This article was originally published in Legaltech News.
You might have thought the title of this article was directed at our children and the status of their bedrooms, but the topic of AI in contract lifecycle management (CLM) is often just as messy. Everyone wants to leverage AI and believe they are incorporating it into their CLM systems, but are they really? For example, is getting a notification of a contract renewal date inside of a CLM record truly AI? Or is that type of helpful ping just the tip of the AI/CLM iceberg?
I would argue it’s the latter. Yes, many of the functions found in today’s CLM systems could be considered a form of AI, but most are just scratching the surface potential of AI and CLM. The truth is that AI and its associated technologies, such as machine learning, have the power to completely transform the contract management process, making it much smarter and more efficient. And it’s just getting started.
Now, it’s time for you and your corporate legal department (CLD) to get started, too, but that can be easier said than done. To help, let’s take a look at what AI really means to CLM and how taking a crawl/walk/run approach to building good data models is key to their success.
What Does AI Really Mean to CLM?
Today, AI is very good at handling basic tasks. No, it’s not going to replace an attorney, or “think better than” a lawyer. But it can handle the mundane tasks no one enjoys—tracking the status of contracts and renewal dates, and even automating the assembly of documents and first drafts of agreements. All of this helps free up lawyers to do what they’re trained for: deliver value that contributes to their corporations’ bottom lines.
As more contracts are fed into the system, the AI is “trained” and becomes more intelligent, mature and capable. Instead of just basic notifications, a mature AI provides insights into common data points like choice of law provisions and helps CLD teams manage instances of force majeure, business interruptions, or limitations of liability. It identifies contracts that include specific provisions and delivers results in real-time without requiring attorneys to contact the head of legal or other business units. It tailors risk profiles from customer to customer.
In short, it makes the CLM process more efficient and accurate, which is what everyone wants. But, the efficacy of the AI depends on one thing: the data that powers the AI engine.
That data must cleansed to ensure high quality, but this process poses its own challenges. First, data should ideally be standardized, but that’s not always realistic since every lawyer writes contracts a bit differently. Second, non-conformity amongst contracts—for example, an NDA that has five data fields vs. an MSA with many more—makes achieving consistency wishful thinking. Third, some attorneys may not want their data included, resulting in inaccurate sample sizes and many disparate data models.
The good news is that while these are notable challenges, they’re not insurmountable if you take a “crawl/walk/run” approach to the AI for your contracts journey (perhaps good advice for CLM implementations in general, but that’s a topic for another article).
Start by documenting the desired internal taxonomy for staff and outside counsel. Be proactive engaging the key stakeholders, including non-lawyers. For instance, talk to contract administrators or business analysts about how they work with contracts. Break down siloes between lawyers and other business functions to ensure everyone is on the same page when it comes to their expectations of the technology and how it will help the entire team.
Next, migrate legacy contracts into the AI system. Not everyone will see the benefits of putting older contracts into the system (ironic, given that these professionals rely heavily on past case law and the nuance of language). But communicating the resulting benefits of migrating legacy contracts—faster development of consistent contract language, fewer mistakes, timely notifications, risk scoring, and less time working with the legal department—can help get them on board.
Before starting migration, be sure to perform a high-quality scan to ensure information is updated and accurate. I’ve seen too many attorneys become frustrated (“the AI isn’t working!”) because the scan of the duly executed contract isn’t legible.
Once documents are in, start walking. Classify clauses for better consistency, beginning with those that are the easiest to categorize—like venue provisions. Once basic information is in, add complex data, such as termination provisions, limitations of liability, and non-standard contract terms.
The AI is then in position to “run” and deliver more meaningful results. Think of AI quickly identifying contracts based on risk scoring for better risk and compliance management. Imagine lawyers no longer being caught flat footed in the event of a lawsuit.
Those are among the many possibilities, but there will be many more as the technology matures. A few years ago, no one would have thought we would be collaborating with colleagues and inputting redline comments in real-time into an MSA, yet that’s now standard practice. Hopefully, AI in CLM will become similarly commonplace and deliver significant long-term value, greater efficiency, and lower risk.
Now, kids, get your house in order—and go clean your rooms (and your data).