Name: How Contract Intelligence Powers M&A Deals From Due Diligence to Post-Merger Integration
Uploaded: 2025-04-22T18:16:09.274Z
Duration: 1 h 8 s
Description: How Contract Intelligence Powers M&A Deals From Due Diligence to Post-Merger Integration

Transcript for "How Contract Intelligence Powers M&A Deals From Due Diligence to Post-Merger Integration": All this banter energy. I just saw them starting. You know? And so as we begin to have folks jump in here, we're excited to have everyone. We're about to have an incredible conversation insofar as talking to folks about how to use contract intelligence and AI to do m and a's. And we're talking divestitures. We're talking acquisitions. We're talking post merger integration. We're talking due diligence. We're talking every step of the m and a process. We're gonna be showing you guys how to tackle, and we're gonna be discussing, yeah, how you maybe have done it in the past manually, but more interestingly, we're gonna discuss how to tackle these things with artificial intelligence. And so I'll go ahead and share my screen. Tom, Mike, making sure you guys can hear me well before we kick off. All good. Okay. Okay. Excellent. It looks like we have a good amount of folks trickling in already. We'll give folks just a few moments more before starting just because there's a lot of folks coming in right now, and do wanna make sure as many folks can start with us at the beginning as they can. What's interesting is that for folks who've come to our sessions in the past, the last time we did a session with Tom, from Microsoft, was before a significant m and a, which was Eversource m and a, and we recently acquired by Workday. So I'm excited to be diving in with Tom and talking about the Microsoft story, but from the first time for my new, you know, Workday home. And it's only fitting that we're gonna talk about m and a. And so without further ado, we'll go ahead and kick things off today, folks. First things first, we do have to share this product statement. We're gonna be showing a lot of capabilities, of the platform, but please keep the notes on this page in mind as you're looking at today's presentation. Awesome. Without further ado, my name is Memme Onwudiwe. I'm an AI evangelist here at Workday, one of the founding team members of Eversort. And I'm joined by Tom Orrison, who's a senior director of legal operations, and Mike Stevens, who's a senior solutions consultant here at Workday. Tom, Mike, thanks so much for joining us today. Thanks for having us. Good to be here. Thanks, Mary. Excellent. Excellent. Well, I'm really excited to have everyone hear our conversation today. You know, we've broken today into three different sections. Though it's really more of a half and half. And so the topic is how contract intelligence powers m and a. And as you can see, we're gonna be talking about things from early due diligence all the way through to post merger integration projects. Right? In the beginning though, we're gonna talk about what contract intelligence is, give you a little introduction. I know it's a bit of a new term. Many folks are familiar with contract life cycle management, but lesser so with actually mining the data and information in your agreements. And so we'll give you a base level understanding of contract intelligence, because once you have that base level understanding of contract intelligence, you'll then be able to appreciate how getting that full visibility into your contractual obligations allows you to tackle m and a more effectively. And then for the second half of today's conversation, we're gonna pass it over to Mike Stevens who's gonna really do an interactive demonstration that's actually showing some of the different processes and projects that we'll be talking about today, but showing you what it actually looks like to use it, to tackle these using AI. Because I know a lot of people folks like to talk about AI, and trust me, I love talking AI as much as the next guy. But here at Workday, we're much more about showing people, you know, some of the incredible artificial intelligence solutions that we've already developed. And so with this agenda, let's go ahead and kick off and just talk a little bit about what contract intelligence is. Excellent. I mean, at the highest level, when we think of contract intelligence, especially here in Workday world, you look at it as all of the things that give you insights into your agreements that have been signed and executed. And so what you see here is that we have the full contract life cycle management tool of Workday here, which includes everything including workflows when you request contracts, but also getting dashboards and data into your existing contracts. And it starts with connecting with your existing repositories, making sure that you're able to get all those contracts into one location. And, I mean, Tom, when we talk about this in the Microsoft context, you guys have millions of contracts. Right? And so this idea of connecting with repositories and then getting visibility and dashboards into your contracts must have been a massive effort. It's huge. And and being sophisticated in how you connect that and how you integrate those things is super critical because a lot of times when these questions come up about a corpus of contracts, there's urgency behind it. So having these mechanisms where things are integrated really pays dividends in the long run. Amen. Amen. And like we said, we're gonna focus on this side of the of the house here when we're talking about contract intelligence folks. So, you know, you might have thousands of scanned PDFs, you know, sitting in OneDrive, sitting in SharePoint, sitting in different tools like that. But you're actually able to take those scanned PDFs and turn them into actionable data, dashboards, insights, you know, really, really quickly. And so we'll be diving into that, you know, with with you guys. And I guess, Tom, as someone who's been doing this for some times, that fact that you don't need to kind of manually key in information and the fact that we can take scan the PDFs blurry documents, how much easier is it for a legal operations leader to adopt the tool when they don't have to do that kind of preprocessing? Oh, it's huge. Preprocessing for a lot of things just this is a very time consuming. So I've I've been in the trenches on ETL drive jobs, extract, transform, and load. They're they're a nightmare. They're a lot of times necessary to get stuff done, but having a system that solves a lot of that on our behalf makes life so much simpler. And also just provides that business velocity that's always so critical to everything we do. Amen. Amen. And so, I mean, that's at a high level. We're talking dashboards and such, but let's talk about some specific capabilities of contract intelligence. You know, one is and I think with the advent of generative AI and chat GPT, people are now more comfortable with chat interfaces. And I think what you can see here is the ability to actually ask questions of a contract. Right? And this is something that has been built out purpose built for contracts and other kinds of legal documents, and you'll see no kinds of hallucinations. And past hallucinations, you'll see that we're actually seeing where it says based on one there. We'll actually be showing you where the answers are being found in your contracts. Tom, has your team been, leveraging this kind of Ask AI feature? We really have. And so I've been in the contracts AI space for a while, partnered with Memme and others. Evisource has been a long term partner of ours, and we were at this before generative AI. And so early machine learning models using neural nets, we unlocked a lot of value, but the training load was was intense. It was massive. And it wasn't something that we could hand off to our attorneys and the practice groups to be really be able to do. It's something that we had to do. We had to spend a lot of time trading models to get to the accuracy we wanted. Again, a lot of upside there. A lot of value was gained by doing that. But being able to go into a corpus of agreements and just ask a simple natural language question really democratizes the power of this AI power of AI intelligence in a very material way. And so a lot of times when we're looking over a corpus of documents, my team will still do we still have, like, a center of excellence concept we call, where we'll refine kind of some of the prompts and make them more accurate. But the the ideation and the generation of what we're looking for, we pull the attorneys in a lot of times and have them just start asking questions. And it's really impressive seeing the light bulbs go off and the moments of how powerful this is. I love that. And I love the idea of having your attorneys ask questions to help understand what needs they have. And to your point, this is an individual level where you can have an attorney ask a question of a document, but we take things further. Right? We allow you to have custom AI, right, where you can actually train a new AI algorithm to track new information across all of your documents. Right? And so, like, here, we're looking at, hey. Can I track how much can a rent a a vendor raise the price annually, and you're able to do that across more documents? So to your point, once you have your lawyers asking those questions in Ask AI, if you see questions that are asked a lot, your team might say, hey. Let's turn this into a model that everyone has access to. That is the common pattern. Figuring out what resonates with the attorneys and then turning that into an actual data field so we can start doing dashboards, connect it into workflows. The the custom AI and and the and the creating data fields out of contracts is incredibly powerful. It's one of my favorite functionalities that have come out in the last three to four years. Exactly. And it's gonna be key and foundational to a lot of the m and a use cases that we're gonna be diving into today. So it's good that people have an understanding of this. I think what's also important is, you know, when you're training these AIs, that can seem daunting for folks. Right? But when Mike does his demo later, he'll show you how simple it is to train AI in our platform. And what's interesting is, one, we make it as simple as just telling the AI what you're looking for. And then once you tell the AI your instructions or what you're looking for, then it will go and find the information for you. But what's really cool is we have a capability called boost. Well, let's say you write a prompt, and let's say that prompt is kind of good. Maybe it catches what you're looking for in, you know, eight of your, you know, 80% of your documents. Right? But we can then have our artificial intelligence review your prompt and then try hundreds of different types of examples of prompts to find the perfect prompt that actually tracks what you're looking for accurately across all your documents. And so although it's easier than ever to train the new algorithm, we're even making that easier by once you've trained it. The AI goes and takes the training from you. And so here you can see the difference between two different prompts, One made by a human, but then one where AI actually understood what they were looking for and then tested hundreds of different prompts to find the perfect one. Tom, has your team been able to use this kind of capability? Love this capability. So personally, it's made me a better prompt engineer. And so I we've done a lot of experimentations with prompt engineering. I would say we're we're ahead of the curve from a department. I've leaned in very, very heavily. But it's almost like a prompt coach and not in some abstract way. It's a tangible tactical, I need to do this today, and it shows you how to formulate a better prompt. And I noticed as I continue to use it, the difference between my starting instructions and the the AI boost, the gap closes closer and closer, but I still get value out of the AI boost. But I can see that it's maybe just a better prompter, and that's that's applicable to a lot of different scenarios. So it's helped with my personal growth. And more importantly, it gets us more accuracy more quickly on the current project we're working on. Amen. Amen. No. That's that's that's excellent. I think that's a key capability. And then just to go over, there's other places AI comes in on contracts. This is less relevant to today's conversation, of course, being able to use AI to redline agreements to, you know, leverage your playbooks and say, hey. This is how we typically negotiate and have AI do that first pass negotiation. And then like I said previously, you know, Workday acquired Evisort. And, you know, I'm part of the founding team as Evisort. We started this company as law students, you know, back in 2016 out of the Harvard Innovation Lab. But one unique thing about our acquisition by Workday I mean, Workday was one of our biggest customers, one of our happiest customers. We joke the happiest customer because they bought the entire thing. But what was interesting was that when we were acquired by Workday, Workday used their Evisort platform to perform due diligence on Evisort. Right? And so, it was a little bit of, a little bit of inception in there. Right? But we actually were being used to analyze ourselves. Right? And, luckily, things came out great, and we were able to be acquired. But I I just think that's it just shows kind of how powerful this technology can be and just an example that we have on m and a. But, of course, no one came here that to hear about my experience, you know, performing m and a, whatever. So we're here to talk about your experience, Tom. And so, I'll pop over to that now. But I guess, lastly, if you are gonna be leveraging any kind of AI solutions, especially for legal use cases when you're exposing your most confidential documents, you've got to make sure that any solution you're working with, because the list of solutions is expanding every day, you've gotta make sure it's one that's steeped in responsible AI. One way of knowing that is by the new ISO certification, ISO four two zero zero one. This is something that Eversource it. It was one of the first companies of the world to earn an accredited certification in and definitely something that you should be looking for as you're evaluating AI solutions for your legal teams out there. Huge kudos to the Workday team for for going through and getting that ISO certification. Getting comfortable with a responsible AI and security and privacy with these systems is a daunting task. And having these certifications make purchasers of these things like myself much easier and much more comfortable that we're making the right decision when we see this kind of third party validation on these things. Amen. And thank you. But let's move on to tackling m and a with artificial intelligence, Tom. And, I mean, you know, to understand that to tackle m and a with artificial intelligence, you almost first have to understand how you tackled m and a before artificial intelligence. Right? Could you speak a little bit from a legal operations perspective? Right? What's it like performing some of these large scale processes like due diligence, divestiture, post merger integration? What's it like doing it without artificial intelligence? I'll start with a simple explanation, and I'll I'll go with painful. Painful was the approach before. And so the old way of doing it was gathering as many contracts as you could gather. It has advanced a long ways to data rooms, but getting that kind of gathering is is a pain in and of itself. But from there, it is the roomful of attorneys, usually junior associates, in a windowless office somewhere, spending twelve hours a day reviewing thousands and thousands of pieces of paper. It's a miserable task. I question how accurate it is after the eighth hour on the third day of reviewing the same documents over and over. And typically, the way it the way it forms is and every deal is somewhat unique for us, but we'll have a handful of things, three to five things that are most important about this. And we'll want that first pass due diligence to really focus on those three to five things. And so mass amounts of manual review, I think that's then changed to more electronic review with a lot of control f's in in PDF and Word. You had the problem of poorly scanned images, that were difficult to even read on a screen. A lot of issues with that. And then that was just the first pass. And from there, issues were identified. Those same contracts were then reviewed by more senior associates for a second pass. That was discussed internally and with outside counsel. And then finally, we would get the partner overlay on there. And that whole cycle would take weeks and weeks and weeks. And all of our m and a activity from acquisitions to divestitures, even the post merger integration, there's always a heightened sense of urgency to all of those things. And so not only was it painful, it was very, very time consuming. And, again, I question the accuracy of that model. I just know human behavior, and after eight hours in a windowless conference room reviewing the same documents over and over, even the most diligent of human gets tired at some point. And so it's it is painful and time consuming. I'll end with a simplistic wrap up. Yeah. No. No. No. And I think that I think that sums it up rather well. And, frankly, you know, even as we were first developing Eversor because, you know, we started with the AI analysis. We didn't build out the CLM capabilities for about three, four years into our company. But it was use cases like this where it's, why am I being paid hundreds of thousands of dollars as a junior lawyer to be locked up in this windowless room and reviewing manually thousands of documents just looking for the word assignment. You know, in a world where I can go on Google and find anything in the world, shouldn't AI be doing this? Right? And so I think this is a, to your point, a really, you know, good example of some of those, just just some of those kind of, you know, issues and kind of how we practice law. And could you talk a little bit more just to the costs? Right? Like, to the you know, would you have to pay outside law firms for this? You know, you know, how many hours of your team is going into this? Yeah. A roomful of associates for weeks at a time reviewing documents is a very expensive endeavor. And so the clock starts ticking, and that ticker goes, and times 10 to 15 associates depending on the size of the load that they have to review. It is one of the more costly aspects, of of an m and a activity. And so it's early on, it spends a lot of money. And a lot of times many times, it's it's wasted money because you uncover something in that review that stops that deal from going forward. And so a lot of money to get to that point. One of the things that was kind of eye opening for me in our first foray into m and a with AI, and this was eight years ago, was a lot of times the deal team's saving outside counsel was secondary. It was the business velocity and getting to those gotchas sooner than later, was huge because we could unwind things and save the company all up a lot more money, not just in legal fees, but overall m and a motions if we're gonna uncover those gotchas really, really, really, really quickly. And so that's always been a kind of our main point of using AI is how do we get that business velocity and strength the time for us to get comfortable that we're making the right decision. Well, let's talk a little bit more about some of these gotchas. What are some of the unique clauses or the data points that are important to track in your agreements, you know, doing these m and a processes? I mean, there's basic things like reviewing the assignment, figure out what customers you might be able to bring bring along with. That's always kind of front and center. For us, a lot of times, and on on when we're acquiring smaller companies, it it's a lot of times around IP and and and talent. And so you have to couple kind of the the the IP assignments along with employment agreements. We have found instances where restrictions on patent assignments are embedded in employee agreements. And those are those gotchas that we're looking for. We wanna make sure that that things are gonna limit our ability to to to monetize our advance and acquisition. We wanna figure those things out immediately so we kind of can right scope the the the upside of of a particular activity. Yeah. That would have been a big gotcha if Workday acquired Everswip, and then I owned all the IP. Right? And so, like, you've gotta you've gotta make sure you've gotta make sure of that of that piece. And, I mean, I think what's what's important here is that we talked about how so much of contract intelligence is passively structuring contract data, passively having visibility into a lot of these insights. And so things like assignment, you might already be tracking. But to your point, there's gonna always be new information you need from your documents. We always joke, you're always tracking x amount of things in your contracts. You always need x plus one. Right? And so in when these unique cases are, well, all of a sudden you're looking for, you know, patent language in your employment agreements, you need flexible tools that can help you track that additional data point that might not have been conceived when you first adopted the tool. I would say it's more like x plus three. Yeah. Let's go. Nice. Maybe three or five. And generative AI really opened up that plus three or five. Because before, when you had to have so many examples of what you're looking for to get the level of accuracy, that took a lot of training time when time was of the essence. And being able to go there and quickly pull out those custom things you're looking for that change deal by deal makes us so much more agile and provides that, again, necessary business velocity that our teams are looking for. Yeah. No. That's still important. I mean, you your team has trained hundreds of algorithms in Everswet's platform, but you did some of those in the before times before generative AI where you probably need to go find more examples of language to train it. But you've trained several in these new times where it's really just as simple as telling the platform what you're looking for, and then you start finding it. And then like we said, Mike will be actually showing folks how to do this within the platform in the second half of this, of this, session today. Excellent. Well, let's dive in a little bit into using AI for m and a due diligence, for the divestiture, for post merger integration, you know, for everything. What does it you've set the scene of having, you know, young lawyers locked up in rooms reviewing documents, spending hours of law firm time. You know, what how would you describe the experience of doing these m and a activities with the benefit of AI? So one, I think we would get quicker results, more accurate results. And I think the attorneys and the outside counsel that we partner with have a better job in the process. Memme. So they're using their their a lot of times, their their very sophisticated lawyer brains, not on reviewing thousands of the same contract, but in analyzing the results of that review. Where they provide the most value, the most unique insights that really help shape our decision making. And that's where they can optimize their time. Same for internal. So Microsoft's a huge company. We have a lot of M and A activity. So the site I'll I'll start with the second question. So a lot of the instances where we're using AI in M and A activities, we're doing that in partnership with our law firms. A lot of this work we've we've handed off to our law firms, and we don't have any intention of bringing that back in house for some of our midsize m and a's. And so we really encourage and have benefited greatly from our law firm partners leaning in and using AI to drive this level of intelligence. So we we've been we've been at this for a very very long time. So a little kind of just a short history of of our involvement in this. And so my foray into kind of contract AI started eight years ago with, it was called the AI school that Microsoft Research stood up. This was before our partnership with Evisort, just before. But we're using a lot of, like, untested model untested neural networks and really using tens of thousands of examples to kind of train the models to be able to mimic what a first pass due diligence review would look like. We were astonished with the results at the time, but looking back, the amount of effort it took to get there, there was still huge, huge upside. But it was just a mountain of effort. And we really had to have, like, true data scientists, researchers at our disposal to be able to do that. So that was not a scalable model. Fast forward probably six months, and I had the great benefit of meeting Evisort, Microsoft's m 12, Investinar, made made an investment in Evisort. They made the connection, and it looked very much like we were doing with a research scientist, but in a platform, we could actually achieve some scale. So one of the first stories I'll tell is is one of our largest acquisitions recently. We had a lot of back and forth with the FTC to get final sign off and approval for it. And so we had we really used it for these post merger activities. I wouldn't even call it integration, but in our interaction with the FTC, they had FTC had made very broad requests for contracts to be produced as part of the review. We used Evisort, and I would say it, in a very novel way to go into that large set of contracts that the FTC asked for and use the kind of requirements they gave us to really whittle that down to the most relevant and responsive contracts. We still had to have attorneys review those and pay attention to those, but we're able to go from 25 corpus of 25,000 contracts down to less than 4,000. So by doing that, we were able to cut about 80% of our outside counsel time, kind of cost avoidance, I call it, from reviewing those nonresponsive agreements. It was huge upside. More importantly, again, for the business team, they they they liked the outside counsel saving and the cost avoidance. But what they really appreciated was the quickness in which we're able to get to that level of insight. Way faster than any kind of human review. We're able to turn it around even with the old machine learning models. In about '45 48, we're getting a lot of really solid preliminary results. So that was an early foray into it. We're very, very proud of. We offset almost, you know, upwards of I I ballparked about 1 and a half to 2 million dollars of outside counsel avoidance in just that one instance. Again, these are large, large m and a's and so the the denominator of what we're saving from is very large as well. But we're able to whittle that down and, again, provide that business velocity. Fast forward to today and the power of Evisort with the conversational AI, and we're moving into a model where we are getting the early intelligence. As soon as we start engaging in some of these activities, we wanna load those contracts into Evisort so that we get the understanding of that rich detail level and also the agility to ask new questions as the deal kind of progresses throughout its life cycle. Stuff always comes up as you move through the deals. And so that having that ability to go back and quickly analyze that with new insights or new questions is hugely powerful. Wow. I mean, that's that's just incredible. I mean, like you said, 2 millions dollars saved. But even past that, you know, m and a is a series of relationships. Right? And that relationship with the FTC is one one that you're gonna need to foster for future m and a's too. But two, the fact that they were impressed by the speed with which you were able to respond, you know, hopefully, help deal dynamics from that it's a there's a there's a lot of benefit to come from that. Thanks so much for kinda sharing, you know, that experience. Yeah. On on that overlay, we we were working with the FTC and our outside counsel. So accuracy was paramount. It was something that we focused a lot of time on, and we were able to achieve a level of accuracy that not just myself but our outside counsel were very comfortable with with that resource. That was that was for a large scale project. That was an moment for me of, wow, we were able to use this and and get value, well, from a cost savings and business velocity, and one of the most important things Microsoft was up to at that time. You know, and that's often overlooked because you can have an AI tool, but if you have to review it manually because it's not accurate, you're not saving any time. You're just, like, you're spending more time. Right? And you're kind of doing that work over again. And so, like, having AI tool does nothing for you if it's not accurate. Right? If it's not if it's something that you're gonna have to manually go over again, then it's just taking more time and increasing cost. And so I I love that, capabilities are at a level where you knew you could trust it for such high level projects because this is really where the impact of AI is felt and if you can really truly automate some of these processes. One of my favorite stories from that from that ABK deal was we were working with outside counsel, and there was three outside three outside firms involved. And our first kickoff meeting with the firms, one of the ladies joined, and she's like, I can't wait to hear how you're gonna show me how AI is gonna take my job. But then in about a week's time, she was one of our most active attorney users inside of Evisort, and in there knocking around on a daily basis. So in a week's time, we went from a full naysayer to, like, a full adopter and almost like evangelist to the power of the tool. That was another moment for me is seeing the attorneys react in that in in that short of time frame. The the value was just so apparent that we were changing mindsets in the process. That's beautiful. Well, then I guess I I've gotta ask, would you ever consider doing an m and a without AI again? If it was my choice, never ever again. I can't say we'll never do it because business circumstances always change on us, but it is something we would care not to do without AI again. Awesome. Awesome. And then for those listening on the the call who only know m and a manually and would love to see the world of what it's like to actually use this do this processes doing, AI, let's move on to a demo. And, Tom, of course, please, I'd love to kinda see your feedback as Mike kinda dives into the platform and shows folks, you know, what this stuff actually looks like. But I think everyone is So let me quick, guys. Yeah. I wanna make a point before we get into the demo just to help the audience. And so if we have folks in the audience who are just kind of embarking on this journey, you're gonna meet resistance of folks that are worried about the accuracy, worried about using an AI system. Think about doing a a few smaller deals in parallel where you're running the AI along with the manual process. You learn a lot by doing that. And more importantly, you shift that mind share of both your internal attorneys and your outside counsel to get more com comfortable that this stuff is actually working, that is a very, very critical step in getting this kind of broad use capability across your departments. Awesome. No. That's a great point. And we should definitely, as we're going through this, talk about strategies that folks can take back to actually get this going because technology is there, but frankly, a lot of times technology trails culture. Right? And so if folks aren't ready for adoption, we need to get them ready. And the technology might be the the last piece of that. And so, yeah, I'll pass it over to Mike without further ado. Awesome. Thanks, y'all. Great discussion so far. So I'm just gonna share my screen real quick and get into it. So I've actually been one of those junior attorneys in the windowless room passing around physical files and doing this the hard way. That's why I do this for a living now because, you know, hopefully, you'll see today in this demonstration, there's just a much better, less soul crushing, more way to do this. It's also a lot faster and more accurate. So wanna kinda bring some life to some of these examples that, Tom and Memme have gone over. We just wanna highlight at the start that, you know, m and a is really the starkest example that I can think of of the challenges that you face in contract management overall. Right? Because we have a limited amount of time in a high pressure environment to identify a large amount of information at scale. And the consequences of doing that well or not so well, are pretty stark. And you can look at, you know, number of acquisitions of major companies over the years that have gone well or not, and see what a big difference that can make. So we're gonna highlight, you know, how you actually on the ground would use this technology. And so I wanna start that just at the highest level. So we've made our models, we have our data, we've dumped our documents in at scale as you mentioned that is very possible in Edisort. And now we can jump right to looking at a dashboard of a lot of the key elements that we're gonna care about on both sides of transaction. Both the seller and the buyer have to worry about these things. I want to focus first on the buyer's experience. So that's first of all in the due diligence phase where we're going to want to know you know what kind of contracts are we looking at that we're going to be inheriting from this company of what different types. Maybe what is our contract value is gonna be critical so we can high grade our analysis, you know, to start with those most important ones first to identify issues. And then even things like we mentioned those IP assignment restrictions. Right? I'm sure when Workday was looking at Evisort, they they were looking at our founders agreements to say, well, y'all do y'all own this or not, right? So these are types of things that we can identify just right out right out of the gates, and rather than this pops up halfway through our evaluation and now we have a fire drill on our hands. So, some of the things you'll wanna look at during due diligence, of course, you're also gonna wanna know what responsibilities are you gonna have pretty much day one once you fully own these contracts. When do I have contracts renewing? How much term is left on those to help me do things like price the agreement or price the transaction? Things of that nature. And so we can orient these dashboards around this AI created data, and I'm gonna go into how we create this data and how we get it. But hopefully, you can see the value of, you know, how would you create something like this manually. Right? It would be a bunch of people in those rooms typing into spreadsheets, generating reports, and just really difficult to do. And it would take further time to analyze that data to generate things like how do we think about our risk scoring of these contracts in terms of the potential liability that we're taking on for these contracts. So lots of different things here that we're gonna get into, but I wanna start at even just a higher level or maybe as soon as I, you know, load these agreements into Evasor, I might just wanna ask kind of a broad open ended question using our generative AI capabilities. And this is called Ask AI. This uses our proprietary LLM, that's been fine tuned on millions of contracts so we can expect really, contract specific and intelligent results. And even ask you to question as high level as, hey. Highlight the risks in these contracts that are likely to affect me, the buyer, in the next year. It's gonna scroll through a large volume of these contracts, your entire corpus, and highlight a few of those things. So this is like, sort of give me the high level here before I dig in more deeply into some of these other models that we've generated. So I just wanna highlight for you, kind of the overarching capabilities and what generative AI has unlocked. And so as we see this response here, as it give it a couple seconds to load, it's gonna give us a couple things to sort of guide our attention to as opposed to we need to think of everything ahead of time that we might need to look for. We can get some early cues, from the generative AI technology just right off the bat. Right? So these are things like as we see here, we may have risks for product delivery and supply risks where, you know, if we can't renew a certain supply, agreement, we may have trouble, continuing to meet the needs of our customers. Or we may have price increased risks that maybe we weren't told about, but are certainly present in those agreements that are gonna help us price this deal, of course, and then we're gonna actually to cut those checks at some point as well. So things we wanna know about. And then things like warranty, financial risks, are you gonna be able to comply with your insurance requirements? Right? There could be a million different things, but being able to ask a high level question like this to just sort of guide our attention upfront to some of these key things could be a great step, that works in conjunction with, you know, our ability to build these more nuanced models that I'll I'll go through in a little more detail, to find those, you know, more particular data elements that we care about. So may I just ask Tom? Just kinda curious, like, you know, how would this change, you know, that ability to ask those overarching questions be kind of a step change even over and above kind of the old world where you had some of those machine learning models, but you had to be, you know, really focused to know exactly what you were looking for. How how does this change the game even further from your perspective? So on various fronts. One, I think it democratizes it so you can pull more people into the process. I think it's really good for ideation as you're exploring having that interactive capability to kind of interact with the models and the contracts. I think this gets the creative juices flowing and we uncover more things that we might wanna look for. And then there's the speed in which we can get to the results. So on many fronts, it really really a step change for us. Yeah. Awesome. And I've certainly seen it every day as a, you know, user of this software on a daily basis that it just unlocks so much more that you can do when we can really ask analytical type questions rather than just, hey, go and pull out this language that I have to then interpret. Right? We can sort of skip right to the bottom line, in asking those questions. And we may not even know what the question is to ask, but you tell me what I should care about here. So, just another example of how we're bringing greater intelligence to this process. And so I wanna dig in, maybe to go a little deeper on some of these things. You know, obviously, I mentioned risk scoring and we're gonna wanna know things about, like, hey, is there a preferential right out there that maybe our seller didn't tell us about that, hey, they actually have to go and offer this property or maybe this whole company to someone else first. Right? That may change the amount that you're willing to offer for that. And I've been involved, certainly in transactions where that's been the case. And I can tell you when, one of these pops up that's important. A lot of people higher up in the corporate ladder start talking to you than ever did before. Right? These are things that can have a huge effect on the bottom line of a transaction or even if that transaction occurs. So this is you know, these are oriented around due diligence, kind of questions of, well, do I want to enter into this transaction in the first place? But, you know, assuming the answer is yes after we get to that, you know, fast no that you mentioned, Tom, that, you know, if we do find these, kind of land mines, we get out, we go to the next evaluation a lot more fast or a lot more quickly. But then we say, okay. Yeah. This is a great deal. We do wanna make this acquisition. We enter the transaction, do all of the, you know, crossing the t's and dotting the i's there. But now we actually it's real. Now we've actually acquired this company and we have to immediately begin to administer these contracts and do post merger integration. So that's a lot of things are involved in that process like, you know, technology wise, building in these new processes and employees, obviously. But then there's also a question, in particular, a contract consolidation. Right? So if you're I don't know how many contracts you had in common for similar services with Activision, but, you know, if you were to buy a direct competitor, for instance, you would probably expect that you'd have a lot of overlapping contracts where, you know, you're providing the same, kind of service. I really don't need to. Right? So maybe I can just walk you all through a quick example of how we would use a dashboard like this, to start to go through that process and identify without opening a million agreements, you know, where we might have opportunities, to consolidate contracts and really get down to what we really need out of a transaction, maybe let some of these contracts go. Right? So Yeah. And, Mike, as you pull that up, let me just answer a couple questions that are coming in through the q and a, and we've got a lot of great questions, folks. Please continue to to ask questions then we got staff who'll be answering directly in there too. You know, there was a question around whether Microsoft is using the CLM or the contract intelligence, which is the more advanced search and analytics tools. Microsoft is leveraging those more advanced search and analytics tools. You guys have a lot of CLMs internally being a planet sized company. And the last thing you needed was one more CLM, which you really needed was to get that kind of centralized view across your kind of data and documents. And, I mean, I think what you're seeing here, this consolidated view, these could be documents in different repositories. They could be some scanned PDFs. They could be blurry and, but we're able to consolidate them and give you that overall view. And what Mike is about to show you, post merger integration, what would we, if we were just acquired by Workday, know of post merger integration. Right? But, it can be a long process, but having tools like this can really help things, kinda streamline. So I did wanna answer that. And if folks have more questions, feel free to put them into the q and a. Thanks, Carmen, for that last one. I'll pass back to you, Mike. Yeah. Thanks, Mary. So, yeah, let's just run with that example. Let's say for something like professional services contracts, which first of all, just getting the header level of what kind of contract this is is helpful. And then we see as we interact with that, we're gonna get a a live updated business intelligence view of this other key information. So now I can see for those types of contracts, I have two where accelerate, in this, case of counterparty, is my counterparty to that contract. So they may be offering the same good or service. That could be another model that we add. You know, what services are they providing? But we may find that, well, this is for exactly the same services as the company that we're acquiring. We really don't need both of those. So maybe now I say, well, I'd love to not pay twice. Right? Because, you know, these aren't free. So I could scroll down and see, well, one of these actually has a termination termination for convenience provision. Right? So I actually now see that I have the means to, you know, conduct this consolidation by exercising that right to give my thirty days notice and get out of this contract. But which contract is that? Right? This is also working like a search. So as I interact, I find that little diamond in the rough. Now I can jump into it and get a more holistic view of that contract to see, you know, is there any other reason why I might need to keep this, or is it truly a good candidate for consolidation? And you can see as we get into a little more granular view of what Evisort can do, we're seeing that this is a super ugly contract. Right? We got handwriting. I put my coffee on it. Got a big coffee stain in the middle. We're still gonna be able to pull out all this information like that t for c date. So I might come into this contract now and just wanna do a little deeper analysis. Like, you know, I could ask a question, like, you know, as simple as, as you see here, I already asked it, summarize termination. And it'll give me sort of a consolidated answer about what could be a pretty complex and spread out, question throughout the agreement. And this just lets me get to that final answer, a lot quick a lot more quickly. But while I'm here, I kinda wanna dig into something that you raised, Tom, which is, you know, what was kind of the old way a couple years ago that we did this where we still use AI to great effect, and now how does generative AI, take it to a whole other level? And so an example I wanna walk through real quick. So let's say you're on the seller side or maybe you understand the buy side, doesn't matter. You know, what are the assignment restrictions? Do I have to give notice? Do I have to get written consent? I can use these models that EBSorts had for a while to surface this information, you know, the text of this language at scale, which is super helpful and incredibly transformative when you have to scale Microsoft. But I still have to read this to actually answer my question. Right? I have to interpret all of these variations of language, which can take quite a bit of time. So what if instead I could just jump straight to the that question and ask, can Greengrass assign this agreement? It's going to actually take that final mile of human analysis that typically humans had to step in and consume these inputs, transform that into the real answer to the question, and then store that somewhere. We're able to do that all automatically going straight to the actual question that sort of prompting this, in the first place. So, Tom, maybe I'll throw another question to you. Like, back in those days, how much time would you kind of ballpark that you would spend on interpreting AI outputs to kinda get you that final mile to actually, you know, get the data that you can act on? As you mentioned, there's still a massive amount. If you're just pulling clause language, there's still a massive amount of review. You're, I would say, successfully directing human eyeballs to the right text to read, but there's still a lot of consumption of that text. And so a lot of times, like you just said, a lot of times you don't you're like you wanna bucket those assignment provisions into, do I have to give notice? Do I have to get consent? How can I turn that into data that I can then action and accelerate my workflows without having to do that additional human review of the clause language? Yeah. So just another example how we're kinda pushing back the frontier of where the human brain's really needed because that human brain's got a lot of stuff to worry about. Right? So as much as we can offload, the better. You know? And that's even analysis like, you know, maybe I wanna give a risk score that, you know, brings in how I think about certain criteria, to give us a very, you know, Microsoft or whoever specific answer. So for instance, on this one, I'm just asking, hey. Rate the LOL provision, on this contract, low to high. And I'm not really giving it a whole lot of criteria, but it's gonna give me a pretty good answer and analysis taking into account all of the factors that go into this about things like the, cap of damages. You know, are there carve outs? Are there special type of damages like indirect that we wanna exclude? And so we can get a nice summary of this kind of information. But then, you know, the the question becomes Memme, you mentioned this. If we have to come into every agreement and touch it, this isn't as helpful. Right? So a really big leap that we introduced right after we actually introduced Ask AI is called X-ray. We're actually able to take this question and turn it into a model so that this question will automatically be asked across all of our contracts without us having to click into them, and we can just jump right to the answer. So I can just turn this into a model, and then I'll show you the more built out version of what that looks like where, you know, I can actually define, my risk criteria to say, you know, this is what I think of as high risk related to the damage cap in this case. This I think of medium risk, etcetera. And so it'll take us through a short evaluation process to kinda make sure that it's capturing the data in the way that we want, and we just tell it where we want that answer to live so we can put it on that dashboard that we already saw. So then after a short evaluation period where you're looking at maybe 15 to 20 of these kind of sample, returns to make sure it's performing like you want. We can say that looks great. We wanna go ahead and deploy this model and get those answers across hundreds of thousands of agreements, and even maybe use that boost feature that Memme showed where it can actually help me write the prompt, which I use this all the time as a solutions consultant in in drafting these models, for our prospects and customers, and it just makes it so much faster. So I I've also been in the space a while where I I had to build those old machine learning models, and it was like, great. I have 200 samples of this language, and I gotta go in and label it. Could take me weeks to build the model. This is something I can do in just a couple minutes now and have it automatically deploy, at scale for me. And then the end user experience for most people is just gonna be viewing that dashboard that I showed you all moment ago where these insights are all rolled up for us. Yeah. I mean, this is so exciting because this in that screen you're just showing. Right? Like, the AI is doing the work of finding that example. Right? Like, before, you might have to go and say, this is where the example is in this document. This is where the example is in this document. But now once you have these instructions, you just give the AI the instructions and says, hey. I found this in here. Is this correct? And all you're doing is saying, yes. It's incorrect. Though it's not correct. And then you're going from there to build the model. Right? And so if you can just identify something which says, hey, is this an assignment for us that's high risk? Is this a low risk or medium? You're able to do that work versus, you know, having to go comb through your thousands of agreements to find every example that you've done yourself, right? And so the AI jumping in here, I think, is just so so incredible, and, I mean, it's Yeah. I I think that another key piece here is like you said, people don't wanna review a bunch of indemnities anymore. Right? If you've got a thousand indemnities and some of them are, you know, high risk, you only wanna review the high risk ones. Right? So why you're reviewing the lowest ones in the first place? So I I just love everything that that we're saying. Pro tip for folks, the more context you can provide. So if you see an incorrect response and you click incorrect, and if you can provide more context, that AI boost is gonna be even more valuable and meaningful because it'll digest that context as well. So if you can tell it why it was wrong, particularly as you get more nuanced in the prompting and you're looking for very specific things, you can really tighten up, I would say, the neck you're casting, by giving context to the risk to the incorrect, responses. Memme. Absolutely. And part of that really gets to, like, how fast is the iteration time. Right? So think about if you gave humans these instructions and said, go look at all these contracts and find that, they would do a good job. Right? They would go back and they would figure all that out. But now maybe roll up that information and take it to someone more senior and they say, no. Actually, I don't really think 25,000 is the number. I think it's 50,000. Guess what? You're handing all those contracts back out and you're doing all that analysis again. Whereas here, I can simply tweak this number, rerun the model, and get fresh data. Right? So the the point of that is that human review does not scale when you're iterating. It just multiplies. Right? If you have if it takes a thousand hours to look at it once, it's gonna take you a thousand hours to look at it the second time. AI, because it works so much faster, allows for this iteration, allows you to get really, you know, accurate and and nuanced and granular with the type of data, you know, when in reality during the course of a deal, new questions are gonna come up, your perspective is gonna shift a little bit, and we don't wanna have to go back to the well in a way that's gonna slow down that deal or be, you know, extremely cost prohibitive. Yeah. And and additionally, the incentive structure of manual review will always have you reviewing the bare minimum. Right? Because it's costly to track just one more thing. And so if you have one more thing you wanna know, you're then weighing. Is it worth knowing that one more thing? Right? I think of one of the due diligence is, we helped out with doing m and a. There's a data point that we track out of the box that had massive impact and something most people don't look for. It was execution status. The company that wanted to be acquired had sent over all of their sales agreement saying, this is the valuation of our company. Look at all our customers. But our AI said, hey. Like, a few of these significant sales agreements don't have a countersigned signature on them. They're not signed. And so the company acquiring was able to go back to them and say, unless you provide countersigned versions of some of these customer agreements, we will have it actually we'll take some of that ARR, that annual recurring revenue that you're recognizing away from your valuation. They were actually able to materially reduce the amount that they paid for that acquisition based off that execution status, but that's not something that lawyers are typically looking for when they're doing a traditional, you know, due diligence. And, I I would even say there's a question from Ken talking about how typically in due diligence when people send over the data, they'll use a virtual data room and they'll put all these restrictions. We're not allowed to download them. It's read only. But I think that in this modern world, when you're performing these due diligence, m and a, it is not a litigation. Right? You guys are partners in this. You guys are working together. Right? And if you go to your partner and say, hey. We actually don't wanna spend millions of dollars reviewing this. We'd rather use AI and do it quickly and more accurately. There's no doubt that they'll be able to transmit the documents to you in a way that allow you to complete that. Right? And so I would think of it as a fact that both sides of this transaction want things to move smoothly, and your partner should definitely be able to help you get the documents in a way that allows you to leverage this more advanced technology. Because if not, they're just increasing your transaction costs, which might be passed through in the deal anyway. Exactly. And, actually, I have a great example that I lived through of, where this really kind of mundane seeming data point of whether or not something's executed is actually a big deal. And it was that prefrite one that I mentioned where we identified early on in the transaction that there was a prefrite. Therefore, this needed to be offered to the owners of that prefrite, which takes an amount of time they get a notice period. And that was, that deal was actually linked to a sort of tax maneuver that they're trying to do, called the like kind exchange where we did this deal within thirty days, you get a big tax break. And so they were afraid that that prefrite was gonna move it past that date to the extent that they weren't gonna get that tax break, and it was gonna cost a hundred and $50,000,000 in tax expenditures if they actually had to offer that. But what they noticed after I spent many hours in that room reviewing all those contracts is that that agreement wasn't even fully executed. So we did just a whole rigmarole and wasted a bunch of time and money reviewing an issue that seemed like an issue, but if we actually had access to all the data in a ready way like this, was actually just, you know, irrelevant, essentially. Right? So to your point earlier, Tom, about, you know, whittling down the contracts that you need to look at, a lot of the power of this tool isn't just reading things faster, it's not reading things that you don't need to read. Right? Finding unresponsive contracts without analysis really isn't necessary. So I kinda shifted over here to the seller dashboard where, you know, if I'm prepping something to sell, I probably also wanna know how hard is it gonna be for me to, make these assignments. Right? Am I gonna have to give notice? Am I gonna have to get consent? And you may find that, you know, this can be a a big deal stopper, and you're gonna wanna not only know what those obligations are, but you're gonna wanna track your status of actually receiving those consents and giving those notifications, which are also things that we can track and dashboard. So it's not just surfacing insights, but it can also be used to monitor the status of us taking action on those things to make sure we're in a really positive stance, when it actually comes time to make these represent representations, to the buyer. But one way I wanna show you, and I know we're coming up on time here, is how these models that I've shown you can work in conjunction with those kinda ad hoc inquiries and kinda like deeper questions that we wanna ask. So, you know, let's say I've built out this model that says, these contracts require that I get written consent. Well, that's great. But I may be asked once I provide this report, well, does that take into account change of control exceptions where you need to get consent unless it's part of a merger acquisition substantial sale of all your assets? Well, you know, my again, in manual review, I'm having to go back and read all of those contracts again, or I'm having to run another model. But what I can do here is actually a combination of those things. I can filter my dashboard or just those contracts where I have an issue, and then I can ask a question like, well, actually, I've got to ask this, but, is there actually, you know, either a reasonability requirement or maybe a change of control exception? Anything that might kinda mitigate, you know, potentially that obligation. And so we can run a limited sort of ad hoc query on top of all of the structured data that we've created such that you may not even have to rerun it. You can just sort of use this, feature on the fly to be able to, you know, find this information, you know, really kinda just at your fingertips. And I think about it as we talk about mining contracts for information because it's hard. Right? You gotta go down there. You gotta sit on your face. You gotta take your hammer and, you know, break the diamond out of the mine. This becomes more like we're just kinda plucking apples off trees. You know? It takes a little bit of effort, but it's pretty easy, and it's, you know, not nearly as, you know, time consuming or as difficult to get to, you know, these types of deeper, more nuanced insights rather than just kinda painting with a broad brush, which a lot of times manual review kinda forces us to do because we have no other option. Yeah. No. That's, I think that's beautiful. I think people I mean, if you could scroll down just one more time. Folks cannot just a little bit like you just can't underestimate the power of these dashboards. Like the fact that you have risk scored data privacy, the high, medium, and low, imagine the effort it would take to manually go through your thousands of agreements, control f to the data privacy clause, and then read and evaluate it based off your own specific schema that your company has to determine is this a high risk, a medium risk, or a low risk. Right? The fact that this is when we talk about out of the box AI, we've talked about we can analyze a hundred thousand documents in under twenty four hours. This is the level of visibility you can get into when all you have is a bunch of scanned PDFs with coffee stains stored in SharePoint, and you can get to this level of dashboards within twenty four hours, that's absolutely wild. And something that if you don't believe us, please challenge us to prove it because, it gets a lot more exciting when it's not dummy data in this demo, but it's actually your own documents and your own insights and you're like, wait, we, I didn't know we had that in our insurance clauses, right, and such and so. I just, I just love to pause on this because this level of visibility, it's just, it has not been possible for so long. Absolutely. Hard to overstate. So that's just a snapshot of what we can do with these tools. As Memme mentioned, you know, for those of you on the call, I'd love to to dig in with you another time to to go a little deeper, but, hopefully, it's a good representation of just the the power of this technology and, you know, how can let you evaluate more deals as you're able to, you know, evaluate them more quickly and make sure that the ones that you do choose to engage with, you can execute on successfully. So, thank you all for, letting me be on today with you to take you through this demonstration. No worries. And I see we have, you know, a question that came in, you know, talking about the high impact this can have, you know, from a clinic in the hospital perspective. And that's one, there's so much m and a activity happening there. But two, from a procurement perspective, my god, they're buying everyone has to buy all these beds, has to buy all these pillows, has to buy all these cushions, has to buy all these drapes. Right? But what if you've acquired six hospital systems and hospital system number three is paying way less for beds? How do you get everybody to pay that amount for beds? You have to have visibility. You can't just acquire these companies and let those contracts be on the shelves. All of your hospitals are buying beds. Find the one paying the cheapest for beds, renegotiate to that contract. Find the one paying the cheapest for, you know, the little neck bibs, you know, get renegotiate that. There's so much money being left on the table by not leaning into contract intelligence in that space. So, Mark Wright, thank you so much for asking that question, and please definitely do follow-up with our last twenty seconds, folks. Tom, Mike, thank you so much for joining today. Thanks for having us. Thank you very much, Memme. Pleasure. Excellent. Well, I'm Memme on Woodyway, and if any folks have any questions about contract management and Workday, please do go to our website and follow-up. Have a great week, everyone. Cheers, all.