Select Page

How AI will transform contract lifecycle management

Dec 28, 2023 | Admin, Latest News

Artificial Intelligence (AI) has exploded in the past couple of years, in large part due to huge advances in AI capabilities and the democratization of very powerful AI offerings like ChatGPT. It has crept into many of the business tools we use to make our lives easier. In theory, it can accelerate the effectiveness of those tools, providing the power of artificial intelligence that is getting nearer and nearer to the natural, creative intelligence that you and I have.

One such toolset that has already embraced the power of AI is Contract Lifecycle Management (CLM). CLM tools help organizations prepare, sign, act on, and manage the various contracts and agreements in their day-to-day operations. A CLM can quickly generate documents; facilitate and track negotiations; manage the signature process; and help organizations convert the unstructured data in their documents into structured data—the kind that helps them make informed business decisions and get a competitive edge. All of what CLM does can be (and in some cases already has been) improved with AI.

What is the state of AI in CLM right now? Where might it go in the near term and not-so-near term? What are the risks and what might be the unintended consequences? What should those using or thinking about using AI for CLM look out for? Here are my thoughts on how AI will affect CLM.

 

It Will Help Create Structured Data

A CLM tool powered by AI will take unstructured document data and create structured data.

Anyone who has used a spreadsheet is familiar with structured data. A spreadsheet is data stored in a very structured format—within columns, rows, and cells. Running analysis and calculations is made much easier because the data is so structured. Do you have a personal finance spreadsheet and want to know how much you spent in November? Just take the sum of the November expenses column. Easy.

Even if it isn’t readily apparent, a legal contract is another way of storing data. Data in this case are the various terms, stipulations, and obligations written within the contract—payment terms, agreement milestone dates, limitation of liability amounts, and more. But contracts are a form of unstructured data. The start and end dates of a contract might exist, but one would have to search for and find them within the text. Want to know how many days separate the start and end dates? You’ll have to figure that out manually. Good luck.

Artificial intelligence can do the searching, finding, and analyzing for you. It could automatically and intelligently pick out the most important pieces of data and store those values as metadata against the document. The unstructured data—the contract document itself—still remains, but now with structured data associated with it. With associated structured data you can more easily run analyses such as comparing contract values or limitation of liability amounts. You can run reports such as one that displays all contracts coming up for renewal so that you can act on them promptly.

 

What to Watch Out For

While the ability to create structured data is great, AI is not infallible. It can make mistakes. Pulling metadata from a document written in natural language is not as simple as summing the values in a spreadsheet column. A CLM with this ability should give you (or perhaps even force you) to review the data it extracted. It should also show you why it picked that data, perhaps highlighting a sentence or two from the original document that contains the data. When choosing a CLM make sure you have this sort of control over data extractions.

 

It Will Summarize Contracts For You

You will spend less time reading and more time on high-value tasks.

Corporate contracts are often very lengthy and full of legalese. Even for an experienced law professional, reading and understanding a lengthy legal document can be arduous and time-consuming. This laborious task is expensive. Ideally, highly-paid professionals would spend less time reading and more time doing the actual analysis and decision-making that follows.

Much like how AI can pull structured data from the unstructured data of contracts, AI can also take the lengthy and perhaps obtuse content in a contract and simplify it, making it easier to read. Think of it as an automatic CliffsNotes generator for contracts. With an AI-powered CLM, legal professionals can spend less time reading and more time doing more valuable parts of their job.

 

What to Watch Out For

AI needs to do two things when it summarizes a contract:

  1. It needs to figure out what is “important”.
  2. It needs to summarize the important content in a way that is true to the content of the original document.

If your AI-enabled CLM fails to identify important pieces of information you will miss out on important pieces of the original contract. And even if the AI has identified an important piece of information, it may misconstrue that information. Just what is important, anyway? And how did the AI decide what is important? How did it consolidate or summarize that information and did it do a good job? Look for an AI-enabled CLM that allows you to understand the answers to these questions and allows you to correct and further train its AI capabilities.

 

It Will Reduce the Need for Legal Support

Legal and contract teams will be pulled in only when absolutely necessary.

Organizations with CLM use the tool to facilitate negotiations, often with a salesperson acting as the face of the organization and the legal or contracts team acting in support. If a client returns a contract with substantive changes to legal terms, generally a salesperson is not authorized to accept or negotiate those changes without legal support.

An AI-powered CLM tool could read through your organization’s legacy contracts—ostensibly all with acceptable terms—and create a term library. After an initial investment of time by the legal team to review the proposed terms, salespeople would have a legal-approved library they could pull from for fallback language. They could support themselves and leave legal to spend time only on the most complex or risky contracts.

 

What to Watch Out For

If an AI uses your legacy contracts to build a term library, the assumption is that all of those contracts have acceptable terms. This assumption may not be true — a term may have been acceptable in the past but no longer is now, or perhaps an exception was made for a particularly large deal with a strategic client, and you’d rather not accept similar terms regularly.

Salespeople are also not lawyers; a term library as a tool is only as effective as the person who wields it. And with the wrong person, a term library could be a liability. You must balance the need for contract velocity with the need for risk reduction.

Look for CLM tools that allow legal teams to review proposed term library terms and rank them in order of fallback preference. Also, look for tools that provide guardrails and guidance for the use of term libraries.

 

It Will Help You Negotiate

AI will act as legal assistants and expert negotiators.

Imagine having an assistant who is schooled in law and could help you redline contracts. And imagine if that assistant was also an expert negotiator who could give you a competitive edge and help you close deals faster. An AI-enabled CLM tool could be that assistant.

A CLM tool could read a suggested edit from a counterparty and give you just the right language to use to maximize the chances of the counterparty accepting your counter-edit AND maximize the favorability of the terms. It may even give you a few options to choose from, or you could ask it to generate new variations until you’re happy.

You could even flip that scenario around, instead writing your own counter-edit and asking the CLM to give you feedback, scoring your edit against metrics like favorability, imperviousness, and likelihood for acceptance.

This level of AI assistance is almost like magic. How does AI even do something like this?

 

What to Watch Out For

Generative AI—the flavor of AI that uses Large Language Models (LLMs) to create suggested edits like the ones described above—has a tendency to have what is called AI hallucinations. When a generative AI “hallucinates” it starts to see patterns where there are none and it starts to generate content that is illogical and/or nonsensical. A hallucination is the last thing you want in an important legal contract governing important business relationships.

And what about the broader problems around LLMs and AI? Is AI nothing but a boon to CLM platform developers?

 

It Will (Continue to) Cause Controversy

The idea of artificial intelligence has been controversial since Isaac Asimov’s science fiction works of last century (and probably before that). But moral (and corporate) controversies aside, there are some practical issues surrounding AI and how AI works.

Though it may seem like it is, AI is not magic. It is trained on large data sets to do what it does. But where does this data come from? It might come from you! Your CLM vendor might use the contracts that you have loaded into your CLM tool to further train and improve its AI models. That might be okay if your data drives only your AI, but a CLM vendor could greatly accelerate their AI improvements if they could use their collective customers’ data to train an AI that spans across all of their customers. Make sure you know how your contract data is being used by your CLM vendor.

Imagine this. You’re using your AI assistant to help you negotiate with a client’s counsel—Jane Smith at ACME Corporation. Your CLM provider thought they would be very clever and track all of Jane Smith’s edits to train your CLM AI, and now it knows exactly how to craft terms that are just acceptable enough to her. It has become an expert at negotiating with Jane Smith. Pretty great, right? Well now look at it from her perspective. Will Jane Smith and ACME want to do business with an organization that has this sort of advantage? Or will they insist that you not use your CLM AI? Is such a tool even moral?

It gets worse. What if your CLM vendor takes the data that you have collected on Jane Smith and gives it to your competitor who is using the same CLM vendor? Now your competitor is also an expert at negotiating with Jane Smith and ACME Corporation, all on the back of your hard work.

And it gets worse than that. What if all of this AI training and data sharing was unknown to you or your competitor or to Jane Smith and ACME? It was all part of the “secret sauce” of your CLM vendor and shrouded in opacity. Then one day the “secret sauce” gets revealed and you learn that ACME Corporation data—along with all of the client data you’ve ever put through your CLM—has been shared amongst every other organization that uses your CLM provider’s CLM tool. That would be a huge controversy, and it would be disastrous for you.

How likely are all these issues? No one knows for certain. But it is important that when choosing a CLM vendor, they take potential issues like this very seriously. Work with them to understand how they train their AI models. How is their customer data used? What are their data privacy policies? What are the moral implications and are you okay with those?

 

Conclusion

AI is an incredibly exciting and powerful technology that can greatly improve the way we conduct business and it can greatly augment CLM systems. But there are also things to be wary of—and some things to be very wary of.

AI will affect CLM in (at least) the following ways:

  • It Will Help Create Structured Data
  • It Will Summarize Contracts For You
  • It Will Reduce the Need for Legal Support
  • It Will Help You Negotiate
  • It Will (Continue to) Cause Controversy

I am excited about how AI will help the world of CLM, and I think you should be too. Look for improvements in AI to accelerate in the coming months and years, but be cautious as well. When selecting a CLM vendor make sure to select one you can trust and you will be well on your way to better contracting.

0 Comments

Authors

[molongui_author_box]