over to the data room, they can manage it there. They can upload directly via Copilot.

It'll bring them a link to their data room to upload there, so they have some flexibility in how they work with these things.

Doug Cullen (28:00)

Yeah. And just to add some color to this, it is not uncommon for people working with that preliminary set of content to have it on a local shared drive, a local hard drive, or a set of material that they're working with in order to stage it prior to loading it onto Datasite. Our recommendation would be to get it up there immediately, take advantage of these tools, and take advantage of the Blueflame-enabled content.

I know if I have an environment just on my laptop, my capabilities on Datasite are vastly superior to what I'm able to do just in a standard operating system.

Caitlin Murdy (28:42)

Definitely. That's what we've seen as we first went to market with the Blueflame AI Assistant. It helps understand, especially for admins, the content that's going into the data room. So if you're a sell-side advisor and you're not as familiar with the company but you need to start to move all of the content around, you can easily understand that based on the Blueflame AI Assistant.

Doug Cullen (29:05)

Yeah, and just a quick question here as well. If files are renamed, will links shared previously to docs still work? The answer, I believe, is yes. The document itself doesn't really change your authorization or permission to that document; it's just going to be refreshed from a naming perspective in the UI itself. Does the file naming system limit the characters to avoid downloading errors? I think there is a character limit in the system. I'm not sure what it is these days.

Caitlin Murdy (29:38)

We have built in the ability, when you download, to concatenate the file names if needed to avoid errors. So we do account for that.

Doug Cullen (29:47)

Okay. We've got one more question. Thank you so much for the participation. I think we talked a little bit about it, but it's probably worth reinforcing here. Which advantages do you feel Copilot can bring to Datasite for workstreams in corporations during DD who are not involved in data structure setup as investment bankers? You want to take a shot?

Caitlin Murdy (30:14)

Yes, because I think it's actually the perfect question. We were running out of time, so we didn't show our last prompt.

One great example of this is responding to buyer questions. If you are on a corporate workstream, you're managing as part of the sell side and you're getting a lot of questions from buyers that you need to respond to based on the content in the data room, you can easily upload that tracker into Copilot or your preferred tool, and it's going to read through the content within the data room and be able to pull out responses. What I love about it is that I've prompted it as well to give a confidence score and to link back to the data room.

So it's really directing me to take a first pass to get a draft of my answer, but then to validate within the data room.

Doug Cullen (31:05)

Yeah. And I think it's going to really transform the way deals get done. I've been doing this for a couple of decades at this point, but I think having the ability for people with all sets of experiences to log into the data room and ask some basic questions, and prepare via natural language, opens this up for corporate users, people who may or may not be as familiar with it, as well as a bunch of senior people who typically may not log into it.

We heard a story from a senior banker who laughed and said, I may log into a data room for the first time in like a decade, because that's been effectively almost pushed down to some of the junior levels. While it is really paramount for advisors to be in tune with all the different shifting dynamics of the content within the deal, what questions people may ask, and how you ask them, we're going to have to leave that here. I want to thank you, Caitlin, for such an amazing job here today. Always a pleasure working with you.

Caitlin Murdy (32:06)

You too.

Doug Cullen (32:08)

Yeah. It's a good time here, and we're going to have to wrap things up. Look, this is happening. We're at the forefront of capabilities with MCP and showcasing what Copilot will be able to do for our users around the world. We're thinking about this through the lens of governance, compliance, and ensuring that the confidential information you have does not leave and is not disclosed unintentionally outside of the secure barrier of Datasite that you've come to trust and love so much.

With that, please refer this to other people. We do have our next session in the series, where we're going through a little bit more in detail with Blueflame. I want to thank people from around the world for tuning in. I want to thank you, Caitlin, for doing such a magnificent job of showcasing these capabilities. The webinars are recorded. Please forward them around. Just an extended thank you to everyone who spent time with us here today, and I hope to see you on the next webinar next week.

Doug Cullen (00:00)

Good morning, good afternoon, and good evening. Thanks to everyone for joining us for the fourth and final webinar in our series. Welcome to the Datasite MCP for Blueflame webinar. I'm joined here by Raj Bakhru and Alice Esmerian. We're going to walk you through several different components of Blueflame, including the Blueflame AI Assistant within Diligence, and we're excited to have you here. My name is Doug Cullen. I'm Chief Strategy Officer here at Datasite, as well as head of corporate development and head of partnerships.

Today, we're going to talk with you a little bit about how Datasite and Blueflame AI work together to bring governed AI workflows directly into deal execution. Our web series has focused on the transformations that are available now with AI technology and how they are really changing the way deals get done. As we focused on in previous webinars, we're going to dive right into the Blueflame experience and showcase the Blueflame AI Assistant in the Diligence application itself. Before we get to those exciting things, I just have a few housekeeping items that I want to cover.

Okay. First of all, we want to hear from you. You'll see a Q&A panel, so please ask questions throughout the webinar, and we will do our best to get to those questions after the fact. If we don't get to them, we'll still follow up with some FAQs to make sure that we address the most important topics. We are also going to put together a survey, so we want to make sure that you respond to that. You'll find additional resources in the console.

These include some of the FAQs from previous sessions, as well as information about some of the things that you're going to see here on the webinar today. Of course, we are recording this, and it will be available on demand. And what webinar would be complete without a good legal disclaimer? All the opinions expressed are our own and not those of Datasite or Blueflame. Welcome. I just wanted to do some quick introductions. Again, we've got Alice Esmerian, product strategist at Blueflame. Why don't you tell us a little bit about how long you've been at Blueflame and what you did previously?

Alice Esmerian (02:25)

Sure. Thanks for having us here today. I'm on the product strategy team at Blueflame. I joined six months ago, so it has definitely been an exciting time with no rest. My background is in investment banking and private equity, where I spent almost eight years. It's very exciting to help bankers and private equity professionals benefit from new AI products in their workflows.

Doug Cullen (02:50)

Awesome. And Raj Bakhru, co-founder of Blueflame, can you give us a little bit of background about Blueflame and maybe some of the things you did before founding the company?

Raj Bakhru (02:59)

Yeah, sure. Thanks. Hi, everyone. Raj Bakhru, co-founder and general manager. My background is actually on the tech side of things. I started my career at Goldman, was later at Highbridge Capital and Kepos Capital, and started a security firm that we sold to a compliance firm. Eventually, I ended up running M&A and corporate strategy as the chief strategy officer of that company. That was a PE-backed company under three different private equity owners. I saw a lot of the pain of the manual workflows that we had in acquiring dozens of firms.

We launched Blueflame because of that. We saw that there was a huge opportunity for AI to step in and make life a lot better for folks in the industry. We knew AI was going to be transformative, and here we are today.

Doug Cullen (03:41)

Awesome. Well, we definitely want to get into showcasing some of the capabilities. We have a lot to unpack on what's going on across dealmaking. AI is becoming central to how deal teams think and operate. Teams increasingly want these AI workflows, but some of these things are not super intuitive out of the box.

So I think we want to showcase that. And then, of course, as we think about it from a dealmaking perspective and an overall strategy perspective, governance, security, and auditability are really critical to everything that we do to ensure that the trusted environment of Datasite is extended into all these different workflows. Raj, from your perspective, what do you think is getting people really excited about some of the things that are impacting deal execution in AI?

Raj Bakhru (04:32)

There are so many things that you can do with the tool in executing either the build of your data rooms, the launch of a process, all of the workflows that happen through the process, and answering questions that are coming in.

And then on the flip side, if you're a buyer dealing with the data room and working within it, asking the tough questions, and running through your analysis and your workflows, whether that is customer retention data or HR analysis, all of that data obviously sits in the data room. It's unstructured data, and the AI has gotten really good at being able to parse through that data, address questions, and find things for you that you might not be able to find on your own, or that would have taken you hours and hours to find. We're going to demo some of that today. Alice is going to show us some of those workflows on both the buy side and the sell side.

Doug Cullen (05:16)

Yeah. We got to know each other about a year ago, met the team at DealMAX, and struck up a great conversation. Ultimately, we bought Blueflame last year. I think we closed in July or something. It's pretty amazing to see some of the things that we've brought to bear already from our perspective. I met the team, saw some of the things you were doing, and I was blown away.

We, of course, had our own organic set of capabilities that we've been bringing into the Datasite platform for six or seven years across the AI landscape. But I was really wowed by some of the buy-side workflows that Blueflame was bringing to bear and had this vision of, "Hey, imagine if we could take some of those capabilities and bring them into a native Datasite experience." This is really what we believe dealmakers wanted: natural-language querying across large data sets. From our standpoint, it just felt super comfortable and very natural. You were co-founder, right?

You weren't necessarily expecting to get your door knocked on by Doug saying, "Hey, man, I really want to buy your company." How did you approach it from your perspective?

Raj Bakhru (06:29)

We knew, obviously, Datasite was the biggest, best, and most credible data room out there. We've used it in prior lives, and we knew that we wanted to be embedded with that content. It's where all of the really interesting stuff in a deal lives. We knew our clients wanted that. They wanted to be able to access that data.

We knew the only way to do that really well was if we were tightly coupled together, able to replicate the permissions and the security that live within Datasite all the way down to the details: what gets watermarked and what is allowed to be downloaded. There is a lot of intricacy in the security and permissions of a data room. There was really no way to bridge that gap unless we were tightly coupled together. We knew that the value would be tremendous, so we figured out a way to make it happen, and I'm glad we did. It's been a fun journey building together.

Doug Cullen (07:15)

Yeah, it has been absolutely awesome. So I guess the fun question that we get to showcase today is: what can AI actually do within this secure data room environment? Before we go into that, I think it's important to set some context about what's going on within this because we've got a few different layers. You've got the Blueflame AI Assistant, which we'll get into, and then you've got MCP. Can you give us a little bit of a 101, Raj, in terms of what's working here?

How are these things interoperating? How are some of these capabilities technologically possible?

Raj Bakhru (08:00)

Yeah. A big piece of what we've enabled with the data rooms is broad-scale, deep, high-fidelity search. When content comes into the data room, we really want to understand what a chart says. We need to understand how the legend corresponds to the bars in that chart. We need to understand logos on pages, charts, graphs, and org charts. All of this is complex, but it is really important in understanding the context of the deal and all of the highly visual materials that exist in a data room.

So a lot of what we brought into Datasite was high-fidelity parsing and search algorithms to work through that massive set of data. That means we can now leverage that search either directly in the data room or directly in the Blueflame platform against that data room. We're going to show you both experiences today. Within the Datasite data room, there is now a sidebar, sometimes called Sidecar, where you have the ability to chat with the data room, and that chat is leveraging Blueflame behind the scenes to ask questions against the data room content. It's leveraging all of that high-fidelity search that we built against the data room.

So this is not like some of the agentic tools you see out there, where the tool tries to search file names and read just a few files to answer a question. This is actually looking through every bit of information in the data room to find nuggets of information that might be buried in files you would not otherwise find. You have the ability to do that directly in the data room with the sidebar chat. You also have the ability to go into the Blueflame platform and do everything from creating a data room to managing the data room to asking questions against the data room.

There are a number of different ways you can leverage that on both the buy side and the sell side, which we'll demo today.

Doug Cullen (09:44)

And because we've been processing content for a long time at Datasite, one of the things I was really impressed by was this next level of processing and extraction capabilities that Blueflame helped bring into our environment. How is Blueflame able to go into things like charts and graphs, pictures, and org charts and really make that information come to life?

Raj Bakhru (10:11)

Yeah. We're leveraging vision models against a lot of that content to pull out all of the details that live within it.

Over the past two years, AI models have gotten much better at being able to read that content. But there are still cases where we have to optimize. Certain models do better against bar charts, and certain ones do better against pie charts, and so on. So it is very intricate data science work that happens behind the scenes to enable all of this. The result for someone interacting with the platform is that when you ask your questions against the platform, it shows you all of the places where it is finding relevant search results, and then it goes ahead and answers the question based on all of those results.

What we've seen happen over the past six to 12 months is that the quality of those answers has become extremely rich, extremely high confidence, and frankly really good at finding the right answers by leveraging a combination of agentic search and high-quality semantic search.

Doug Cullen (11:10)

Awesome. Thanks so much for that. I think now let's just get into it. I love this showcase in a Diligence project. I think Alice is going to share her screen here.

Why don't we just dive right in? If you don't mind, walk us through what you can expect now as a Diligence user when you have the Blueflame AI Assistant available in the Diligence project.

Alice Esmerian (11:38)

Yeah. We're going to cover a few workflows here. There are obviously many more that can be done. Feel free to drop any workflow suggestions or workflow questions you have into the question panel. But just to get started here on my screen, everyone will recognize the Datasite environment. In case you haven't enabled Blueflame AI within your data room yet, you can simply do that by clicking this button at the top right here.

When you do that, you'll be able to open more information and, if needed, get in touch with your Datasite rep so they can help you add Datasite AI to your VDR. If you created a new project, you'll automatically have the option to have that. Now I'm moving on to what it looks like once it's enabled. I want to start by showing you how you can leverage it to make sure that when you start your VDR, all your content is organized and cleaned the right way. Here, I simply ask the Blueflame AI agent, "Can you help me run through my index, the folder names, and the document names?"

I want it to be very clean so that when I let buyers into my data room, everything seems very neat and they have an amazing first impression. Here, the model has the ability to look at the existing names, but also the real content behind every document, to inform how it should actually be labeled. That really helps me get useful feedback, such as, "Here, you're not necessarily using the right date because the date of the document is actually different," or, "You're not using a standard naming convention for all your contracts," or, "The date looks different." So it comes back with many different levels of feedback for me that are very helpful in quickly pointing out the issues.

As a banker, that is typically a process that you would do manually, scrubbing everything, and that's time-consuming. Now you really have this level of feedback from a single prompt, which is pretty amazing. Another workflow that I really like is the risk audit capability that we've built with the help of Blueflame and Datasite working jointly. Here, I'm asking the model to look at potential risks and create a risk report. As a banker, you are typically heads down in your deal, and there are some oversights that may happen that buyers will be very focused on, and you may not necessarily have prepared a sufficient level of defense.

So here, you really get this third pair of eyes, an extra person taking a look at the entire content of your documents and screening every single document that you might have uploaded without actually reading every single line. It will give you a risk snapshot for both strategic and financial buyers if you want it to, and it will look at very different aspects of your deal. It will scrutinize everything it can, including add-backs, change-of-control clauses that have different types of risk around them, and customer concentration across different verticals.

So it is really going to rerun a lot of the analysis and data available so that you can come prepared, get your defense where it is needed, and then open your VDR. Although it might look like a lot of content, you can then get that into a simple table with risk, buyer type, and defense. I'll copy and paste that into an email. I think that's a lot of work that has been done very quickly thanks to Blueflame AI being able to go through all your content and retrieve information for you.

Doug Cullen (15:09)

And Raj, from your perspective, how is this able to do such a high-quality job looking at all of this content? I mean, we have a very large corpus of information potentially here, as well as across Datasite. I always feel that Blueflame has this innate ability to really understand things from a dealmaker perspective, and I don't think that's accidental. It feels like it has been by design.

Raj Bakhru (15:33)

Yeah, that's always been what we've been going after. The vision has always been: how can we build for the prompt-first dealmaker?

Someone who wants to come into the platform naturally and go there first to get their work done. What you saw Alice do here could obviously save you hours of time, and you didn't need to custom-create a really difficult prompt to get to that. I chuckled briefly when Alice said thank you to the agent and then had her prompt there. It was very natural language. It is able to ascertain what the user is going for, what the investment professional or the dealmaker is going for, because we designed it to do exactly that.

What we've seen happen over the past six to 12 months is that the agents themselves, the agent harnesses, have become tenacious, which is what we like to call it, in the way that they search for information. In that last example Alice showed you, there was a lot of information that it had to dig through to generate that table of potential issues. That's exactly what you want it to do. I don't know how long that prompt took. It might have been a few minutes of work to go do all of that, but that's exactly what you want. You want it to spend a lot of time working through all the information.

The agent harnesses have become really good. I mean, we've designed this agent harness to be really good at reading through the VDR, understanding all the content, and finding all those potential issues really well just from the natural-language query.

Doug Cullen (16:52)

And effectively, I'm asking almost a pretty basic question, but underneath that, it's firing off dozens of additional questions in order to orient its way through the data room, find the answer, and pull it out. I love that sort of resilience that it must have in the models themselves.

Raj Bakhru (17:11)

Yeah, from a technical perspective, it's leveraging subagents.

It can kick off dozens of subagents in parallel to read through content, ask multiple questions at a time, and aggregate tables of information. Again, it is tenacious and resilient. It will always work really hard to find the right answer for you.

Alice Esmerian (17:29)

I think what's great on top of that is that this is all done within your Datasite instance, so you don't need to drop any document outside of Datasite. Everything remains safely stored within Datasite, and that additional level of security is very important.

Doug Cullen (17:44)

Yeah. I mean, we really have taken a lot of time thinking about the guardrails. We are the trusted place where deals are made, right? So we wanted the ability to take advantage of these capabilities, but in a super guardrailed environment. Can you talk a little bit more about that, Raj? How difficult is that? What would be some other approaches? If you wanted to take things out of the data room, obviously that would create a bunch of risk and exposure points.

Doing this within the four walls of Datasite is a pretty important thing for our dealmakers around the world.

Raj Bakhru (18:20)

Yeah, and it has a lot of value, right? The download-upload process itself is clunky. It's also error-prone, right? Sometimes you just have old copies of information. You haven't brought in the latest copy. You start asking questions, you realize that you have bad data, and you ask poor questions because you're working off the wrong data.

This is all directly built in. It's native. It's fast. You can just go directly and ask questions against it. The permissions are obviously really important, so that was baked in from the get-go. No one is getting access to anything that they don't otherwise have access to. The agent is only working through what you have the ability to see, and if you have the ability to download, any watermarks are preserved. We even provide what I think of as guardrail citation: the ability to understand where a specific figure or data point is coming from, and to go back to that source document to verify as needed or to call it out in any of your follow-ups.

So we think the guardrails were paramount. That's table stakes. We had to make sure that it got that perfect. And then, obviously, on top of that, we wanted to build a lot of the value-added functionality.

Doug Cullen (19:25)

That makes a ton of sense, and it's just so powerful to have these capabilities in the platform itself. This is kind of what I always felt search should be, right? The technology just was never there to allow for this type of interrogation, natural-language querying, and ultimately getting these incredible answers right within the Datasite data room.

Raj Bakhru (19:49)

Yeah, and it's not easy. I will say that even with a download and upload into any of the LLM tools that are out there, you're not going to get the same level of tenacity with the agent or the same level of high-fidelity search through the charts, graphs, and logos. So replicating a lot of what Alice just did through the core LLM tools that are out there is not necessarily doable. You'll hit token limits, and you'll hit a number of other issues in terms of the fidelity of content that's coming in and the ability to search through it.

Doug Cullen (20:19)

Yeah, because if it's just the LLM, it is limited to the context window that it has. With the Blueflame agent experience, you're not really limited or bound by those traditional constraints. You're really able to look at 5,000 to 500,000 pages' worth of information in a data room.

Raj Bakhru (20:40)

Exactly. If you were to leverage some of the project tools that are out there, there are limits on the number of files you can have in there and the size of those files. It doesn't directly index all of those files. So we can obviously work against much larger or more complex data rooms much better.

Doug Cullen (20:55)

Awesome. So I think we're going to shift gears a little bit and maybe showcase the Blueflame platform itself. Part of the connectivity here is that you've seen the Blueflame agent sitting within Datasite, the AI Assistant, and now we have this other incredible experience in the Blueflame platform itself, which is leveraging the MCP capability. It's awesome to see what other customers get if they've invested in Blueflame and have it as an enterprise platform, which we recommend for everyone, of course. You get an even more enriched experience.

Alice Esmerian (21:32)

Yeah, absolutely.

We've always had this connectivity into Datasite, where our clients were able to query Datasite content from their Blueflame instance. Now, the addition that we're releasing with this MCP connection into Blueflame is the opportunity to push content into Datasite. In the example we're seeing on the screen, I'm starting fresh and just want to start a new project without ever leaving Blueflame. So I'm going to ask Blueflame, "Can you help me set up a new Datasite project?" It knows that it's going to need a bit of context to help me as best as it can.

It immediately asks me: what is the industry of your deal? What type of deal is it? What is the size? Where is it going to be hosted? That is obviously very important for our customers. By providing this high-level context, it automatically returns a sell-side data room index that is tailored to my particular deal because, as you can see here, it came back with some IT-specific sections and things that you'd expect to see in an IT services data room. So that is great, and I can push it to Datasite. But I want to take advantage of the fact that all my Datasite data rooms are actually flowing into Blueflame.

So I push it a little further and say, "Actually, can you look at this other deal where I really liked our index and inform from that what should be the final index for that new Pegasus deal?" The model does that immediately. It looks at everything in that other VDR, comes back with a plan, and then, once we iterate together, finalizes a very detailed index that has all my main sections and folders. Everything is prepared from a single push of a button to be accessible. Then it also provides the link so that I can open my Pegasus sell-side data room within Datasite to see that it actually created all of that, which is pretty insane.

As a banker, it would take a lot of time to move a clunky Excel spreadsheet into Datasite to have your index populated. You obviously need to think about how you want to orchestrate that. Here, you can leverage all your firm context from all your data rooms and build something very tailored, all from a single natural-language prompt.

Raj Bakhru (23:42)

Yeah, it's amazing how fast you can do things now.

Doug Cullen (23:45)

I mean, it's not just that, in and of itself, getting access to a data room via Blueflame or MCP and setting it up is super powerful.

Then you take it to a whole other level by doing a comparison against another deal. You'll be able to dynamically operate in this Blueflame environment, taking advantage of your previous work and best practices, and ultimately easing that setup for the project itself. You can do it in minutes, as opposed to the days it probably would have taken back in the day.

Alice Esmerian (24:19)

Yeah, definitely. Another thing that is quite time-consuming when it comes to setting up a data room is inviting parties. You're receiving names and email addresses in different spreadsheets, emails, and unstructured formats, and you need to manually push that into Datasite.

Now, the way it works with the MCP is that you can drop a spreadsheet, and the model is going to identify the roles, names, and email addresses of the individuals. Here, you can see on the screen that I pushed an email list, it identified the right people, created the roles, and then had the invitations ready to go to their inboxes. Obviously, it is done for me to go into Datasite and make sure that they have all the access permissions and that I can keep control of that. We're not pushing that all the way to the other end quite yet because we feel that it's super important to make sure that the permissions are still set up manually at this point. But at least you can invite them and locate each individual role from a single prompt again.

So that's very helpful. I'm also conscious that, having been in the bankers' shoes, you don't always receive a nice spreadsheet with everything in an organized way. You receive several emails in the same day saying, "Hey, can you add Raj? Can you add Doug?" What I love with Blueflame is that because it's connected to your email system, you can actually pull the data from your emails without having to sit here, and then push it into Datasite without ever leaving Blueflame. So that's another way to leverage the fact that all your systems are working jointly within Blueflame AI.

Doug Cullen (25:53)

And your CRM, right? One of the big value points for Blueflame is connectivity to all the primary systems of record. You could be a banker who has a CRM connected, and you're connected by Office 365. How is that working, Raj, to enable Blueflame itself to have access to all these critical business systems?

Raj Bakhru (26:12)

Because we're connected to all those systems, and we have proprietary integrations in all the major CRMs, in Office 365, and obviously Datasite and Grata, you have the ability to have the agent correlate data across them.

So again, you can pull information from one and push it to another. You can do some pretty cool things there, such as asking, "Who do I have in my CRM as potential buyers for this project who have not been set up in Datasite yet?" And you can reconcile a list across the two as well. Obviously, Office 365 comes with email and calendar access, and you can automate quite a number of things by leveraging the data room with the Office 365 integration. You can create your process letters and auto-send your process letters. You can point people to specific files in the data room with links. There are so many things you can do by pairing these systems together with the agent.

Doug Cullen (27:03)

It's how I think people have always wanted it to work. You just had to go to these different places to do the different things, and now you're able to almost sit in this beautiful environment and allow the orchestration to be at your fingertips.

Raj Bakhru (27:17)

Yeah. There was another webcast we did where Alice created a whole CIM through prompting, right? A 55-page CIM created through prompting, and now you can upload that into your data room directly through the MCP connection.

Alice Esmerian (27:29)

Yeah. And I feel that another way to really benefit from this integration is around the process that comes when you prepare your data room. As a banker, you work with your client to make sure that they're populating the right information and that you're going to have everything covered. Here on the screen, we're looking at a data request list, or DRL, where you're going to have all the requests you submitted to your clients. You never know if the process is really tracking what has been uploaded and what information is there or not there.

It changes for every single process, so it's quite manual. Here, I just uploaded it. The agent is going to read through all 241 items that I requested from my client. It's going to crunch the entire data room to see what's in there or not and come back with very detailed feedback. It's going to produce the materials I requested for me to run through these conclusions, which include the detailed DRL with annotations, sources, and page numbers for each of the different information requests, as well as a summary in PowerPoint that I can share with my client and tell them, "Here is what I need from you."

Those sets of documents are produced by Blueflame. You can see here the DRL tracker, as well as the PowerPoint. Sorry, it's just loading for a little second. The DRL tracker is in XLS, as well as the PowerPoint. On top of that, what I'm asking Blueflame is that it put it into an email that I can send to my clients so that I don't even need to move back, download the files, put them into Outlook or whatever app you use, draft the email, and send it out. It's already sitting in my draft folder, and I can just push it to my client so they can easily navigate where they need to prepare and submit more.

So that's really a great way for me to leverage AI and save time while keeping all the traceability and auditability of everything you're doing within Blueflame.

Doug Cullen (29:26)

Yeah, incredible value for the advisors and the clients, right? Ultimately, having the banker or other trusted advisor run a process for you, part of the difficulty, as you know from running strategy, is trying to navigate who needs to get access to it and what files need to be provided. Of course, there are always going to be gaps, but being able to really focus on the missing pieces as opposed to all of the pieces is crucial for customers and really allows you to have that optimal conversation, kind of trusted advisor to the underlying corporation.

Alice Esmerian (29:58)

Yeah, definitely. And we've covered a lot of sell-side workflows today, but this is obviously something that is going to benefit our buy-side clients. We just released within Blueflame what we're calling Spaces, which is a way to, on top of getting access to your data room within Blueflame, create a space around it where you can put your SharePoint data and really any content that is flowing around a particular deal or around a particular industry that you're focusing on, where you can work in this closed vault. You can invite team members you want to collaborate with and leverage both the content of the data room and all the analysis that your team is working on.

Here you can see my project's homepage. I love the fact that I have an activity thread and I can track who has been doing what on that project. Here, I've been working quite solo, but team members will see everything everyone is doing if they decide to share it. You can add people here, and you can see that I have everything concentrated, with my chats around that particular project and notes that I take. Those can be notes from a call one team member is having with a CEO, while another team member is talking to an expert and another one to the CDD provider.

So you can just share all of that and then query your data room content with this additional layer of context to really optimize your work around the deal.

Doug Cullen (31:26)

Yeah. And just to take a step back, because I think this is so profound and we've gotten a ton of questions around this, what we are covering on the Datasite side of things with the Blueflame AI Assistant was really highly oriented toward the sell side and many of the things Alice was showing there, again, for sell-side advisors. We're jumping to the other side of the coin here because now you're really thinking about most of the interaction that happens around the deal is not necessarily on the sell side.

It's really around the buy-side and all the interested parties. So as a Blueflame customer, yes, I get access to the Datasite data room, which is a super crucial part of it. But now I have my own space to operate. We've got this project, and I can add so many different contextual components. I can invite my team members, I can invite my advisors, I can invite experts from third parties, and so on into this space. I'm tracking it and really working through the diligence from my point of view, which I think is just going to transform how buy-side deals are done.

Alice Esmerian (32:31)

Yeah, absolutely. I think buy-side people love the data room space, but they also love having their things working in their own instance. They're creating a lot of analysis on top of the data room. They might want to leverage previous deal material as well to inform the decisions they're making and the risks they're identifying. Here, it's really the opportunity to work side by side. You open a note on one side, and you open a chat or a document on another, and everyone works with this shared context that is powered by Blueflame.

Raj Bakhru (33:00)

I think what's really nice for the buy side is that it's a place to collaborate, right? The data room has historically been a place to share information. This is a place where I can go to my colleagues and work with my colleagues. As Alice has shown here, you could take your call notes. Everyone on your team can see that same set of call notes, and that all feeds into the same context that's being used in the chats in the space.

So everyone is working off the same set of information. You no longer have these little silos of knowing what's going on or digging around SharePoint files or whatever it may be to find the internal context on a deal.

Doug Cullen (33:34)

Right, because you may be taking this to investment committee, or IC. I think that's one of the great use cases within Blueflame: taking all of this context and all of this information to that next level to present to investment committee.

Raj Bakhru (33:49)

Exactly. And getting to that ready draft of information, Blueflame already has all of the context of how our firm likes to build PowerPoint decks.

This is the branding guidance around what we do for Excel models and PowerPoint decks. So it just makes building that material so much faster and easier when it's already coming into the format and the style that you like.

Doug Cullen (34:13)

So we're going to go to the next portion of this, or we're going immediately into some Q&A. All right, I am going to read from the questions here. We probably don't have time for everything here, but I think the most important question is: Alice, how do I get the Blueflame AI Assistant on my Datasite project?

Alice Esmerian (34:37)

So for any Datasite project, you'll always have this button that I showed at the beginning of the webinar. It is the Datasite AI button on the top right of your screen. Once you click on it, you can get more information around exactly how it works and what it's able to do, and you can also activate it or get in touch with anyone at Datasite, who will be able to help you very quickly and enable it. That's for existing projects. When it comes to new projects, it's something that you should immediately discuss with the person at Datasite helping you set up a new VDR.

Raj Bakhru (35:10)

And I think what's worth noting is that you don't need to be a Blueflame subscriber to leverage the Blueflame AI Assistant within Datasite. That can be something you do on a per-VDR basis if you so desire. There are lots of reasons to become a Blueflame subscriber, as we showed, but you don't need to be one to leverage that.

Doug Cullen (35:25)

Yeah, awesome. And then I guess, maybe this is the paranoid side of me, Raj, but how do we ensure that redacted and/or permissioned content data isn't accessed improperly? That's part of the trust of Datasite, right?

Raj Bakhru (35:39)

You're using Datasite because Datasite puts a premium on making sure that that's always right. That's part of why we paired together. We wanted to build tightly together to make sure that we're never showing information to folks who are not permissioned for that information. We're never allowing people to download things that can't be downloaded. We're always redacting information the same way. Translation still applies the same way, and watermarking still applies the same way it always does. So we took all of it into account to make sure that we were putting out something that was safe and adhered to all of the expectations that everyone has of Datasite.

Doug Cullen (36:19)

This wasn't necessarily entirely new to Blueflame, right? We basically have to do the same thing from a Blueflame standpoint by respecting SharePoint permissions and content availability and visibility. This is almost core DNA for Blueflame.

Raj Bakhru (36:34)

A lot of our team comes from cybersecurity, privacy, and compliance backgrounds. That is a big piece of why we knew Blueflame, a big piece of what we knew we needed to deliver as part of Blueflame in this space, right? The sensitivity of the content is extreme, just as high as it gets.

We wanted to make sure that we were protecting that.

Doug Cullen (36:55)

And then the question here, which I think is super important, is around audit trail. How do I know what has been asked? How do I know what has been presented? We want to make sure that we maintain this crucial set of information around the deal so that there are no dark spots in the room, if you will.

Raj Bakhru (37:15)

Yep. Again, no different. We wanted to make sure that we were doing things consistently with how the data room has handled things historically. All of the queries are archived. We have the ability to surface that, and all of the file accesses still go through the same auditing, so the audit trails still apply.

Doug Cullen (37:32)

Yeah. Needless to say, kind of table stakes, right? We couldn't do this any other way. We wanted to make sure that we extended all of that monumentally important governance, compliance, and auditability, just like everything that happens on Datasite. That is paramount for deal success.

Raj Bakhru (37:51)

Exactly.

Doug Cullen (37:53)

Okay. Alice, talk to me a little bit about some of the advantages for sell-side workstreams that may or may not have a deal admin.

Alice Esmerian (38:01)

Well, it's true that today we've covered a lot of workflows that are focused on the person really setting up the data room, so the team that is going to be in place to do that. But there are a lot of sell-side individuals who are going to be added to a data room and do more focused work for VDD reports, or maybe other folks at the bank who are less involved in the day-to-day process but still want to enter the data room and be able to understand what's happening.

So I feel that this ability to question the content with Q&A, understand where documents are, and identify whether there is any information that you might have missed is important. Let's say you're a commercial diligence provider. You're going to look at the commercial contracts, but there may also be other bits of information in an IP folder that you actually might want to be aware of. So that's a very good way of orienting yourself and getting very quickly into a deep level of information for someone who is less involved in the building of the data room.

Doug Cullen (39:00)

Great context. We covered this, but I think it's super important, so I'll read this question aloud, and maybe Raj, you and I can probably tag-team this a bit: what is Datasite MCP only versus when do you need Blueflame AI Assistant, and is the Assistant included in the Datasite license?

Raj Bakhru (39:21)

Yeah. So the Assistant is that first experience that Alice showed today within the data room. There is a chat on the right side. That's the Datasite AI Assistant, the Blueflame AI Assistant within Datasite. That has its own separate commercial terms on the VDR. There is an extra piece to the SOW, or you can upgrade as Alice showed to get access to that. Blueflame clients have access to that by default. It's enabled for them to ask questions there. That said, they also have the ability to ask questions directly in Blueflame, and it's the exact same tooling behind the scenes. It's all going to the same place and running the same queries. Oftentimes, you want to be within Blueflame itself because we have wider integration access. You have the ability to go out to your Office 365 and CRM.

In fact, FactSet or CapIQ, or what have you. So you can do a lot more across the ecosystem directly in Blueflame, but you have the ability to leverage it in either, if you're a subscriber. You don't have to be a subscriber to get access to the Datasite sidebar chat that we are. That is the Blueflame AI Assistant within Datasite.

Doug Cullen (40:27)

Awesome. Yeah. So if you have, and I think Alice covered this as well, but just to reiterate, on net-new projects that you may be engaging Datasite with, it will be an option to add this as a set of capabilities.

And then if you actually have active projects, which we have tens of thousands of before this capability really was available, we can add it to those projects as well. So any active deal that you have, you can do so either directly within the app itself or, of course, you can always contact either a service professional or a person in the sales organization, and we can turn it on for you. Okay, we have one more question, and this I think is great as well. Maybe Raj can talk about it: what LLMs are being used to create this magic?

Raj Bakhru (41:18)

Yeah, great question. That changes potentially all the time. As everyone knows, models are constantly leapfrogging each other. We leverage all of the best-of-breed models and open-source models. We'll leverage whatever we need to leverage that can be secured the right way. Our data science team is constantly benchmarking models to understand who is doing the best at what actually matters. Some will do better at certain types of questions or certain types of content, and we will get specific about what we point to for particular circumstances. As we go out to certain integrations, like CRM or Office 365 in particular, we might be using different models for that.

That may be different from what we leverage for asking a question against data room content or for running the agent harness itself. Sorry, I can't tell you the exact single LLM because there is no exact single LLM. It is multi-model by nature, and that is intentional. The models that we're using for different things change all the time.

Doug Cullen (42:20)

Part of the benefit to the user is that you don't really ever have to think about it, right? It shouldn't be your job to decide. We have plenty of people who benchmark the models, test the questions, and figure out what's best for which types of questions. I'm asking one question. And really, the Blueflame agent is routing that, if you will, to a variety of different places to get the absolute best response possible.

Raj Bakhru (42:46)

Exactly, and that is more than a full-time job for a lot of people to figure out.

Doug Cullen (42:52)

Yeah, we're testing the models, we're doing performance benchmarking all the time. It's a crucial thing, and we certainly take it super seriously here. Okay, well, I think with that, we'll have to wrap up. Thank you both for such an amazing discussion.

This shift is amazing. Things are happening so quickly. I love your notion of a prompt-first dealmaker. I think this is just changing the way deals are going to get done. We showed a lot of sell-side use cases. We showed the Blueflame AI Assistant within the Diligence project. Then we jumped out to the enterprise Blueflame experience. Again, you really got that same type of admin flow, and we got a little peek into the future, I think, on what we believe buy-side diligence is going to look like.

We've got this MCP availability, the enabled content, and all of this is coming together because of our partnership, right? You can't do this type of groundbreaking innovation without being super connected across our strategic teams and our technical teams. I really believe the best is yet to come.

Raj Bakhru (44:04)

Yeah, it's getting better every day. It's been exciting to watch, and it's been fun to build.

Doug Cullen (44:09)

Thank you to all the people around the world who were able to join us today.

Thank you, Raj. Thank you, Alice. As a reminder, this is recorded, so please look for that on-demand replay and forward this to your friends and colleagues. We'll probably be back with future series, but this is a wrap for our MCP, or Datasite MCP, series. Once again, I want to thank everyone for tuning in. Thank you to my esteemed guests, and really enjoy and take advantage of this next level of dealmaking. We think the best is yet to come. Thanks, everyone.

Alice Esmerian (44:48)

Thank you.