CCBJ: Would you give us your perspective on how contract lifecycle management technologies have evolved over the past 5 years, and where we are now?
Ziad Mantoura: Over the past decade we’ve gone from a
small number of CLM technology players to a plethora of
players. You can’t go to a conference now without seeing
new names, new products and new launches. Part of that
growth is because of people recognizing that technology has
a meaningful role to play in unlocking the value of contracts
across the enterprise. And part of that is also the fact that it’s
easier and cheaper to build modern technology on a cloud-
based architecture than it has been historically.
How would you describe the current state of generative artificial intelligence within CLM technology and what exactly is it helping contract professionals do?
There’s a lot of buzz around generative AI, both in society
generally and in legal, and contracts is one of the areas
where I expect generative AI is going to have a meaningful
impact. When we talk about generative AI as opposed to
more traditional AI, the “generative” part refers to the
creation of new content that doesn’t currently exist. That
can be words, images, music, etc. In the legal industry, and in
contracts more specifically, we’re talking mostly about words,
and we’ve seen related generative AI capabilities come in
different waves. If we rewind to early 2023, we started seeing
some of the first generative AI features coming out within
CLMs. What they were essentially doing is helping with the
drafting aspects of contract negotiation; the idea that you
could [Right Click], highlight a clause, ask it to “Make this
mutual” and it would do so.
That approach was helpful to show what was possible with the products, but it didn’t deliver that much value when you think about the needs of many large enterprises. If you are in a B2B, large volume contracting business, you want as much standardization as you can drive in your contracts, but with the flexibility to get the deal done with your counterparty. So, if every time you’re asking to make the clause mutual, you have no standardization across process.
What we’re seeing now is the next evolution of CLM, where the providers are linking live negotiations back to your contract repository by way of VectorBDs and/or a RAG-based methodology. If we take the same “Make this mutual” request, the LLM will go back to your repository to make it mutual in a way that you’ve made things mutual in other approved contracts.
The other approach we’re starting to see that is showing some real promise is saying, “Well, the problem with my repository is I’ve got some documents there that are great, but I’ve also got some documents that I wish I’d never signed. We had to sign that agreement because the counterparty had all the power, but that doesn’t mean I want that in my next agreement.” So we moved on to saying either you have a golden source repository of the contracts you want to refer to or you go through a playbook process. We’re seeing CLM providers build out both of these alternatives in their product. At this point it’s not clear which is going to prove to be the better option.
What we’re now seeing— an emerging 3rd wave—is the ability to use generative AI to help you produce a playbook, or at least get it 70 or 80 percent of the way there. So rather than saying, “Hey, we need lots of time from some senior attorneys and we’re going to start with a blank piece of paper or a template that might or might not be appropriate,” we’re saying, “Okay, we’ve got the template. We’re going to use generative AI to populate it with a first review based on what we’ve seen across a sample of agreements from your starting points, first markups and final versions, getting you to that playbook world much quicker, which then allows you to have much more effective negotiations.”
The world we’re going to be in is not one where you’re saying, [Right Click], “Make this mutual”. It’s got to be [Right Click] “Give me the same concession as I gave to this customer six months ago.” Or [Right Click] “Is there a pre-approved compromise position I can offer this customer?” that follows your playbook. Firstly, it is a time-saving device. But more importantly, it allows the legal department to get away from negotiating every agreement and to a place where they can confidently say, “Contracts below X value can be self-service by the salesperson, the product lead, the business lead, whoever it might be in their business, because they’ve got the right parameters and technology assisting them.”
What does the timeline look like for the regular adoption of gen AI that is generating these playbooks and helping create golden sources?
The underlying technology is here now. It’s not 1, 2 or 3 years out. That said, the timeline for adoption of new technologies in legal has always been slow. I think anybody looking at a CLM today should be looking and saying, “What is this company doing on the AI and the generative AI fronts? Where are they going? What does their product roadmap look like?” There are companies out there that already have in play, at least in demo environments, everything I’ve described to you. But of the dozen or so CLM providers I’ve spent meaningful time talking to, it’s on everyone’s roadmap.
Of the CLM technology vendors you have spoken to about this, are they hesitant? Motivated? When do you think they’re planning on testing use cases?
I’d say there’s a spectrum. There are some folks who view it as a little bit of a novelty; something they’ve got to have because everyone else has got it. Then, there are those who have been ‘all in’ for a while, and they’re a little bit ahead of their competitors. We are also seeing CLM providers make acquisitions of smaller companies with AI expertise. I expect that trend to continue as part of a broader consolidation of the CLM market.
For most genAI technologies out there today—not just in the CLM space and not just in legal—the underlying LLM has been OpenAI’s GPT models. That’s the most prevalent. CLM vendors are looking at alternative models, at what’s available from different providers or open source. They’re experimenting with what’s more effective and making sure that they have an underlying architecture that allows them to quickly move to new models, as well as address customer concerns on data security, and make sure the price-point of the LLM works for the CLM’s pricing and customer business case.
You need to have an architecture that can take advantage of different LLMs as they get released because we simply don’t know who’s going to win that race. And there’s literally billions of development dollars pouring into the next generation of LLMs.
How will generative AI’s impact on contract management technology benefit your typical legal departments? How will it change the way organizations look at their contracts?
When I think about the impact generative AIs can have across legal—law firms, in-house counsel, alternative legal service providers—there’s a huge wealth of data that we’ve had in legal for some time, but it’s been trapped, locked and often unstructured. Generative AI, combined with modern cloud infrastructure, presents an opportunity for that data (and the value therefrom) to be unlocked. In the contracts space, there is huge value that sits in executed agreements and knowing what is in those contracts from both a business and risk point of view.
There’s also value from the finance point of view. I can’t tell you the amount of times we’ve spoken with finance teams about a contract problem, and they want to remove key financial data such as payment terms, penalties, automatic price raises, etc.
From a large organization point of view, being able to know what’s in your contracts, having a good contract repository that you can run analytics across, that you can ask questions and get answers, is going to be important. That’s what I see the future model looking like.
Where else do you see generative AI having enterprise impact?
Outside of contracts and CLM, there are a number of areas where generative AI is showing promise on both the corporation and law firm sides. On the corporation side, a lot of organizations look at genAI as a large opportunity for their core business or, in some instances, as a significant threat to their business. So, there are many corporations spending large sums of money around generative AI because they think it will bring transformative efficiencies, or because there are potentially competitive threats out there.
Microsoft is ahead of the pack here. We are their 2023 “Partner of the Year” and we’ve been working with them and dozens of the world’s largest companies on how to roll out the Microsoft Copilot product line-up in a safe and secure way that legal teams are comfortable with; which addresses such concerns as, “What data is Copilot going to have access to? And are we comfortable with it having access to that data?”
We see this big wave of generative AI happening across the corporate world, and legal has a role to play in that. They’re looking at how to leverage that Microsoft technology, including how it can be used for the legal departments. We’re seeing that play out in terms of more effective search and knowledge management. It’s not just about drafting but also being able to find and ask questions and get answers, and we’ve seen multiple use cases across organizations and legal departments.
Within law firms, it’s slightly different. The main reason why is there’s a significant economic opportunity for law firms to unlock the data that they’ve been sitting on. At its core, a top tier law firm is a collection of incredible legal intellects with knowledge of different practice areas, subject matter expertise and experience. Generative
AI is providing a interesting opportunity in this broad knowledge-management space. In the past several years we have seen KM departments in large law firms become increasingly tech-forward recognizing the unlocked value in the data they sit on.
Large law firms are investing significant capital in generative AI because they see it as an opportunity to have the data they’re sitting on elevate themselves ahead of the pack.
Generative AI enables law firms to do their work more efficiently and make higher margins, or to justify charging higher rates. Ultimately there’s a big incentive for large law firms being able to unlock the power of their data, because if they get it right, it will make them even more successful than they’ve been over the past decade.
We’re starting to see it with some of the big firms. They’ve always been okay with embracing third-party technology that doesn’t touch their data. Now the competitive advantage is not only in using a third-party tool better than other law firms—which will continue to happen in the same way it’s always happened—but in unlocking the power of the knowledge these firms have been sitting on to deliver additional value.
It’s early stages. No one is saying that they can 100 percent guarantee they’ll be successful. But there are firms that are saying, “This is a real opportunity for us to differentiate ourselves from our peers,” and they are prepared to put some money behind it, which I find really exciting in an industry where too many firms have taken a wait-and see- approach to new technology.
Any final thoughts?
Anybody who is in legal and plans on being in legal for the
foreseeable future, should learn some of the basics around
generative AI, think about some of the use cases
that could be effective for their organization or
their law firm, be prepared
to experiment with it,
and don’t be afraid to try and fail. Generative AI is going to have a real
impact on our world and understanding the
fundamentals will form
the foundation for creating
an effective generative AI
strategy.
Published July 18, 2024.
