Welcome and Introductions
Kate Bridle: Hello everyone and welcome to another IHC event. I am Kate Bridle, your IHC events host, and I am so thrilled to have you with us today. Special thanks to our wonderful sponsor, Pramata, thanks to whom this session is free for all of you. I do want to be clear, it is not a CLE session, but we’re still going to learn a lot of great stuff about AI contract intelligence for in-house teams, which is going to be very interesting and important. Thank you to our fantastic presenters today, James Chou and Justin Schweisberger, who are going to talk to us about this important topic.
James Chou: Thanks Kate. Welcome everyone to this webinar. We are very excited to present this topic to you, and we hope that by the end of it you will actually have a practical framework that holds up in the boardroom. My name is James Chou. I am a contract management solutions expert here at Pramata. In my past life, I negotiated contracts with different research institutions and then moved to implementing CLM solutions, which led me to Pramata where I sit on the product team. I work very closely with our customers and prospects to try to make our product better day by day.
Justin Schweisberger: Thanks James. Thanks everybody for joining. My name is Justin Schweisberger. I’m Chief Revenue Officer here at Pramata. I’m coming up on my 18th year with the company. I was one of the very first employees in the early days. The first 12 of my 18 years I was in ops and deployment, doing contract intelligence projects and working with large Fortune 500 clients. This is a really important topic because we’re seeing more and more companies say, I really need to understand more about my contracts. We’ve made investments in CLM, or maybe we haven’t. How can I really translate that into value that hits at the C-level?
About Pramata
Justin Schweisberger: We’ve been around for a while and we work with some of the largest Fortune 500 clients. We offer end-to-end CLM, which includes request to approve, and we have AI redlining tools, which I know is a big focus area for a lot of people right now. The types of companies we work with have large, complex commercial relationships with their vendors, partners, and customers, and this leads to a lot of challenges.
Why do companies choose Pramata? First, we really emphasize elevating legal’s strategic profile. We solve problems that frankly used to be thought of as unsolvable, whether it’s understanding your data rights and obligations across hundreds of thousands of contracts, or responding to tariffs quickly. Things that can actually make a huge impact in the organization. Second, we have a very flexible deployment model. A lot of people have spent a lot of time and effort deploying a CLM and feel like they’re in good shape with the request and approve workflow. We have those capabilities, but a lot of people come to us saying, I don’t want to rip that out. Can you come in and complement that investment and make it better?
The C-Suite Challenge: Why Contract Intelligence Matters Now
Justin Schweisberger: If you take a step back and look at the last five or six years, it has been a once-in-a-lifetime series of events crammed into a very short window. Think about how many of those resulted in people having to go understand their contracts. Can we adjust prices based on inflation? Who owns our data? What rights do we have with our client data? Can we roll out AI products? Now there are new tariffs. Where are we exposed? Can vendors pass those costs through to us?
Add on top of that the normal day-to-day things that are strategically important, like M&A integration and diligence. How can we combine two companies, look at vendor overlap, customer overlap, contract risk? These are real, challenging questions that a lot of companies face. Either they’re unanswerable, or the answers are really slow, labor-intensive, and frankly a headache to get at.
When we engage clients on this challenge, one of the most important things you can take away is this: even before any evaluation framework, what is the problem you are trying to solve? What is the most important thing? Is there a C-level challenge? Can I hook this initiative into something that will ensure it gets funded, showcase the value, and focus the evaluation efforts? Because if you show up and say, I need to extract data from contracts, there are a bunch of ways you can do that. How do you figure out the right one for your specific problem and your specific business?
The Framework: You Don’t Have to Tear Down the House
Justin Schweisberger: Before we jump into the pillars, there is a really important point I want to hit on. We look at this like building a house. Let’s say you need contract intelligence and it’s not there. You have a house that was great for two people, but now it feels too small. You face the choice of adding on to your current house, which you love, or going to buy a new place and starting from scratch. One of the big takeaways here is you don’t have to tear down the house. There are really good opportunities to add on to a strong foundation that is already in place. If there isn’t one in place, we can speak to that too. The question is: how can I make the biggest impact as quickly as possible?
With C-level visibility comes C-level pressure, expectations, timeframes, and the need for agility. You could invest a lot of time, energy, and even select a vendor, but if you don’t do a thorough evaluation and build a really strong foundation, you could end up in a very tough spot. We always say let’s showcase what things look like, not with a generic set of contracts, but with yours. Pick things that are challenging. Most C-level problems involve your customer or vendor contracts, and those are complex. Pick a gnarly one or two that has 10 amendments and a couple of MSAs from the same counterparty. Let’s make sure you can test the quality of that. Not just, is it right generally, but is it right for my business? Does it reflect how we look at the world?
That is really the purpose of today, and why we put these four pillars in place: to help people work through the process in a rigorous, focused way.
Pillar 1: Foundation
James Chou: Here we are going to go into each of the four pillars. By the end of it, folks should have a comprehensive yet practical framework they can reference whenever they are evaluating contract intelligence solutions.
You can’t build a dream home on a cracked foundation. I live in Florida, and we often deal with sinkholes and swampy areas. Imagine trying to build a house on that. The same paradigm applies to setting up a contract intelligence solution. If you’re trying to get insights, extract meaningful data, and solve C-level challenges, you don’t want to do that on a platform where your data is unclean. You still have everything in a mess and you don’t know what’s in your contracts.
This is why the foundation is Pillar 1. You really need to get it right to be able to benefit from additional intelligence and take advantage of generative AI on top of your contracts.
There are three main buckets to think about when building the foundation:
Contract organization. How do you automatically identify and eliminate duplicate contracts? In my past life, one of my first jobs on a legal team was to go through a filing cabinet full of contracts and put them in a spreadsheet. There were duplicates everywhere because people used to make manual copies. My job was to figure out what the duplicates were and which files were only partially signed. You obviously want to deal only with fully executed contracts, because those are what drive your contract intelligence. That is a key question to ask any vendor: how do they handle this?
Contract hierarchies and relationships. Can the system automatically handle contract hierarchies? Before, you might have had all your files in different folder structures with no easy way to visually see, okay, this is a parent MSA and these are all the child order forms and amendments. As you do business with a company or client over many years, you accumulate a lot of contracts, and trying to piece that information together manually becomes very challenging as your organization scales.
Setup effort. If a contractor came to you and said, I’ll give you a shovel, but you need to pour the concrete and level the land yourself, that seems very counterintuitive. The same applies here: what parts of the process are automatic versus what parts require manual effort from your team? In the age of AI, you can partner with AI, but we are not at the point where it can do 100% of the work. It is really about making sure you leverage the technology available while keeping your team involved to ensure the data is highly accurate and the foundation is strong.
The outcome is a single source of truth. Your team is no longer scattering through different sources, going through drives, and checking people’s desktops. You have one place where you can derive all your information and allow everyone beyond the legal team to access it.
Demo: Building a Foundation in Pramata
When working with any new client, we tell them: point us to your files. We don’t ask you to organize things or put together a spreadsheet. You point us to where your files are, even if they are in multiple locations, and we have methods to ingest them from different sources.
We run everything through our contract AI knowledge engine. The AI does a very good job ingesting all the documents. We OCR every document so that a PDF version is available for each file. One of the important things the AI is doing is filtering out the noise: duplicates, partial agreements, even files that are not contracts. By the end of it, all you have done as a client is pointed us to where your files are. We do all the heavy lifting to provide you with a subset of fully executed contracts that your organization actually does business off of.
We also have the AI do all the attribute and metadata extraction, and we use that information to start building out your accounts and parent-child relationships. We are not at the point where AI can do 100% of the work. There will be documents with issues. If something has been copied many times and the image quality is so bad that even the machine can’t read it, there will be exceptions. That is where our human-in-the-loop system comes in. We have a team of legal professionals who handle those exceptions and work with the client to rectify any potential issues with documents the AI was not entirely confident about. That is how you build a foundation with data that is highly accurate, upwards of 99% in some cases.
Once you have the foundational data set and all the noise is removed, we start building out your accounts. We also have technology to consolidate subsidiaries into one main account so you are not looking at six separate subsidiaries when they all roll up to one parent company.
Diving into an example account like Graystone Corporation, you can see the parent master services agreement and all of the subsequent child agreements such as order forms and amendments related to the MSA. If you dive into any one of those contracts, you have the fully executed version available. Because we built this foundation, you can start doing very interesting things like chatting with your contracts, doing custom reports and dashboards, and on a C-level perspective, tackling very complex business problems.
Pillar 2: Trust
James Chou: Now that we have discussed building the foundation, the second pillar, which is equally important, is trust.
You don’t just trust the builder’s word that everything is up to code. You would hire an independent inspector to show you exactly what’s behind the walls. That is no different from the framework we are building here. When evaluating contract intelligence solutions, you really need to know where every data point is coming from.
A lot of the concern these days is that AI can analyze documents at scale and do metadata extraction, but where is the source? How can you verify the accuracy? You are not going to blindly trust a vendor. You want a mechanism to actually verify accuracy, and in cases where anything needs modification, you need to know how you can easily correct it without submitting IT tickets and waiting a long time.
There are three key areas here:
Accuracy and transparency. Can you see exactly where each extracted data point comes from in the source contract? If you’re extracting the effective date from an MSA, you probably want to know where in the source document the AI found that. You need to be able to audit the AI’s decisions on the spot. AI really shouldn’t be a black box. You need to be able to trace every answer to its source. And if you can’t trust it, you’re not going to trust the solution for making critical business decisions.
AI hallucinations. This is a very fair question to ask any vendor: how do you minimize hallucinations and allow users to quickly verify outputs? If the AI is giving you different or incorrect answers every time, that is a problem. As lawyers, you want consistent answers. You don’t want the AI giving you one answer the first time and a completely different answer the hundredth time.
Quality assurance beyond the AI itself. We are not at the point where AI can do 100% of the work. There will always be edge cases. Businesses are inherently unique depending on the vertical and industry, and some contracts are very specific to your business. You need to make sure there is a QA mechanism in place. At Pramata, that is where our AI plus human-in-the-loop paradigm comes in. While the AI can deal with the majority of cases, there will be exceptions where you need to work with our QA team to ensure anything that needs to be verified actually is.
Demo: Verifying AI Accuracy in Pramata
At Pramata we have two main mechanisms for establishing trust. The first is our digital snapshot, which shows you the source documentation and citation for where terms were extracted. Going back to the Graystone account example: if you click on the MSA, you can see all of the metadata. If you are ever curious how the AI extracted the other party as Graystone, you can hover over the camera icon, which is our digital snapshot. It shows you the citation for where the AI extracted that attribute, and you can click through to see it highlighted in the actual document.
This becomes especially powerful for something like term renewal, which has historically been very difficult to track over time because commercial relationships are never static. You might track an MSA that was signed years ago, but there was a series of amendments and continuous order forms. At Pramata, we can actually show you how certain fields change over time across the entire family of documents. A series of amendments can actually change the material picture of the relationship. If you were only looking at the MSA with its original fixed term, you might think this is an inactive relationship, but in reality that is not the case.
When you sign net new deals with an existing client, this picture updates, usually within the next day. If another amendment is signed, it will be reflected in the repository. The AI does all the metadata extraction, and if that changes the material picture of something like a term or renewal date, that information is automatically updated. You are no longer manually keeping track of this information.
The second mechanism is our TrueCheck dashboard. We actually show you the AI accuracy scores across your entire repository, including precision, recall, and accuracy rates for every attribute that was extracted. Yes, the AI will be confident in most extractions, but there will be some where it is not entirely sure. If the AI is not confident in an extraction, it will kick it off to human validation. Extracting signatures from signature blocks, for example, has historically been very difficult. In those cases, the AI is smart enough to flag it for human review.
You can also have an audit trail showing who went in and modified any data point, ensuring the data remains highly accurate. This combination of source data citations, full accuracy dashboards, and the ability to work with your team to rectify any issues means this is no longer a black box. It is fully transparent and auditable on an ongoing basis.
Pillar 3: Adaptability
James Chou: Now we have talked about foundation and trust. The third pillar is adaptability.
A custom home is designed for how you live. You can adapt the space as your needs evolve without tearing it down and rebuilding. Think about model homes. They look beautiful, but is that house actually configured to how you want to live in it? More often than not, the answer is no. You are going to make a lot of modifications to fit your actual needs. And you can think of your cross-functional partners as roommates. Does everyone in finance and sales want to live in a space built exclusively for legal? The answer is almost certainly no. You have to make sure the house accommodates everyone’s preferences and needs. That is no different from a contract intelligence solution.
There are three areas to consider:
Playbooks and standards. Can you upload and maintain your own playbooks and make changes without IT involvement? Your organization has specific business practices and negotiation guidelines. Can you inject those into how you redline documents? You don’t want some generic LLM making its own judgment when redlining. You want the AI to take into account your business considerations and processes, and use your playbooks as guidance.
Customization capability. For non-technical users, can they build their own agents for specific tasks? Someone in sales might have a question about their contracts. How easy is it for them to access the information they need without submitting a ticket to legal, waiting a couple of days, when they could have gotten the answer within seconds?
Growth and evolution. Can the solution actually adapt to your business? You do not want to be stuck in a cookie-cutter solution. The ability of the solution to adapt to the way you do business is ultimately what drives adoption and whether people actually use and benefit from the system.
Demo: Adaptability in Pramata
Going back to the Graystone example, your legal teams and cross-functional teams will have questions about their contracts. What is frustrating is that a lot of times your contract managers are inundated with work and have to prioritize their deal flow, while on the side your sales reps might be asking things like, do we have an agreement with this other party? Can we sign this NDA as is? Or in the case of Graystone, how did the amendments change the MSA terms over time?
You can see these blue buttons in Pramata, which are out-of-the-box AI agents. Our customers have used these, but they have also created their own. That is really the beauty of Pramata: we give you the foundational building blocks to benefit from our out-of-the-box agents but also build your own. You can have a price escalator agent, a tariff risk agent. The sky’s the limit with what you can do with AI agents.
Here is an example. Manually, you would have had to read through the MSA and all the amendments to piece together how the relationship changed over time. With a click of a button or a natural language question, you can get answers in seconds. This is the power that comes from having that foundation and being able to trust the data is accurate.
We also have integrations with CRM systems. For your sales teams, they don’t have to learn an entirely different system. Within Salesforce, a sales rep can check whether you have an NDA with another party. That question would have previously required an email to legal that might have sat in an inbox for a day or two. With Pramata, the requester can get that information in seconds. The whole workflow is fully customizable to your business needs. We have a drag-and-drop interface for modifying the contract request workflow, and our folks within legal have found they can manage and own this process entirely, building new workflows and modifying them on the fly.
In the case where the other party insists on using their own NDA, you can build out a quick checklist so the requester can check whether you can sign the NDA as is. You attach the third-party NDA, click the agent button, and the AI compares it against 10 or 15 things the legal team cares about when reviewing NDAs. In seconds, your sales rep can see whether you can sign it as is or whether modifications are needed. If it needs to go to legal for further review, it might take three to five business days, and the requester knows that upfront. The checklist is fully configurable to your legal team’s needs, and your requesters get all this information upfront so neither they nor legal is wasting time.
For playbooks, you can upload and maintain them directly in Pramata. If you don’t have a playbook, we have a Playbook Creation Wizard where you upload a representative contract with terms that reflect how your legal team negotiates. Through a series of guided prompts, we can build you a playbook within a couple of minutes. You can then use those playbooks in your negotiations and easily modify any terms on the fly.
In our AI Negotiator for Microsoft Word, which is a Word plugin, you can set which provisions your organization cares about when running a playbook analysis. Because the plugin is connected to the repository, it can actually show you your precedent analysis. If you want to analyze limitation of liability, it will pull up what you have agreed to for that provision in the past across all prior agreements. The AI then does a term-by-term analysis of the limitation of liability provision in the current document.
Based on all that context, you can apply redlines directly to the document and add counterparty-friendly comments justifying your positions. You can do account-level research, run your full playbook analysis, and inject redlines directly into the document. After completing your redlines, you can save the document in Microsoft Word and we handle all the versioning for you. With integrations with DocuSign and Adobe Acrobat Sign, you can kick off execution directly through Pramata. Lawyers can just live and work in Microsoft Word, but all the data is being captured and is fully auditable within the Pramata platform.
You can also do a diff compare in the Word plugin. If you have a redlined version and want to compare it against your standard template, you can do that comparison, then systematically go term by term and see where there are material differences, apply redlines for those, and make your contracts compliant against your organization’s standards.
Pillar 4: Impact
James Chou: The final pillar is impact. We talked about foundation, trust, and adaptability. At the end of the day, it is really about impact.
A well-designed home should solve your actual living problems for the entire family. You don’t want to buy an expensive system for only a couple of users. Ideally, your entire organization should be able to benefit from a contract intelligence solution to drive better business decisions from their contract data.
There are three areas to consider here:
C-level executive value. What questions can your CEO, CFO, or board ask and get immediate answers to? In an M&A use case, you have a very finite amount of time to do due diligence on a large number of contracts. Can you do that evaluation quickly and surface any questions that executives might have?
Cross-functional access. How do sales, finance, procurement, and operations teams benefit from the platform? This is not just a legal-only tool. Whatever solution you go with, you want it to drive value and impact across your whole business.
Complex scenarios. Can this support M&A due diligence, renewals, tariff risk, or portfolio-level consolidation at scale? These are your intractable business problems, and you need your solution to actually be able to tackle them.
The right contract intelligence solution creates value across your entire organization. It enables executives to make strategic decisions, allows sales to close deals faster, and enables your legal team to become more strategic. They don’t want to be doing hundreds of NDA reviews every day. They want to impact your business and work on highly complex deals, which is where the real value is.
Justin Schweisberger: The thing you don’t want to have happen is that you’ve invested the time, done the research, selected something and deployed it, and then an executive comes and says, can you tell me where in our contracts we have exposure to tariff increases? Or we have a breach issue, what are our notification timelines? And your answer is, it’ll give me a while. The immediate followup is, I thought that’s what we bought this platform to do. It is important to stress-test these scenarios. They are also the types of things that, as you build a business case for change, are really critical to include.
I also wanted to address security, which is something that comes up a lot as you look at any AI solution using LLM vendors. We are SOC 1 and SOC 2, Type 1 and Type 2. We do not use client data to train the system. We take a modified RAG-based approach. There are a whole set of questions you can ask on this topic, and if anyone has specific questions for Pramata on security, feel free to reach out.
Demo: Impact in Pramata
James Chou: For the M&A due diligence use case or any kind of complex issue, you are not just thinking about doing an analysis on one contract or even two or three. You are getting into the realm of portfolio-level analysis. Because we have that foundation built with highly accurate contract data, you are actually able to define the parameters of what is important when you are doing something like an M&A due diligence risk assessment. You tell the AI what matters in your analysis, and it performs that across an entire portfolio, whether that is a few contracts or hundreds.
Imagine trying to do this contract by contract manually. It takes a lot of time and can be error-prone. With Pramata, you can very quickly build customizable dashboards. This is just one example of an M&A risk dashboard. Each client’s M&A dashboard looks very different because each organization cares about different risk exposures. You can build this out to report on exactly what your organization cares about, and you can ensure the data is highly accurate because you have that foundational data set built out. You can start getting these kinds of customizable dashboards in minutes to an hour depending on your use case, and they meet your reporting and audit requirements.
This applies to many other scenarios beyond M&A, including tariff exposure analysis and data rights usage. Because we can do this at scale, you can get answers very quickly while ensuring data accuracy is high.
Out of the box, we also have an auto renewals dashboard. Because we track how terms change over time, including end dates and renewal dates, you can keep track of all your auto renewals in one place. Finance and procurement teams can access this information in Pramata without bothering legal, and they can also export the data as a CSV or Excel file for their own reporting.
We also have email notification mechanisms. Depending on how much notice you need to provide to your vendor, we can send automatic emails to your contract managers or procurement teams so they are never caught off guard by an auto renewal. This lets you systematically go through upcoming renewals, run a vendor risk analysis, and ensure those contracts are compliant against your standard terms, all while driving cross-functional value that goes well beyond legal.
Closing Framework and Key Takeaways
Justin Schweisberger: Hopefully you got a couple of nuggets to take away. A few of the big takeaways for me before we even get to the pillars: really understanding the problem you are trying to solve and knowing your business, knowing how contracts are done and may be done in multiple ways in different places. That is one of the very first things that sets a good foundation. If you can zero in on the right problem, you can pick the right solution.
Cross-functional buy-in is really important. Think about all the different people in a company who need information from contracts to do their job: a sales person looking to upsell or renew a customer, a services person trying to meet SLAs, a customer success person getting called in, someone on the finance team running a price uplift analysis, and the legal team itself. Help me understand what’s in our contracts really spreads across the whole organization. You don’t need everyone to spend hours on this, but make sure you understand their needs and can represent those needs, then bring them in at the right time.
Can you set a good foundation? There are a lot of companies with really great dashboards, but the underlying data is wrong. Amendments aren’t factored in. Can you trust that information? Transparency into what’s happening is really important.
Can you move quickly in the face of change? We have had what feels like 10 once-in-a-lifetime events over the last six years. Are you ready for the next one? Are you ready to keep up with the pace of AI change?
And finally, but most importantly: what is the real impact you are looking to make in the organization? It may cause people to stretch a little out of their comfort zone. But there is a lot of good work being done by people on this topic, and sometimes it is being done through sheer brute force. If you don’t stretch yourself to make that case and really look at the overall impact, people won’t see those efforts. They’ll just see how long it takes or that you had to hire 20 temps. Really appreciate the time today. Hopefully you found some things useful, and feel free to reach out if you’d like to discuss further.
