Legalweek and AI: Revolution, Evolution or Same Old, Same Old? 

Our team’s conclusion was that generative AI is almost making it too easy to develop apps that leverage ChatGPT to create any kind of document, including legal contracts. AI was everywhere in the conference but is it anywhere in the world of law? Is using AI to generate contracts with the push of a button really so much different than using Word templates some 20 years ago?

Lawyers have always been slow to adapt to technologically driven change.  This is due to the profession’s focus on expertise and experience-driven, hand-written editing, email, and color-coded folders that still rule the day in most law offices. Yes, the legal profession has taken to Zoom teleconferencing, not necessarily because they wanted to,

 but they had to during the pandemic. But age-old issues, like discovery, still have not been fully solved with technology. Neither has the concept of charging per hour.  Its interesting that the legal profession complains about the unpredictability of software costs, yet, thinking about the per hour business model, where is the predictability in that? 

Reflecting on the “AI is everywhere,” message, it was very difficult at Legalweek to differentiate one company’s AI from another. In other words, based on the high-level messages we saw, you couldn’t understand how hard the AI would be to train, implement, change, manage, and so on. It has devolved into an almost meaningless—but required—marketing term, like “scalable” or “easy to use.”  Additionally, the business value was, almost without exception, focused on making users and processes more efficient.  Clearly, cost always matters, but what about business outcomes for clients.  The real question – , “Can AI generate a more efficient database for legal documents?” – was not in the discussion bubble.    

AI-driven e-discovery was also a hot topic at Legalweek –So, what’s the big deal? Tech companies providing e-discovery solutions have been using forms of AI – primarily for supervised machine learning – for years, yet at Legalweek, it was positioned as “new and improved.”  There is a growing frustration among corporate e-discovery users around the cost and quality of data.  Our conversations with corporate legal managers centered on issues of accuracy and over-discovery.  Maybe it’s part of the cost model for using an e-discovery solution, where every megabyte is counted each and every time repositories are indexed.  

One of the key conclusions of our EncompaaS team at Legalweek is, as Mr. T would put it, “pity the fool.”  We greatly sympathize with the tech managers at law firms and corporate offices whose work responsibility is to make the right decision on what AI-driven solution to buy.   That’s because the AI solutions are difficult to differentiate and the end result from a business outcome perspective is very difficult to predict.  At Legalweek, the primary benefactor of AI was focused on efficiency 

The rush to adopt AI specifically for the legal department is creating yet another data silo that is isolated from the core of the enterprise information infrastructure.  We don’t envy the task of having to explain the benefits and workings of AI to very highly paid lawyers and corporate counsels.  

The biggest issue of all which impacts the successful deployment of Generative AI and Large Language Models, data quality, hardly surfaced at Legalweek.  Without a strong data quality foundation, it’s going to be difficult at best and impossible at worst to effectively implement a data pool that can fully represent the totality of the enterprise knowledge base, to provide accurate and complete answers to queries and the effective generation of content. That same foundation is also required to deal with the discovery and management of personally identifiable information as it relates to the growing number of privacy laws across the United States.  

We had a lot of good conversations at the conference; don’t get us wrong, we met with some very smart people who were very much interested in finding out ways to not only be efficient but also to use AI to provide better business outcomes.  The three takeaways that stand out from our week in New York:  

1. E-discovery still has a long way to go, especially from a cost modeling perspective.  

2. Data quality should be mission No. 1 for technology teams supporting legal operations.  Without the full use of data from across the enterprise and the use of AI to vet, label and tag information, using Gen AI and LLMs is a risky venture, in terms of accuracy and safety.  

3. Focus on using AI to produce better business outcomes.  Legal documents have tremendous business value.  And with the use of AI, these documents can provide insight and ideas just as important as data lakes.  

If you would like to start a journey addressing these three takeaways, give us a call at EncompaaS.  We can help you create the data quality foundation to ensure the effective and safe use of AI apps that will connect the dots between legal obligations and terms and business performance.