Transcript: Datasite MCP for ChatGPT

Good morning, good afternoon and good evening for 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 Data Site MCP for ChatGPT webinar. This is the second in our series where we kind of go through the opportunities that are existing now that we've enabled and connected via MCP, their Data site, VDR, your best VDR in the world, to the various different frontier model platforms. And today we're going to be talking about GTT, which is a fantastic platform.

Before we get going, I just want to remind us of a few things. OK, So first of all, we want to hear from you. Please ask Q&A throughout the panel. We have reserved a little bit of time towards the end of this session to get to that question and answer. So please ask that. We also have a survey and we'd love to know what you think. There are additional resources around the console as well, some PDFs, a little bit about document and frequently asked questions, etcetera.

So check there for the resources. And this session will be made available on demand. So if you have colleagues that weren't able to attend live or you want to replay it to understand some of the things that we were able to demonstrate, 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, and then all opinions expressed are our own and not those of data site or Blue Flame AI. So with that, Alice, thank you so much for joining. I'd like to introduce everyone to Alice's Marion.

She's a product strategist at Blueflame AI, and I know you've been here for a few months, but if you don't mind just giving the audience a little bit of a background about what you're doing, a Blue Flame AI, and then what you did Prior to joining us. Sure. Thanks for having me today. So yeah, I did join Blue Flame six months ago in the product strategy team. My background is in investment banking and private equity, where I spent almost eight years. And so it'd be great for me to have the opportunity to work on products that really enable our investment makers and private equity members to leverage AI on their workflow.

So super excited to be here today and talking about a few workflows around the MCP product. Release awesome like we want to get into this is definitely show not tell. Real step change moment happening in deal making. I've been in deal making for almost 20 years now. Today myself been watching a lot of the transformation across how deals get done, run corporate development. We've done 9 acquisitions in the last couple years and I think this is the greatest time to be a deal maker and you know, really thinking about what we, we can take advantage with some of these, these new technologies.

But before I think, you know, one of the things that we wanted to do is we did launch MCP at deal Max. This was a a big deal for us here at data site allowing for the first ever connectivity from a data room, our data site diligence platform acquire platform to be connected to MCP. We'll be demonstrating enterprise ChatGPT, one of our favorite platforms here the day. And you know, it's kind of this a new age as we're talking about. Raj introduced the concept of a prompt first deal maker, which we think is amazing.

But you know, one of the things we also focus a lot on here is making sure this is done in, in a proper way, right? This is a serious business, right? We want to make sure that the it's offered in a compliant, resilient and effective way because you know that matters a lot in deals. Yeah, totally. And we're seeing it a lot today with our clients at Blue Flame. 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's at in their organization, what workflows were applicable or not.

Today's things have changed. Every organization is really trying to deploy AI consistently across the firm, across our workflows. And so it's really important that we enable them to power AI while staying totally safe, maintaining all compliance and government that are pre module to executing a deal in a Safeway, right? Like the data they are dealing with is so important and valuable that we really power them to to do that. So very exciting to leverage this MCP connector to bring this safe AI connectivity into they decide for the first. Time.

Yeah, it's, it's an incredible time. So let's just, we'll transition here quickly. I think it's, you know, we want to think about how AI is actually used. We want to demonstrate some of that, talk about it securely inside the data room. This is sort of one of the most important things. We, we launched the Blue Flame AI assistant within diligence today, actually a few weeks ago. And that is sort of the vehicle that we utilized to connect MCP. So we just wanted to ground some of the things that we're talking about for, for the audience because this came up, you know, earlier this week.

But I wanted to make sure there are a couple core concepts. So one is MCP. That's what we're talking about here. This is the connection. We've built an MCP server. We've published it via the ChatGPT through our great partners over there that makes an available connection for any user for enterprise ChatGPT to be able to connect and enable their data room. So this is always going to respect the role that you have, whether or not you are an admin, whether or not you are 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. So you're seeing here the Blue Flame AI assistant. And if you have not seen it yet, within diligence, our core platform, you have your standard diligence platform. There's a little right panel and a little icon on the upper right. You click that, that slides out and that gives you the full agentic experience of Blue Flame within the trusted environment of data Site. So you've got the data site project, you've got Blue Flame AI Assistant running within the data site project, and then that project itself can be connected to via MCP in enterprise ChatGPT.

So just wanted to kind of anchor those concepts. We'll be showcasing the MCP connection today because that's the sort of purpose, this is sort of the new innovation. But just to reinforce and under score that, that is powered by the diligence project that has the Blue Flame AI assistant enabled. And then there is Blue Flame, the enterprise company, right? So this is used by private equity firms, investment banks, corporate development executives have a separate Blue Flame subscription. So these are the sort of three dimensions that we want to talk through. Just wanted to clarify this is this came up on our first webinar and we wanted to try to give a little bit more grounding in terms of what we're going into.

So I think, you know, with that we can probably, you know, shift into the show not tell aspect of this and sort of highlight and talk through some of the powerful use cases that a platform like GPT brings the bear. Right now, one of the most stringent things when you get going is you have all this content, you're trying to figure a way to set up the data room. And you've done it many, many times I think in your life. And I think 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 what we can do with ChatGPT? So there are plenty of things we can do today. We're going to run through 4 workflows that are essential to the admin work within the data room. There are plenty more workflows that are applicable. So if you have any questions about more ID, more suggestions, feel free to add them to the Q&A section. We'll, we'll look at this afterwards. But today we'll power we'll look at four different workflows on top of delivering this MCP connectors and this blue flame enablement. We've built 8 skills, which are essentially for those not too familiar with this concept, almost 8 Chef recipes that helps the model being even more empowered with additional guidance as to how to understand queries around their room specific workflows.

And so these are going to be available at the GitHub link that will be shared with you in the resource of this webinar. And so I just wanted to pre phase that because these skills are going to be powering some of the workforce we're going to look at today. So the first one here is a simple one. And and when we go back to this deal making prompting, you can see that here we're not starting with anything sophisticated, right? It's 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 on. And so here we're just setting up the stage with a bit of information around the company, its industry and its size. And so immediately after asking for that, the models 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 immediately can provide some feedback around these. So you can see how a task that would have taken a lot of time when they're putting together a spreadsheet index or just reorganizing folders and naming and making sure everything was clean and with 0 typos is done from a simple prompt in a few minutes.

Yeah. And this is something that, you know, very often deal makers are sort of baptism by fire, right? They may, may or may not have ever set up an index. What does a good index look like? And 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 that the information is reflected about the underlying company. And so it's sort of that adage about a tidy house, like you really want to make sure everything's tidy, easy to find, easy to locate, follows a very logical setup.

And this is just a super powerful capability that the AI can help us with and GPT in this instance and just give us a great clean index. Yeah, absolutely. So you enter and you have automatically a great first impression of what you're going to experience within this data room. And so the index is skinned 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 and the model counts burning 53 seconds.

You can see on the screen that my phone, 14 folders and all the subfolders had been pushed to date site. So massive time save. That then gives me a lot more time to do everything that is more valuable for me as a admin creating a data room. Yeah, I mean, and just, it looks kind of magical. It is kind of magical able to get that content, put it in that very logical place with the logic, logical order. And now I think we're going to kind of go into maybe some some file naming conventions.

Yeah. So 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 is either looking at the existing name and clean and cleaning that up, which is totally doable with just the MCP without before you enable blue flame, because the MCP connectors has access to every single file names. And so it can see, for example, that if the company were working on here, it's called voxel matter, if it's cap letter, no cap letter VM instead of voxel matter or if it's just clunky names, it's going to be able to clean that up, clean the typos, etcetera.

But it's not going to be able to do until you enable blue Flame AI is actually access the content of this document and leverage this content to in inform how to rename a document. And a good example here is that I wanted all my contracts to have the year at which these contracts had been signed in the name so that I can clearly just orchestrate the the content. And so to do that you obviously need to enter a document, read it and and pull the name the year in the name, which the model was able to do because this data room is empowered with Blue Flame AI.

Yeah. So a couple steps here just to rewind the tape. I mean, one is to get the name of the document consistent, right, which in and of itself doesn't always happen. You know, many a data room I've gone into and I've seen the the famed book 007 dot XLS, right? What the heck is that? Who knows? Is it a financial model? Is it part of a customer queue? Like we have no idea. So getting the name into a logical place, super important, right?

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

And we're utilizing the ability to look when the contract itself was executed and apply that as part of the naming convention. So which again, as a, 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, spending a lot of time checking for typos or any inconsistencies. So again, this is something that you can automate so that you can spend more time on more valuable and informative task to get your process ready and prepared in the best form possible. One thing potentially that we wanted to highlight for people watching when we say that Blue Flame enables you to access information within the document, there's absolutely 0 information retention that is being done.

And This is why Blue Flame is particularly powerful and safe in a deal context, because any highly confidential information will never ever be retained by Blue Flame which. Yeah, no, no retention and then no training, right. So early days, people were really worried about this highly sensitive content being put into the LL Ms. to allow for the LL Ms. to be a trained. And so there's nothing of that going on here. We are a financial services kind of powerhouse and we understand that the content is highly confidential. So no training on the content and absolutely 0 retention to any of the content that is is being made available via this platform.

Yeah. And moving on to another workflow that honestly cost me a few hours and nights back in the days is AQ and a process. I think this is a great addition and a great way to leveraging this access to the content of your documents, right. So in my example here, I attached a list of Q&A question, which could be easier because at the at the beginning of your process, your VDD providers start asking you a question and information request around the asset. Later on when you start receiving plenty of questions from different buyers, all their advisors and everything comes to you in an unstructured manner.

You can then just drop the question. So here you can see I attached an Excel file with unstructured data. They're not organized either by SIM or even advisor or buyer. And then I'm just asking the model to pre populate draft answers to all of these questions. Now this is also powered by one of the skill that we're providing and something I'm very adamant about in everything we do with Blue Flame for our clients is making sure we have 0 hallucination. And so this skill prompt your agent to really look at the context and the content and make sure that it provides an honest opinion about its ability to answer to the question as well as very detailed, excuse me, VDR link and a section of the document that it used to pre populate the draft answer.

So 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, direct yourself in the VDR while having all the guardrails that enable you to trust and feel safe while using it. So this is, I actually got the benefit to be able to showcase this to a managing director here in New York for a very large investment bank. And he was sort of reflecting he's like this used to crush my weekends. You know, I mean this the availability of taking questions, basically uploading a set of questions to be able to allow those to be answered by the content of the data room, right?

So questions are being asked, they're being uploaded, they're being effectively answered with cited content, right? So we are literally putting the citation to the document that supported it as well as just to reiterate what Alice said, which is super fantastic, which is a bit of a confidence impact as well, right? A confidence score. Some questions they're going to be highly confident about reacting to, responding to based upon the document, the framing of the content. Other they're not going to be as confident, right? So we're able to really discern these two in the categories to be able to narrow in from a preliminary standpoint of 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 that.

And then of course, when you're live, the orchestration of Q&A is probably one of the most complicated tasks that we have. By doing a deal, you've got various questions, you've got various stakeholders, each person has their own series of experts, HR, legal compliance and we need to orchestrate that question from the buy side to the sell side to the expert and back. Totally. And you can see here on the screen that the model to all the questions, you then batch them by seams. So here on the screen, I'm showing you the commercial section for each question and topic, it could provide a status whether complete or partial.

And then for each single draft response you have, the documents are document that it used to address it as well as the page reference. And so this is then all agglomerated by seam into a single spreadsheet that you will see here on the screen in a few seconds. Really organize a nice by workflow with a number of question and their level of completeness. And then the total Q&A tracker, which for each single question, a draft response, a statues, the source reference, a citation and a follow up question required or additional data needed. So to me, as a banker, this is a great way to really start.

This Q&A process and accelerate it while keeping it absolute like valuable and of very strong quality. Yeah, just frames this problem. I mean, super valuable for the, you know, the person helping from the advisory perspective, whether that's a law firm, whether that's a bank, whether it's another financial advisor, but super value for the corporate having been, you know, selling the company a couple different times and selling different assets. You know, as as a senior exec, you're super curious around how the process is going, what types of questions are coming up. And so being able to be on top of it, I think is the sign of a great advisor.

And then being able to give this type of report to be able to give to to someone like me or Rusty or CEOII just really can't overstate how valuable that is to the underlying corporate. That's that's the kind of person behind the deal. Yeah. And so accelerating reporting and communication with your client, as you mentioned is something that is now accelerated with this MCP connection because you can edit activity in the data room extra. And so one more thing that I wanted to show today that ties to that is I think that a lot of people here will have some sensitivity to this topic.

But publishing a data room for the first time, when you add buyers in, you always had this little moment of anxiety to just wonder. I really hope I didn't leave anything contradictory that don't tie or just a document I won't open properly. Or is there any risk or what are the risks that potential buyers are going to flag immediately? Which is why we put together this skill that is called Pre launch readiness Report. And it's going to focus on confirming 3 main areas of your data room. The first one being it's going to run a gap analysis.

And so it's going to look at all of your sections, it's going to keep in mind this context it add around your deal and its industry and its size and provide feedback on how complete each section is. Is there any area where you might be lacking a bit of substance or context? And so it's going to provide feedback around that. It then also going to look at the quality of the document. So it's just going to run through. Is there any document that is blank contract that is unsigned, any PII information that has been not redacted? And so that could actually create a lot of risk around your data room.

So it's going to do that. And then finally it's going to run through all the potential risk that we provided in this scale around every single area of tax, finance, operations, ITIP. So it's going to kind of cross check all of these and gives you feedback around where are your largest risk and where buyers are going to pay most attention, which as an admin really gives you time to anticipate, OK, what do I need to prepare in my defensibility approach? How do I communicate this asset to my potential buyers? And so all of that is then consolidated into a single Word document that you can download and share with your client, with your team, and make sure everything is ready.

I know this. I remember to hear a story from a managing director that described to me every time before she launched the deal, she had to literally go through every single document that would take hours because ultimately it was her that was representing the asset and taking it out. So we think about the value from a managing director perspective, a senior banker or maybe even your analyst or your associate that has to do this. I'm sure the Alice 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 blue flame AI perspective or in GPT.

I mean, what power for both really. Everyone across the board, whether you're the corporate, whether you're a senior advisor, whether you're one on the on the deal team itself, I mean it's fantastic. You have this extra pair of eyes as if someone was really working with you while you're actually doing something else and reviewing the content for you. So that's very exciting and I really believe that everyone working on on the outside age rooms will find great value from from using those. Yeah. I mean, this is just, I think it's changing a lot.

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

It's incredible. It is great and although I think everyone is starting to get used to the speed of AI and GPT, everyone is so concerned about the accuracy and getting confidence around the content. So this is really a great way to to get in there. Yeah, this is it's incredible capability. We're having a lot of fun. We're early days with this. Just to reinforce this capability was launched I think about 3 weeks, 3 or 4 weeks ago. The feedback has been tremendous.

We're looking forward to a few questions here, which I think are great to be able to get. But you know one way to think about this too. From a good perspective, this is kind of the worst the models will ever be around answering these types of questions. Yeah, it's changing every day. So we are keeping up with all the innovation here and making sure that our prompt enabled deal makers are getting the best of of the tools. Yeah, you were saying even you're to leave today and she's going to go test one of the the latest models from from ChatGPT.

Open AI has been great with us in terms of sharing their most their latest models and we usually get our hands on those a bit early in a preview session just to uncover. It's sort of like this is like your Christmas Day, right? You get to open up new presents. Which happened now more than once a year, so it was great. So hey, let's, let's take a look at some of the questions. We have a a few here that I'm going to try to fly through.

I saw someones earlier, what kind of audit trail is available available on these auction? Oh, what, what type of audit trail is available on these actions? And the answer is a full audit trail. So we're able to capture everything that's been done, all the queries, all the information, and that will become part of the reporting package that is available in the data site. Let's see. So another one, how strong do you need to be with LLM prompting to use something like GPT and Datasite? It's going to depend, but honestly, the way we've build it is, and you've seen all my prompts on the slide, they're fairly straightforward as long as you're able to convey just the same way you would ask someone in 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 something that is very powerful when it comes to handling AI products. So try something. If it doesn't quite work, refine your prompt. But 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. Yeah, I mean, I think like everything with these alums, it's a little bit about you just got to kind of jump off the curb and, and, and try it. You know, I think your framing is really good in terms of hey, 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.

And then you have the ability to kind of do things with follow up questions, which are super important. Let's see, is it possible to give an give, for example, buyer advisors limited access to GPT output if you do not want to share full access. So these are dynamics that we're thinking through right now to be to reinforce these are capabilities right now, a lot of which you're seeing are really for the administrators of the projects only to get ahead of it yet. So do we intend to bring some of these capabilities to the reviewers and the buyers? The answer is yes. We've launched this, as I said, about 3 or 4 weeks ago.

So we're still early days. So these are administrative capabilities that we're demonstrating right here. There are a list of capabilities that you have no matter what your role is asking questions. How many projects do I have? And things, they're all very well defined and articulated. And honestly, you can just ask ChatGPT, say, 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 right in there. OK. So OK, so last one here and this is certainly, certainly worth reinforcing, but my question came for after reviewing the material fact sheets and FAQ you shared, it seems that the client's data does not leave the four walls of the data site environment exactly.

We know we took meticulous care in making the client content stay within that. I would sort of do this is like the secure boundary and barrier that is data site and all of this is still within that a secure operating in bottle. Now their extension out to MCP also kind of is governed and needs to operate by those same very kind of sacred set of rules. And so that is how it is. Anything else to add from that perspective? No. Again, it's a semantic search that goes into the data site without pulling anything from it.

And so this is how you operate the safest way leveraging AI into your data room with data site. Yeah. And we've gone through all the security reviews, etcetera, as as you might imagine, and we've taken meticulous care on bringing in this capability into the incredibly secure operating environment that is data site. 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. So with that, we're at time. Alice, thank you so much. It was amazing.

Thank you GPT, our great partners at at Open AI. And you know, 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 US and probably bank holidays all around Europe and and the UK, but we'll be back with our next series. So pay attention to those. A reminder, it 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 post webinar wrap UPS in the form of blogs as well as FAQs.

But you know, look, it's never been a more exciting time to be a deal maker and the capabilities that are available via the large language models and of course via incredibly focused almost vertical specific capabilities of someone like a blue flame and then bringing those into the most entrusted environment of diligence. I think is is really changing and transforming the way the deals get done. We're getting this sort of prompt first deal making and ultimately we want to make sure that we're delivering these results in a way that that people can trust and depend upon. You know it is AI so you got to double check things before making them always available.

Those without saying human in the loop always, but it's really a way to accelerate your your did making. Thank you once again, Alice. Thank you for those hundreds of people around the world that attended this webinar series. We're so pleased to be able to bring not only the innovative technology, but really this practitioner perspective to help you explore the world of AI. With that, I will thank you and have a wonderful day. Thanks everyone.