Name: Mitigating risk with AI-enhanced Contracting
Uploaded: 2025-12-06T11:46:02.898Z
Duration: 28 min 7 s
Description: Mitigating risk with AI-enhanced Contracting

Transcript for "Mitigating risk with AI-enhanced Contracting": We are going to jump right into our webinar today. I'm really excited to talk to you all about how Agiloft can help you identify factors of risks within your contracts. Right? Contracts are, of course, key business documents, and there are a lot of little elements in them that can slip by, you know, no matter how hawkish you might be in reviewing those documents, no matter how detailed you might get in trying to identify every little aspect, there can always be something that just slips by. And so we're gonna see a variety of tools, AI tools that Agiloft provides to help you identify that risk and ensure that everything within your contract that you want to, make sure that you mitigate or identify as a risk is, identified, tracked, and taken care of. So what we're gonna do today during our demo, we have our our little agenda up here. We're gonna see how we can quickly understand, contract risk with screens. Screens allows me to automatically analyze a contract based on a predetermined set of standards. These are, core concepts that I can use to essentially say, you know, I want the payment terms to be x. I want the governing law to be y. I wanna make sure the limit of liability is z. Right? And you can write that as a playbook ahead of time, and then screens is gonna analyze that document in order to ensure that you are meeting those standards, and that the contract you're analyzing kind of adheres to those requirements. And once we see kind of how that works, we're gonna start off with seeing how that works. We're going to then move into understanding kind of from the back end How are those playbooks actually built? How can we make it easy for you to build a playbook, whether that is built manually or generated automatically from an existing contract? So we've got some really cool technology that can help us automatically generate those playbooks, and we will see how that works as we jump into our demo here. So without further ado, I'm gonna I'm gonna go ahead and jump right into Agiloft, and we are going to see how we can analyze a document with our screens that we've built either manually or from a template. Now when I jump right in here, we can see that I've got my contract dashboard. For those of you unfamiliar with Agiloft, we have some fantastic out of the box dashboards that give you the capability to see what's going on with your contracts, see where things are at, monitor your upcoming expirations, etcetera. And when I need to, I can jump into my contract repository. And for our demo today, we have an example contract right here, kind of example contract in draft status right here at the top of our repository. And we can open up the contract document and start to review that contract and analyze it using screens.ai. Now this contract is a third party contract. Right? It's counterparty paper. In this example, we we don't have the, leisure of working on our own internal templates. We've been given the task of analyzing this counterparty paper and understanding how it is risky for us, using both, you know, our brains and using AI. So we're gonna go ahead and open up this document. And one of the great things about how how Agile works is that when I open up a contract, it does open right within Microsoft Word. So I'm gonna pop that open, and we're gonna see that right here within Word. What I love about this is that you can continue using all of the tools that you're familiar with. Right? Whether it is the, track changes feature, the comment feature. Right? All of that. We're not taking any of that away from you. You can continue to use those Microsoft Word features as you are used to. And so I might come into this document and I might, you know, review this and make some changes. I might turn on track changes again using those those features that I like and make some more changes here. But what when it comes down to it, what I really want to do is analyze this document with our AI. So I'm gonna use my add in here for screens to open up my screens AI add in and start to review and analyze this document. Now when I use screens to analyze the document, I've got two easy options that I can take. One of them is to use my own screens. Now just to kind of level set us all here, a screen is a playbook created either from your standards, your existing playbook, or your contract that gives that guidance and those requirements on what we need to do in order to analyze and make sure this contract is in compliance. So you can either create your own screen, which is, you know, fantastic. I can just select one of my own existing screens that's full of, you know, my personal or my organization's standards, or I can use a community screen. Community screens are cool because they're expert built screens that are reviewed by our legal knowledge engineering team. When I talk to a lot of people about the community screens, they have a number of questions. First, they ask, are these vetted in any way? And, yes, they are vetted by our legal knowledge engineering team. And then they they wanna understand if the screens they create automatically get added to the community, and that's not the case. Your screens don't get added to the community. So these are just some fantastic resources if you don't have a playbook today. For example, for our services agreement here, I could scroll down to the category of screens here for services agreements. And you can see we've got a variety of, community playbooks created by experts out in the industry. I like this one by Andy, services savvy procurement of professional services. And you can see that in building this, Andy has specified what it was designed for, what the purpose of this is, and what some of those limitations of the screen are. So a lot of capability to ensure that not only do we have the opportunity to analyze screens based on our own internal policies, standards, playbooks, etcetera, but experts out in the industry are providing us with great best practice starting points in order to begin our risk analysis of our contracts. And we'll actually see a little bit more later, how we could use this to actually, you know, build our build out our library of playbooks, even more thoroughly. But for now, I'm gonna go ahead and select and run Andy's screen on our contract document here. And as this screen is running on this document and analyzing it, we are going to, have three great results from the output of this screen. So the first and, you know, relating to our topic of risk today. The first is going to be a risk score. So all of the individual standards or topics that are analyzed against this contract, they're going to have risk ratings of low, medium, or high, and they're all going to either pass or fail against this contract. And I'll go into a little bit more depth on what that means in a moment. But based on their risk rating, low, medium, or high, and their pass fail, that will get aggregated into a standards score here, which is a weighted risk score of all of the things we analyzed about this contract. Second, screens can automatically ask questions about our contracts. So if you, you know, just wanna understand what are the payment terms, is it a one time purchase or something ongoing, you know, other questions like that, We can ask and have those answered. And what I love about this is when I jump into one of these, we can see what question was asked. We can see the AI answer. But most of all, you can see it jumped me to the section of the document where it found the answer. So that way we can ensure that the answer we're seeing isn't a hallucination of any type, and we can kinda double check the answer against what was found in that document. So that provides a lot of a lot of comfort for our users around the answers they're getting from that AI. But now to kinda get into the meat of everything, we wanna start to review the pass fail results of the standards that make up our, that make up our playbook. And you can see we've got kinda two results right here at the top. We have a service modifications restriction, And this one is a pass. So it says the contract shouldn't allow the vendor to change or modify services, and it passed in the AI reason because, you know, it doesn't do that, but it passes due to the absence of that text. Whereas our next one down here, our payment terms extension, and I'll filter down to my failed items, which are the ones that need, the ones that really need my attention. This payment terms extension says the contract should provide the purchaser net 45 or more to pay fees due. This failed, and the reason it failed is because the document specifies that payment is due within thirty days, which doesn't meet our requirement of forty five days or more. And, again, we can see that citation right here in the document highlighting for me, where that failure was found. And so that's really great. Right? That starts to build up, and you can see this is a medium risk standard. And that starts to build up that risk score and help me identify kind of the risky hot spots in my contract. But even more than that, this provides me a red line, an automated red line that gets generated to resolve this failed standard, as well as a Microsoft Word comment that I can apply into the document. And then to take it a step further, we can see user guidance or preferred language as well. So with a single click, right, I can actually apply that red line and comment into the document, and we'll be able to see here just like that using the normal Microsoft Word, track changes feature. You can see it's even added under my name. We've got the red line fixing our issue, bringing it up to forty five days. We've got our comment showing why we did that. And now we can move on to another, another standard. I can snooze this one so it doesn't show up in my list anymore, and we can maybe filter to a high risk standard. So I might take a look at this, you know, data privacy liability standard. And this one results in a bit more of a complex red line. I know that first one was easy. We're just striking 30, putting in 45. But here, we'll do something a little more complex. For our data privacy liability, we can see this is a high risk standard. So this failing contributed a lot to that weighted risk score. It says the contract should specify that claims related to data privacy or security are either uncapped or subject to a secondary higher cap. And this failed because there is a general cap, but there's no uncapped or secondary cap for data privacy or security. Now, again, it jumped me to the section of the document where it found that answer. And then down below, we can see a much more complex red line has been generated here with some surgical strikes and additions. And what I really like about this is that we have a, notification that this is a multi fail red line. So multiple failing standards all reference the same document text, and the system automatically creates a single red line to address all of them. And when I click into this, we can see what those multiple failed standards are. So I can see which standards are being addressed by this one red line. So multiple are being knocked out with a single click. And, again, we've got that red line. We've got the word comment added. And one thing I do wanna highlight here, which I think is really cool, is that we have this customized red line approach. So if the system generates a red line and you read through this red line and you're kinda not really pleased with it, you can actually come in and just write a prompt to say, actually, I want you to red line like this, and it'll redo the red line to match your prompt, which I think is really cool. But, again, with a single click, I'll be able to apply both that red line and that word comment into the document. And now we can see all of that has been added into the system, into our document here. The last thing I'll mention here about, fixing our failed standards is that once I kind of tune in my playbook and I start to get to the point where I feel like, yeah, this the AI is kinda nailing it. I don't need to review this every single time. Right? Then we can start to use our resolve all fails with one click button. And so for this, I just click go, and it takes every single failed standard and redlines the whole document using all of those failed standards. So that's a really cool capability that we have that, saves a lot of time once you're kind of rocking and rolling with a playbook that you feel you can rely on. Now when it comes to risk, as we mentioned, we do have that standard score. And I'll show you in a moment how all this information syncs back into Agiloft and could be used for other purposes as well. But one thing I really wanna highlight is this boosts button down here. So the boosts button has a few really cool features. What we were seeing was that people were taking, you know, chunks of a contract, copy pasting it into ChatGPT, asking it to rewrite or change something or make it mutual, and then putting it back in the document. And so we figured, you know, why not give you the capability to just do that right here in the document? And so, for example, you can redraft something, make it mutual, clarify, qualify, etcetera. But what I wanna highlight here today is this generate risk summary, boost. So with this one, I click it, and I can provide the details of which elements of risk I wanted to focus on. So, you know, I wanna give it a cons if I want it to give me a concise risk summary, and I want it to include the high and medium risk items. And then I can even provide a prompt. I can say, you know, write as an email to an exec. And I'll click generate, and it's going to take all of those failed standards that were identified, in our initial scan of the document. And it's going to write up a risk summary for me as I requested in the style of an email to an executive. So, I mean, it even wrote an email subject for me. Urgent review required key risks in the services agreement. Dear executive, I hope this finds you well. Here's all of our high risk concerns. Here's all of our medium risk concerns and, you know, a nice little summary for us here. So I think this is pretty cool, especially because I can now just take this, you know, copy it to my clipboard, paste it right into Outlook, and send that to whoever I need it need to send it to. So not only am I given the opportunity to, understand, you know, based on a standard score, the risk score, how risky a contract that it is to quickly rectify that fix, but also to generate a quick risk summary identifying those key risks. Now finally, I'm gonna jump back into Agiloft, and we'll see that all of that information about our contract has flowed right into Agiloft. So that overall risk score is living here in Agiloft, and our pass fail results are here as well. The thing that's really cool about this is that this allows me to unify the data that screens generates. Right? These pass fail results, this risk score. I can unify all of this with the rest of my Agiloft automated CLM workflow. So what that might mean is that, you know, if a contract is submitted that has a high risk score, we automatically, you know, bring in some additional approvals. We automatically notify someone from the risk mitigation team. Right? There's a number of automations that could be kicked off based on a high risk contract or, alternatively, based on a specific element here, a specific standard failing. We could automatically notify someone. Hey. Our data privacy, compliance standard failed on this contract. Please make sure we include data privacy, in the review process. So it allows you to really automate a lot of those elements throughout the process. Now I'm gonna jump back to our agenda slide here because what we just saw was how it actually works. Right? How we actually have a third party contract, how we analyze it using either our own screen or a, community screen, and how that results in risk analysis to information in various red lines. But now we wanna actually understand how those playbooks are built. And it's it's actually a really cool process. There's a few easy ways to do it. So we're gonna jump over to our screens web interface, and we're gonna cover a few of the, really cool and easy ways to build a playbook. First, you can build a playbook from scratch. So this is really, you know, saying from scratch kinda makes it sound like it's gonna be a lot of work, but, actually, it's really easy because all you have to do is type things in plain English. Right? You don't have to train an AI. You don't have to do anything crazy like that. We'll say this is gonna be, you know, Austin's cybersecurity playbook, and we'll save that here. And then you can see just like we saw when we scanned the document that the results were standards and questions. And so now I can build those standards and questions right here. So I I might start with a question. I might just say, where will data be hosted? And it's gonna automatically generate a title, And I can save that question, and boom, I've got my first element here of my playbook. I might also add a standard. A standard is a simple pass bill requirement that a document should meet, and I can add a standard here and we'll say, data must be hosted in compliance with GDPR. Right? And so, again, it's gonna generate a title for me. It says GDPR data hosting. And then what I can do is I can put in my preferred language. I can put in a default comment so that when the system automatically redlines, it's gonna redline against my preferred language. And I can choose, you know, whether I want that to be a surgical or a heavy redline against that. Heavy would be essentially, you know, almost completely replacing what's in the document with my preferred language. Excuse me. And, again, with the default comment, I can have the system either just adapt that comment or put it in verbatim. And then, of course, I can define the risk level. So it's a really straightforward process of starting to build my playbook between my standards and my questions, and I can add as many standards as I need. So that's that manual approach to building a playbook. I just come in and I start to type in each additional standard. You know, we'll say governing law must be Delaware or something like that. And so it's a really, really simple process to build my playbook. But if I wanted to get a bit more sophisticated with the creation of my playbook, we'll go ahead and jump back. You see I do have two other options on how I wanna build a playbook. First, I can start with a community screen. This allows me to go to the list of, community screens here. You can see they're sorted kind of by categories, so I might jump into, you know, vendor contracts here, and there's a lot of different options that I can use. And I can start with one of these screens. You know, we'll say critical issues for software development, and it'll import that screen into my instance, and then I can go through and review and edit these as I'd like. So we think that's a really powerful approach. It really gives me the capability to take an existing community screen and pivot it to my needs. So I can come in and just, you know, edit the standard text, edit the additional details, provide my own preferred language, etcetera. But I think one of the coolest ways to create a screen is to generate it from a contract. So what this allows me to do is to take my template, right, my preferred template, my standard template that I that I always use when I have the the opportunity to. And I'll go ahead and make this a little bit bigger for everybody so we can see it clearly. But I take my own template, and I put double brackets around the key contract language that I wanna focus on. So you can see in this little screenshot, you know, if I have a clause that's really important and I want to create a standard from this clause, I can just put double brackets around it, and the system will identify that and use it to create a standard. So I actually have a document here that I've kind of done this with as an example so you can all see what it looks like. So here is another similar services agreement. And what I would just do is we can see in this fee section, we've got, you know, services will be billed monthly and payment will be due within thirty days. Right? So that kinda lays down the payment terms there. So I can actually put double brackets around that section, and the system is going to see that as a standard when I upload this document. And you can see that in this document, I've just started to put double brackets around other key items. And so I can just go through my out of the box template and start to bracket some of the key elements. And when I upload that document into screens, it's going to use that to create an automated playbook. So here is the playbook, and you can see I I just generated this today. Here is that playbook that was generated from that document I was just showing you. So we can see we've got monthly billing in net 30, termination for breach without cause. So all of these different standards were created based on me just bracketing certain sections of the document. And what I really like about this is that not only does it create the standard, it also takes that bracketed language and puts it in as your preferred language. So that can help you drive that redlining approach. So it it makes it really quick and easy to create a new playbook from your document. Now, one last interesting question I wanna address here that I get a lot is people ask me, they say, oh, can we give it, you know, 5,000 successfully negotiated contracts and have it generate the playbook from all of those contracts? And, you know, we don't support that feature, but I think there there's a really good reason that we don't support it, which is that if you look across your last 5,000 contracts, you might find that you agreed to a spectrum of negotiated positions, right, from something that it aligns with your preferred position all the way to something that you kind of, you know, agreed to because you had to, and it's not really your preferred position and everything in between. And so if we were to use a broad scope of documents like that to create a playbook, the playbook would probably result in the average of all that, and you would end up with a playbook that's essentially like a middle of the road position. When I think what we hear from a lot of people is that they want their playbook to be their preferred position, and they wanna negotiate towards their preferred position. So we think it's much easier both to gather the documents and to generate a playbook that makes sense and contains your preferred positions to use kind of a single gold standard template like we just saw bracketing an individual document to generate a playbook from a template document. So just to recap on everything that we have seen today, we started out by seeing how screens can analyze a document for risk and identify not only a risk score, but also areas of the contract that need to be redlined to be brought back into compliance with our playbook. Then we saw how all of that data became searchable and reportable when it got synced back into Agiloft, and it can also be used in our workflow to ensure that the right people are brought in when a high risk contract is identified. And then finally, we saw how easy it is to build our own playbooks and identify those risk levels from right within screens. As we saw, there's three easy ways to do that, to build a playbook from scratch by typing, you know, plain English in into the system, by starting from a community screen and pivoting to make that more specific to our needs, and finally, by generating a playbook from our contract. So we can take our preferred positions that exist in our template, our template contract, and use that to generate our screen to analyze, redline, and understand the risk of contracts we are working on. So that was everything I had on my docket to show today. If there are any questions, we'd be happy to take them now. We'll kind of allow a minute or two here for any questions in the q and a box, and then we'll wrap up. Thank you all so much.