Name: Start Optimizing Agreement Lifecycles: Meet the Contract Workflow Agent
Uploaded: 2026-10-08T18:16:14.149Z
Duration: 1 h 44 s
Description: Start Optimizing Agreement Lifecycles: Meet the Contract Workflow Agent

Transcript for "Start Optimizing Agreement Lifecycles: Meet the Contract Workflow Agent": Welcome everyone to start optimizing agreement life cycles, meet the contract workflow agent. We've been so gratified at the response that we've gotten to this webinar. We're very excited to share a lot of new product developments with all of you over the next hours. Thanks so much for joining us. Let me do a little bit of housekeeping. Today's webinar is being recorded. I'm gonna make this available to you afterwards. You can use the same link that you use to register to access the on demand webinar. So if you wanna share it with anyone else in your organization, we certainly hope you're gonna wanna do that. There's a very easy way to do so and we encourage it. You probably saw the note on the right, but we're gonna encourage you to ask questions as we go. You'll see the q and a box. Please don't hold back. Ask your questions. We'll address them in real time if we can. We'll hold some of them for the end. We'll also answer some of them in writing as we go. So please don't be shy. We are a public company. We're gonna be talking about some new product development as well as some future product development. So on that basis, here's our product statement in terms of what you should base your decisions on. And with that, I'm gonna introduce myself and invite my two colleagues up onto the stage so we can all say hi to you. My name is Hal Marcus. I'm the head of product marketing for Workday CLM and contract intelligence, formerly known as Evisort. We became part of Workday in October 2024. My background is as a lawyer. I know we have a lot of, folks from legal on this call. We have a lot of folks from procurement and various other roles around the organization. I started as a practicing litigator, became a general counsel. I negotiated hundreds of contracts. But quite some years ago, I got into legal technology, and I've been focused on that ever since at companies like, DocuSign, Thomson Reuters, LexisNexis. And, for the last couple of years, I have now been at Workday since Evisort became part of this company. Very pleased to be joined today by Adam Lafrance, who's our head of product management. Adam, say hi to the folks. Thanks, Hal. And hi, everyone. So like Hal, I am a former attorney. I was a practicing lawyer at a large corporate commercial firm. Unlike Hal, didn't go to court. I was always on the transactional side, and I lead our product team. So, you know, for me, it comes from a place of, deep empathy with our legal users and, you know, audience members here today. So I have walked a mile in your shoes. I can understand they can be uncomfortable at times, but we're here to make that journey better for you, and excited to share, you know, a little bit more about what's coming. Thanks, Adam. And you're gonna see a full demo today by, Tyler Nevins, one of our best solutions consultants. Great to have you on board, Tyler. Thanks, Adam. Ira. I'm Tyler Evans, solution consultant. I've been with Workday for about five years. I focus on our spend management pillar, which includes our product Workday CLM. So, I guess I'm the black sheep out of the group here. I'm not a former lawyer. My background really was more on an implementation side. So I do share Adam's sort of empathy for our customers and really just understanding that implementing software can be difficult. So it's been very exciting to see the evolution specifically of this product and in this space. So looking forward to showing you more in today's demo. Thanks, Tyler. I think people are relieved to have one non lawyer in the mix. Alright. On to the next. This is a very exciting time. We're so glad you joined us at this moment for what we are covering today because we have just introduced Workday CLM r two twenty six. I know that sounds like code language. It's our second big release of twenty twenty six. We actually roll out new features and capabilities on a rolling basis, but we group a lot of this... The the conversation into two big releases for the year. And this one is a doozy. This is the biggest new release in the four years that I've been with this company in terms of the overhaul of the look and feel, the Aginta capabilities, and the breadth of all the changes that are happening here. For those that are familiar with Evisort or Workday CLM, same thing, just carrying across our our historical name for those that know it under that name. For those that are familiar with it, this is gonna look very different when you first log in. We have rolled out this new interface now or are on... Are about to. Some of you have already gotten it, some of you will have it in the next few days, and that is being rolled out to all customers. So you will be seeing this new interface, this new home page, new kind of navigation, bringing Ask AI right up top, putting important information that's personalized to you and customizable by you right on your desktop. So that is something that's rolling out to everyone. But that's just the surface, I mean, quite literally. The range of new capabilities in Workday CLM are really expensive, and we're gonna run through all of that. But let's do a little bit of level setting first. What is this all about? What was the evolution of Evisort, and what is the technology we bring to the fore? Well, it's all rooted in this idea that contracts are the lifeblood of business. I think a hard thing to argue. Anything really important that happens in an organization, you hire someone, you buy something, you sell something, these all involve contracts. But contracts aren't just documents, they're data, and that's the kind of, the issue that CLMs historically did not really give ample consideration to. Contract Lifecycle Management Solutions have been around almost as long as I've been practicing, and in that time, they were focused on moving a document through the stages to get it signed. It became more and more digital, of course, over time, but it was still about getting that document through to the final end of that cycle and maybe getting it ready to come back around for the next phase of that cycle as it renews or becomes amended or becomes terminated or expires. What's left off is the fact that all the data that's in those documents, in those contracts is absolutely critical for running your business, and providing visibility into that data was the mission of what Evisort was all about. Well, we do three core things. We connect to all your contracts wherever they are. Whatever contract repositories you use, whether they're just lost in emails, whether they're in personal drives, They're scattered all across the business, and we've done lots of surveys, by the way, that validate that if you're in that situation where your contracts are all over the place, you are not alone. We hear that from almost every new customer that we talk to, and they are looking for a way to bring all those contracts together so that they can know what's in them. So we connect to all your contract repositories often with productized two way syncs so that any new contract in either location will sync with the other, And we apply our AI and OCR process to all of those contracts. We ingest them at scale, accurately. We can recognize over 270 types of contracts. We recognize, I believe, over a 140 languages in terms of identifying what the language is in those contracts. We pull all sorts of core metadata points around those contracts. Now beyond that, we then give you all the tools that you need to analyze those contracts, know everything about what's in them. So we are extracting dozens of core data points, and I'll cover that more in-depth in a couple of minutes, But we're creating a broad base of data about all of those contracts and then providing you advanced search, customizable dashboards that you can create yourself with click and drag and check boxes so that you can know everything about what's in those agreements. These dashboards are interactive. They're always current. They're lightning fast. So there's never any kind of delay in having access to all that information at your fingertips. And we also provide Ask AI, the ability to ask conversational questions. More on that in a minute as well. However, we also provide the full range of end to end CLM capabilities like I was describing before, Getting a contract as a document, if you will, through all those stages it needs to go through to get it to execution and then into a post execution cycle. All of that is in the product as well. Now that said, I'm gonna be talking about contract intelligence for a little bit. And a small subset of our customers just use us for contract intelligence. This is a term that, as far as I know, I believe, that was actually coined quite some years ago before I joined the company. Now it's become commonplace. But this idea that you can just know what's in your contracts was, at the time, relatively novel, and it's still something that we excel at. So we have two aspects of what we do, post signature contract analysis, which delivers our contract intelligence agent. And we're gonna focus on that first and then talk about some other AgenTic capabilities. So the contract intelligence agent is all about using AgenTic AI again to give you that deep visibility into your contracts. Conversational experience we introduced well over three years ago, and we've been enhancing all along. It provides highly reliable, quick, fully sourced answers to your questions. And people often don't realize how extensively they could be using this. Firstly, it goes across one, all, or any any specified subset of your contracts. That means you could be looking at a dashboard, filter down to the contracts you actually care about, and then of those contracts, ask your question, and it will automatically limit to that group. Secondly, this is not now just about extractions. This is about understanding and interpreting the data within those agreements. It can calculate dates. It can calculate amounts. It can help identify risk points. It can provide some nuanced analysis that can be immensely valuable. Now it's great to be able to ask questions of a contract at any time and then click a link and validate by going not just to the contract but even to the language that's supporting that answer. But what's even better is when you don't have to keep asking. And that is enabled by going a step further and turning your questions into custom AI models. Some might call these agents or prompts or any number of terms, but it really goes beyond that. It is about sending a fixed model and publishing it when it meets your standards of what it's going to be able to deliver. And again, far beyond extractions. This can classify information like you see in this example here. You want to get all of your data, points into one of five categories. It will make the judgment call for you and put that into the appropriate category. It will extrapolate beyond the four corners of the contract and interpret some information, interpret a date, interpret an amount, or calculate a payment amount. It can identify risk factors from a pull down based on your guidance, and it can summarize information as well in valuable ways. So what is it you're doing with all of these models? Well, you're creating structured data. You're turning the wealth of all those contracts, all those, you know, potentially millions of pages and, you know, of artful, nuanced text into structured data that you can actually work with, that teams across the enterprise can work with. As I mentioned, all of these dashboards are a great way to have all the different personas around an organization have access to exactly the data points they need at all times. However, there are data points that don't display that well in dashboards. These would be things like summaries. And here's a popular one. I know we have a lot of folks that do spend management on this webinar right now. Would you like to be able to summarize all of the financial benefits you've negotiated across all your supplier contracts in one simple sentence, even if they're coming from three or more parts of the contract? You can do that. You can, you know, give it a a designated length of how long you want that summary to be. You can put it into a different language. You could put it into the voice of a pirate if you want. You know, this is generative AI. It can do a great deal whatever makes sense for your business. And then those data points travel with the contract just like all the other metadata and extractions from those contracts. That means you can share them with other systems, use them wherever they're valuable. And increasingly, we are enabling that within the Workday financial suite because we are integrated both with Workday Finance and accounting and with Workday Spend Management source to pay systems. So increasingly, this information can be available through those connections as well. So it's a rich data layer that gets enabled. Here's an example of what that might look like for someone doing a lot of analysis of lease agreements. Here are three fields that no contract AI would automatically be programmed to identify, but ours can because of the custom AI capabilities. So now you have your lease address, property, square footage, a combination of base rent, and increases summarized in very brief language. And all of that appears in your reports, which you can export to CSV, put them in Excel, or send them off to any other product that you'd like through API. Alright. What are these models and what is changing about them? Because as I mentioned, we just had a huge new release and some big enhancements just came to those models. Well, we've always had standard models. These are dozens of common terms and clauses that we would extract across contracts. This was the genesis of Evisort, quite some years ago. These models were pre trained to a high degree of accuracy. They were not customizable. And one limitation of them, they were English only for the most part. They could sometimes extract in other languages, but not on a supported basis. However, since we introduced custom models several years ago, we have had the ability for you to take this much further and have customizable self-service models that were multilingual and could extract, calculate, classify, summarize, as I mentioned. Well, what's new with this new release? All of our models now are customizable. All of our models are multilingual. So this gives you a broader way to share all that information across entire organizations. Now what are those models? I'm gonna run through this very quickly. I wanna leave lots of time for the demos and what's new, but these are our standard models. These are the clauses and fields that will automatically be extracted and populated from all of the agreements that you bring into our system. And these are common terms as you can see across, different kinds of contracts for the most part. So this is the baseline. This is a series of custom AI models that our legal and AI experts have already created in our custom AI model library, and there are over 120 of them. And as you can see, they cover a range of much more specific use cases, like employment agreements, sales contracts, revenue contracts, customer contracts, lease agreements. There are some that are industry specific like healthcare, financial services, construction. There's insurance information which pertains more to HR type use cases, and then there are energy service contract models, data privacy models of real interest to those in IT, ESG standards, and more. So this is the current list of models in our library. That means these models are already trained to a high degree of accuracy on our datasets, ready for you to just browse through and adopt as you'd like, and you can configure them further. They are the baseline. And now all of our models are a baseline because you can configure all of them to the exact needs of your organization. Now how do you create these models or for that matter, edit them? Well, it's really very straightforward, and it is not something that needs third party services. Although we have many third party services, a big ecosystem of partners that are happy to do this work, I think you'll find it's very straightforward. You can put in a question. It'll run that question across a small set of your agreements. You go ahead and give it a thumbs up or thumbs down in effect, correct or incorrect. And if it's incorrect, if it's not the answer you want, tell us why. Fill out a box that just says, this is why it's off point. Then it does the next round, and the model learns and can even enhance itself automatically. So your very simple prompt becomes a more elaborate prompt specifically designed by the AI to get the highest possible accuracy. When you're satisfied with the level of accuracy you're getting, go ahead and publish that model. And then that will run across all the appropriate contracts. Now everything I've been talking about so far is contract intelligence. I wanna move on now very briefly before I hand it off to Adam to, talk about contract negotiation because we have many agenda capabilities here too. We have the ability to do full document review and redlining of entire contracts that come into your system against the appropriate playbook. So it will review a full contract, identify all the risk points, give it a degree of risk, and tell you why, and then recommend the red lines, specifically pinpoint red lines, not full clause swaps unless they're absolutely necessary so that you can efficiently have your negotiation move to the next level and keep a degree of compliance and consistency across your contracts. Now the biggest issue we hear with this is, you know what? That's great, but I don't have a whole lot of playbooks. They take forever to create, and I have a lot of different kinds of contracts. And that's why now we've introduced an agentic playbook builder with our new release so that you can take your notes in the format that they're in about what your guidelines are and upload a sample contract. And you put those two together, and it will automatically create a draft playbook for you with all of the guidelines, all of the the issues that you care about with the degree of severity on them, and then you can edit that as desired. So this is a great way now in a much shorter time frame to create a whole series of playbooks for the different kinds of contracts you need to negotiate. We also deliver analytics around all this so you know how much these playbooks are being used, how well, the answers are being adopted, whether they are on point or not, and you can track all of that as you go. Now one more important development here. All of these things I've just been describing, no. Well, let me go back a step. Not the playbook creation, but all of these things I've been describing about editing your, contracts, negotiating them, having access to those AI tools, all of this is now being enabled in Word itself. Because we've heard so much from customers that they want to keep editing in Word, and they wanna have all of the tools, from Workday CLM at their fingertips. So we have built all that in together into our Word connector so that you have all of those tools available to you. Alright. That is not all that is though in our new release. I'll hand it off to our head of product, Adam, to walk through the next round. Thank you, Hal. And so with that, you know, we really wanna talk about how we're bringing our AgenTA capabilities, as the slide says, across our entire contract life cycle. You know, when we think about these enhanced capabilities, we're really thinking about the life cycle holistically, and the ability to both augment and accelerate some of those tasks, some of those core processes that rely currently on on very, manual human effort, you know, high cognitive load, high, you know, transactional costs in terms of, the time and effort involved to complete it. And so with that in mind, Hal, can you take us forward to the next slide, please? Sorry. I don't have the the controls there. Thank you, sir. You know, just looping back to what Hal was describing earlier, historically, we had thought about this as kind of two sides of the coin, our... Both our presignature and our post signature, capabilities and, you know, creating the the agentic skills and capabilities in service, of those kinda two stages of the life cycle. But with the contract workflow agent, what we're really able to do is have those capabilities that span across the entirety of the life cycle. And so on the subsequent slide, I'll get into a couple, but we understand the need, for those capabilities, you know, from initial request through to, you know, execution or or signing of a contract. And equally, everything that happens, you know, post, execution in terms of obligation management, in terms of the ability to have those, you know, reminders or escalations when a contract is coming to close and you need to decide, you know, are we going to renew? Are we going to amend? Are we potentially... You know, what actions do we need to take to terminate? So if we can go forward, I just wanna double in, or double click rather on a couple of those key skills. The first one that I'm I'm really excited about is around, the ability to build those workflows agentically in the first place. So historically, you know, for our larger enterprise customers or even some of our smaller ones, you know, the ability to define your business process as a workflow, involved a lot of time, a lot of effort. This this could take, you know, days to get those configurations correct. And it put a lot of effort, on, you know, the individual folks, within the company to really set up those configurations. With these new agentic capabilities, you can provide a simple prompt, you know, basically, a a high level outline of of your business process, upload some template documents, and generate that first draft of your workflow in mere minutes. What's more, it's not a one shot, you know, engagement with, the workflow builder. You're able to have a conversation, continue to refine, add assignments, add additional tasks, or review, phases as you go. So it really is, a living workflow and one you are able to to adapt and refine, to meet the needs of your organization. So we have a couple of, posts, both on LinkedIn and and some of our marketing blog that that showcase this functionality in more detail. Tyler's gonna give it a quick run later, but I'm truly excited, about this functionality and, the customer response we've seen, thus far. Building on that, the next thing is within the, the new request. So the kinda agentic intake capabilities, to both enable your business users, the kinda end requesters, to simply, you know, provide, a natural language, question or, you know, clear prompt about the type of workflow, the type of contract request they're looking to initiate and be automatically routed to the correct one. So this greatly streamlines that process. It reduces, you know, many of those asks that may be directed towards, you know, legal audiences or procurement for that matter, because your end users are able to self serve in a much more straightforward and intuitive manner. And we're actually, continuing to, enhance that that intake functionality, the Agenta capabilities and and support thereof. So if you were to provide... You know, let's say you're on the procurement side. You're, buying goods or services, and you're doing so on third party paper. You know, you're you're uploading a vendor's template to have the, agent effectively review and extract key data from that to populate that intake form, to to validate, you know, the key information that's required to generate that new request and to do it automatically on your behalf. Just doubling down on one of the, questions that was asked in the chat earlier, you know, our core philosophy is that the human in the loop is central to the experience. So the AI will always give you full visibility both into what it extracted and where in the document it found. You know, much like my legal professors, Ask AI will always cite its sources. So what it will do is it will always provide you with that visibility to increase trust and confidence in the accuracy of the data it's providing, but ultimately surface it back to the human, in the loop to to to be able to make those informed decisions and and move the business process forward. And then if we go on, what is also, in active development and we're working towards is one of the Agennta capabilities where I think Workday is uniquely positioned, to deliver just based on, the broader organizational context and and the data that we have about your organization. So one of the things that we can, lean into is data that already exists within HCM. So within our human capital management product offering, this can include things like, direct reporting lines, managers, cost centers, so that from a review and approval standpoint, we can leverage that data to have the agent, both provide reminders when, tasks, are nearing their their deadline, but ultimately escalations. Somebody goes on vacation, we can leverage that data to be able to escalate, to a manager or a peer who needs to take that on to ensure that contracts, you know, in flight do not get bogged down in bottleneck because deadlines are being missed. Okay. So with that, I wanna take a step back. You know, I know we... We've talked about a lot of the the functionality that we have, you know, in isolation. But if we really think about this, holistically, we're showing many core agentic skills, across the entire contract life cycle, and really trying to meet our users where they work at every stage of the process. So this is, you know, some of the key highlights that that Hal and I, wanted to emphasize with you today, but it's by no means the whole set. So if you have questions, please continue to drop them in the q and a, and we're happy to take it from there. And with that, Hal, I'd like to pass it back to you. Thanks, Adam. And, yes, we've been getting a lot of questions. We've been answering them in writing. We will start to address some of them at the end in in, you know, so that everyone can hear, but I encourage you to keep the questions coming in. Just one more thing, I wanna call out. You know, it's one thing to hear from us the experience we're having, but Adam and his team work very closely with customers on all the developments and all the priorities that they have. And so it's great to get their early feedback on the new development that we've been doing in this particular... This was an an unsolicited email that came from one of our customers, and we were just so delighted to see it. And they said, yes. You can go ahead and share it. So it'll give you a sense of the kind of response that we're getting, and we hope to see a great deal more of this as we roll it out now to a much broader customer base in the coming weeks and months. I wanna call out one last thing before I hand it off to Tyler to take us into a demo, and then we follow that up with q and a. And that is if you're finding this webinar valuable so far, and I hope you are, then I wanna encourage you to check out our meeting of the minds legal AI podcast, where every month, my colleague, Meme and Woody Wei and I interview customers just like the one, that we we cited on the previous slide and other thought leaders within the legal AI space. And we talk about, what's going on in terms of the most... You know, the latest, most proactive uses of agentic AI in this space. I hope you find that valuable. There's a barcode there for you. And with that, let's hand this off to Tyler and let you see all of this in action. Tyler? Great. Thanks, Hal, and thanks, Adam. There's gonna be a lot of ground to cover here, so I'll go ahead and start sharing my screen, and we'll jump right into it. I'm sure this is going to spur quite a few more questions than we've already seen so far. Alright. So right now, I am on my home page, and this is actually, a part of this overall new user experience prior for any customers that are currently using our CLM tool, and we're happy to have you doing so. You're probably more familiar with the landing dashboard, the documents page there, which of course is great for operational reasons. But, for the average end user and especially our stakeholders that are not engaging with a contract on a day to day basis, a simple home page that is going to prompt them to drop in their questions about their contracts, prompt them about recent activity, or the work that's required of them is just much simpler and more effective for that end user. So we'll talk a little bit more about this. One area I'll just highlight is these various reports or search options that you can quickly run. This is just going to allow the user to quickly access a subset of contracts for them or take them into the search engine and capabilities to start filtering down even further beyond that. For this demo, we really wanna focus on some of those new agentic elements that were covered a little bit earlier in the slide. So we're gonna start with that concept of the workflow builder. Now from a legacy perspective or a legacy solution, we're thinking about workflow, in the sense of process and having to consider our different conditions, different individuals that are being brought in. So this can be a pretty time extensive activity in general. Leveraging AI is, of course, just gonna allow us to expedite that process, meaning we can develop more workflows. We can make sure we have a very specified process for these different potential branches that we might wanna take a user down. So when we go to create our workflow, we have a couple options. Of course, we wanna be able to create with AI wherever possible. We still can start from scratch. We can clone existing ones. And here you can see that I have a couple options. I can start with some instructions and I can also just upload a document as my starting point, which is gonna help by extracting the different fields from that agreement. So I can leverage that in my conditional logic and I can really start to expand on what that process will look like. So let's go ahead and upload a couple items here. Go ahead and grab, two documents. This is actually just gonna be our Workday NDA and our UMSA. And I'm gonna go ahead and drop in an additional prompt here that's really just going to say that we're designing a multi document workflow. So it's not just for one contract. We wanna have the NDA and the UMSA go out as a part of this process. We probably want an intake stage. We're gonna have our stakeholders provide just a couple initial details for us, and then we're gonna wanna take this through a couple review rounds. So if we define those basic instructions, we can have this created. And then from there, we can continue to layer on AI still as we might wanna make tweaks or changes to that prompt in addition, of course, to any of our updates that we want to make ourselves. Now this will flow through. It takes usually just a couple minutes for this. So I'm actually just going to open up another one that I sort of preset for us here. And in this way, as we open up the the workflow itself, there is this new additional value that you can really define these stages. So prior to this release, we had a fixed stage where you go through a draft, you go through your review, go to signature and execution, which did cover the common workflow requirements there. But of course, every contract has some slight variations to that. You might wanna go through reviews in different rounds. You might wanna engage with the end user as part of an intake, gathering new data as you move through these different stages. So based on my prompt, it would create for me this starting intake and then our subsequent reviews. And then lastly, we would have our signature execution. I'm gonna go ahead and delete this financial review stage to start because I wanna show for you the builder AI. So this is the additional layer of AI that we can continue to leverage to make any of those manipulations or those modifications to this. So if I wanted to say, let's build or add an additional financial review and assign it to a specific individual. I can drop that into the tool and have it continue to manipulate the existing workflow here. Now, of course, within this overall workflow, we still have the ability for us to designate or prepopulate the way we would want signers to go. We can start to work and create our conditions. We can review the document templates that have been created and see how those different fields have mapped into our various forms. Our builder AI completed the elements there that we had. It added the financial review into my overall compliance and risk review phase here. I could define it to create a separate phase if I wanted to do so in more of a serial fashion. In this case here, I wanted this parallel review to drop into my document... Or excuse me, into my review flow. Again, I just opened up this document so that we can see our document that uploaded and the various fields that were extracted from our document and then maps to an appropriate field value so that we can quickly leverage these in our conditional framework and start to build that out. Now in combination with our workflow, we need to have that playbook. So the two that work together in the sense that workflow's gonna define our step by step process for our end user for that document, our playbook is gonna be our sort of rules of the road, our basic guidelines that we want AI to leverage and utilize when we are looking at the document that is going through that draft cycle. So for playbooks, we also have our builder available. And this is, of course, more of an administrative layer for the individual to use. So we can create our playbook either manually or using AI. And at this stage, we can either upload an existing playbook document. We can also leverage contract templates as a way to just give more context for AI to determine what the different negotiation recommendation should be. And when it goes through its red line optioning, it has some reference material to leverage. Again, initiate or upload some documents here. Let's take an MSA negotiation playbook. We'll generate this playbook, and this will initiate the agents to review through the documents, to extract that information, and to start generating or creating the framework for those recommendations. Now this can take a couple of minutes. You all don't wanna just sit here and see this load, but this can run-in the background as well. So while that is running in the background, we're gonna go ahead and just look at what an executed playbook would look like. So if I open up this mutual NDA, you can see that there are 19 total requirements that I would look through. There is a way for you to define the definition or the criteria of it. So low, medium, high priority should focus on this, should always identify this language within this document type. And when it does identify that, leverage the different review criteria. So we can look at the reasoning that is going to be behind the scenes for this as it comes to give us those feedbacks for redlining. And then for any of those medium, low, or no priority items, it'll also understand that variation in its priority scaling. So with that in mind, let's go ahead and look at the example of both of these being leveraged together in an active drafting of a contract And to the intake portion of our AI functionality here being there to help drive the user to the appropriate workflow. Under our ask AI layer here, we have our workflow routing. So I can initiate workflow routing from here. Of course, I could just browse these, but then I can also start to just drop in a natural language prompt. So let's let's take the example of a stakeholder that is working with a new counterparty, could be a vendor, could be a customer. Maybe we haven't onboarded and we haven't fully collected their information. So we know likely we need to initiate an NDA as a part of this starting process before we move into, contracting on the various terms. So I could say something like I need to create an NDA. Pretty straightforward, but we might have different NDA processes. So we might also drop in other context or information related to the type of NDA that we have, the type of relationship that we're working with. And based on either the workflow name, the description, as well as some routing hints, AI is gonna use that as the guidance to help address or identify what process the user should be following here. So I noticed in my environment here, I'm gonna go ahead and follow this NDA workflow. This is tied to one of those different playbooks that are already created and published. So click into our workflow. And as you saw in the workflow flow builder, there is that intake optionality or element for that. So this is where we are determining for the end user, are there any fields or information that you should provide upfront that might manipulate our workflow with different reviewers? It might change the exhibits or the language within the material document itself. And so these fields are really driven from a workflow perspective. On the back half of executing this agreement, we're gonna be leveraging AI to extract the data and store it at the record level. So let's go ahead and populate some values here. I'm based out of California, so I'm just going to be following that. Oh, actually, I think I selected the wrong wrong NDA process. So let's let's go ahead and start that over. Here we go. I'm gonna go ahead and do a mutual NDA. These are just some demo values here. Of course, your fields will be configured to the purpose or the, information that you need to gather. Required fields. Go ahead and populate that. We can continue to fill out this information. Optional fields versus required fields. That's something that you get to dictate, of course. Get nice validation prompt here that I need to choose which contract I'll be leveraging. I could select multiple here for a multi document scenario or I could look through the different template options that I want to leverage. So I'll go ahead and use this, consolidated one and we will go ahead and create this request. Now just calling out here some of the different tag elements. This will populate the information that I also provided upfront in my form based on those different fields that were extracted from the document as we set up the workflow. So just making it easy for the user to pre populate some information into our documents. Alright. Now that we've created that, we have our overall workflow defined for us above here. We have our various tasks that are pending with the appropriate user engaged for that. We can see our upcoming phases and the work that we need to complete before this document would be considered as executed. Along the right hand side here, we have a request summary. Something very helpful for say, an approver, an executive coming in trying to understand what this purpose of the document is, where we are in progress with this request. So it summarizes for us where we are in process here. And then, of course, we can start to jump into this document to make our changes or execution to that. So let's go ahead and do that next. We have our built in connection to Microsoft Word to make our changes as well as our AI features that you're gonna see as an add on to these different functions here in Word. So I'll go ahead and just open my Word application. This might open up on my second screen, so you'll just have to give me a moment to drag it in if it does. Looks like we're all good here. And along the top right hand side, I have my Workday CLM add on. So I can open this, expand this screen here so you all can see a little bit better. Let me just make sure I select the right one. And I can start to leverage my AskAI functionality. Now this, of course, is available in the platform so that we could look at drafted documents. We can also look at any of our executed agreements to search, filter, identify really from a reporting perspective or just ask any ad hoc queries of an individual document so we understand maybe what we're buying or what services we expect or there's any sort of breach language we need to be aware of. But we also have it available to us here in the Word application. As I go through these different optionality, on the right here, we have editing options. We have the ability to look at the clause language that's identified in the document, and then the playbooks are gonna be able to be analyzed against the document we're working on along with, of course, some of the more traditional elements of working on a contract like comments, the back and forth, notifications for that, and track changes in history. So to run our review, we can simply select the playbook. Now keep in mind, playbook should be focused on a particular workflow or contract type. We wouldn't wanna be running our MSA playbook against our NDAs. So every workflow has the associated playbooks towards that. In this case here, because we're running a mutual NDA, let's go ahead and run this process. And now while this takes a couple minutes, I'll just highlight again for you these other elements. So Ask AI, as I called out, you do have some, more generic AI editing capabilities where you can run a couple quick prompts against this polishing a theme of the document, flagging some key risks, checking for consistency. And then as I mentioned, comments which will then show both the document comments, something that would be exposed to the counterparty as well as any internal comments that we have for our team members to work back and forth with on this document. While I give this a minute, I can also just take some time to go through any of the q and a that I'm seeing here, maybe answer some of these live. So, Brianna, I see your question there along the sample playbooks that we could download. We don't have sample playbooks in the sense of, we would be giving you advice on how to go through and negotiate your NDA. That would certainly be something that your legal department would want to develop. But the purpose of that playbook builder is, of course, to make that easier. So, of course, in a legacy scenario situation, that's a pretty timing extensive activity as well, to go through each of your contract types and determine what recommendations should be made for that. But the ability to just simply import documentation and then have AI extract that information makes it much faster to build out a variety of different playbooks. Yeah. I just wanna underscore... this process... Oh, Oh, excuse me, go ahead. I. just wanna jump up here to to underscore interesting question. And I saw a couple of other questions come in that were sort of related to this. So what I wanna call out is some future development that is happening right now, and we hope to actually have out even before end of year potentially, so not far off, where playbook creation can just be a function of existing contracts, where we can suggest to you what we think is an appropriate playbook based on a series of contracts that you upload even without any notes or guidance. That's something we're working on now. It's certainly subject to the product statement at the beginning of this, presentation, but we're very excited on this development. We think it's gonna be very helpful. Thanks, Hal. Yeah. Appreciate that. So our AI review has completed. We can see that it breaks out into some nice easy user experience identification of what action is required of me. And so I'll have this human in the loop. We, of course, want to ensure that our subject matter experts are still engaged with the material that we're working on. They're applying their strategic logic and knowledge and understanding of this process alongside the recommendations that we have provided using AI. So working with both, both parties here in that sense to look at the suggested red lines, review the reasoning, and make the determination. Yes. Are we gonna move forward with or no. Am I not going to? Or maybe somewhere in the middle. Yes. We do need to redline this. I'm gonna make an adjustment, from what has been written and then modify this suggested redline before I move forward with that. We can also confirm against the no issues found. So understanding that there's language in this that looked against our guidance and got a thumbs up or a score of past, which is identified here and confirming that, yes, this is true. This does meet the the gold standard or the terms that we have written into this and we're ready to move forward with that. Now Now when we're happy with the materials here, we wanna make the the final steps to this. Of course, we can edit and apply those selected red lines. We can still just confirm the location of where to apply this. So, again, really focusing on the navigation or the organization of the document document as well. Not something that AI is always going to be able to do 100%. So it needs that person in the loop to determine where it should be populated. Now, as you make any of your changes, of course, to the document, you go through your multiple internal reviews. These all sync back into the platform so that you have all those changes readily available here. If you're working with, a supplier relationship, we do have a supplier portal so you can send this across to your counterparty through that so that they can manipulate the document and you go through those multiple rounds. But ultimately, what we're getting to is the signature and the execution of the document, which would lead us back into more of our contract intelligence layer. So that's everything to do with extracting the information from the contract, making it available to you in a structured data format so that you can report on it, so you can drop it into a dashboard, you can do a variety of visualizations for this, and then you can use this to filter and manage your contract over its life cycle. So understanding that your contracts will come up for expiration, having that information up to date one hundred and eighty days before, you can determine is this counterparty still upholding, their contractual agreements? Do we take them through some sort of corrective action plan? Do we wanna go back out to market? If it's a customer, do we need to understand are there changes in the pricing and the revenue? Is it our opportunity to go back to them and see if we can sell them additional value and bundle that into their contract as part of that real effort? So lots of different activities, of course, that come out of understanding the details within our contracts, making this more readily available to our end users so they don't need to go to those SMEs in procurement, you know, in your different revenue teams just to understand the language within their actual contract. So for this instance here, let's go ahead and run a prompt in Ask AI. Let's say, that I want to maybe see all of my contracts with a particular counterparty. So let's say with the Bank of New York. And Ask AI is really great at identifying language within a document and pulling back a, say, list of documents. But the format in which I, as an end user, would see this information and have to scroll through, say, thousands of documents wouldn't be very beneficial. So instead of displaying all that information, for you in this sidebar here, When it exceeds a certain number of records, we wanna actually take you into our search engines so that you can continue to layer on additional filters or additional search here. So we've exceeded what we wanna see and you can see how as I scroll through these 20 records, it already starts to get a little cluttered. So I can jump right into my search engine and look at the total 42 results that came back based on my search prompt. That prompt is also then just built into the filter, the identified filter, and I can start to already see a summary of these various agreements or this group of documents. So from here, I could add additional layers of filters to drive down into the document that I'm truly looking for. Just doing a quick time check here. How, I can continue to go on with some additional items. I know there were some questions just related to the identification updated within a document. So if I actually open up this, I'll hit on that, but otherwise, feel free to come back on if we wanna start addressing any other questions. This is really just going to be an individual record, but here we can see that we have some of the field values that are being extracted from a document. From an end user perspective, we can validate this information using a magnifying glass. And if we ever needed to modify the field that was extracted, we can still do so. So human in the loop again, ability to control for any potential errors, although our out of the box models come highly pretrained with a high degree of accuracy already. Alright. Thank you, Tyler. Anna, I just came back on stage so we can address more of the questions that we've been answering mostly in writing directly back to the people that have been putting them in. We're definitely seeing a few themes emerge, and so there's probably a few good things to call out that I think will be beneficial for everyone with our remaining time. One... Oh, two more just popped up literally as I was saying that. One that I want to, to bring up. There have been a couple of questions about accuracy of the models. I shared before how we have standard AI models that do core extractions across of common terms and clauses across different kinds of contracts. And then we have a custom AI models that you create yourself or that you start with our custom AI model library where we have a 120 plus pretrained models. You can use them as is or you can customize them as you go. And just to recap, everyone now has full ability to customize all of our models. That's change that we've made with this r two release that I think is great. So what does it mean in terms of accuracy? Well, we typically train our models to 95% or better accuracy just as a baseline, but that's what it is. It's a baseline because you have the ability to take it further, apply it to your own datasets, run a few tests, and refine the models further with our AI even helping you prompt more effectively as you go. So you decide when that model is at a level of accuracy on your datasets that you want to publish. So you have a lot of control over what these models are doing and the level of accuracy in them. Adam, did you wanna add anything there? No. I think that's a a great summary, Hal. Actually, I take that back. The one thing I would add, with both our standard and custom AI models. So Hal gave you a great example earlier, around contract type and that being trained, to identify over 270, discrete contract types. We are continuing to enhance so that you will always have the ability, to refine those models. So even our standard ones can be customizable, depending on the needs of your organization. We, the way I often describe it to folks is, we give you the floor, you set the ceiling. Right? And so with something like, a contract type, there may be ones that are specific or unique to your industry, to your company. So you will still have that added degree of configurability and customization, but these are ones we want to provide, to all our customers without the need, to create them, them yourself. However, you still retain that ability to customize, to improve the accuracy, as needed to meet the needs of your organization. We've had a few other questions also, and this is a kind of question we love getting. It's sort of like when people ask about pricing. That means you're very interested, obviously, in what we're offering here. And this particular series of questions from different people were about using us in conjunction with or to replace other CLMs. So let me try to answer that in a in a broad enough way that might be helpful. We have been implemented, in conjunction with various other CLMs over the last number of years, sometimes by partners. Again, we have a vast partner ecosystem, you know, major, you know, providers that can help you with all sorts of specialized needs. We also do some implementations ourselves, and we have been implemented with, you know, various other CLMs as the contracts in... Contract intelligence layer to augment what they have to offer. So that is possible and has been done. Now in addition, however, we're getting some questions about, what if I move all... Everything out of this other product? Will you recognize all the data, the workflows, etcetera? That's a little more nuanced answer. It depends on the product. It depends on what information you're bringing in. And some of what we'd be bringing in, I think you would probably find you wanna replicate anew, like workflows, which are now, as you saw today, much, much faster to create, much easier to create than ever before. Adam, anything you wanna add about how we recognize data from other systems, including things like groupings, parent child relationships, etcetera? Sure. I I think the first thing I would offer is that, you know, we understand that, both our software as well as, you know, your workflows do not operate in a vacuum. So from, the the data configurability, the extensibility, through our public APIs, through those integrations, you know, our guiding principle is to truly, you know, meet users where they work, but equally to get the data where it needs to go. And that can be both upstream in terms of the ability to initiate requests through things like Salesforce as an example, but equally where that extracted contract data needs to pass through. So in addition to what we support, in terms of native integrations, there are also extensive API, capabilities, to build your own custom or or bespoke ones, to meet the needs of your organization. There are a couple of questions that we left open in the chat, not because we don't wanna answer them, but because I was hoping to get a little bit more clarity to make sure we can answer them as accurately as possible. So without, you know, calling on folks, to come off mute or, maybe we can add some some clarifying, questions. But, Loretta, I'd seen a couple around, you know, the approval process, and I think this ties in with, Allison's question as well. And so, Tyler, if you're able to just take us to an existing. request, So this would be an open request. One of the things I wanna emphasize, Allison, is that the view, you know, the... What a user would see on this given stage, or on a given contract request, can look different, can be tailored to either the individual requester, you know, the... Who submitted or initiated that contract request in the first place versus those that are approving. And so one of the things, if you, Tyler, just go over on the the lock icon on the the right side bar there. So there should be, sorry. Just one down. There's the ability to configure the, the access controls. Oh, we're we're looking at it. Yep. Sorry. Permissions. My apologies. It's a little bit small. But here, you'd be able to, essentially, Allison set different views, for different folks. So that could be by role. That could be by, you know, individual department. So how it looks for, folks within, say, legal, or finance, procurement for that matter. So you would have that ability to customize. And, again, it could be view. It could be full edit access. You do have that configurability. The other thing that I wanna emphasize is that when you come into a given request, it will show... It'll provide visibility into all the required tasks and approvals, in order to complete this contract. But what we're showing at the very top would be that any approver who comes in will be able to see my tasks front and center. So we're always gonna prioritize, that kinda user centered experience where any individual reviewer would see immediately, what is required, for them to action in addition to kinda having that holistic overview of, all the broader tasks involved. Okay. Loretta, can you just clarify what you mean by a DOA? Because I'm pretty sure it's not dead on arrival, but I wanna make sure I answer your question as appropriately as possible. And feel free to just drop it in the q and a, and I'll I'll make sure. Just looking back through. Ah, delegation of authority. Thank you, Loretta. So one of the things, that we're able to do is is provide that. I realize we are coming up on time. So maybe, Loretta, we can follow-up with you offline. I do wanna make sure you get the full answer, but I just wanna give everyone a quick thank you. Tyler as well, thank you for walking folks through. I think the level of questions we're getting is indicative of the incitement, of the excitement rather, that we have. And so thank you all for for being here, for engaging with us today, and look forward to more to come. Thanks,