chevron-down Created with Sketch Beta.


Moneyball and the Business of Law: A Whole New Ballgame

Kevin Walker

Moneyball and the Business of Law: A Whole New Ballgame Kubeš

When ChatGPT launched in November 2022, it appeared that Shakespeare’s quip about killing the lawyers might actually come true.  Since then, there has been no shortage of doomsayers predicting the end of the billable hour and the demise of junior associate ranks.

Rumors of the legal industry’s impending death have been greatly exaggerated.  The prevailing narrative is overlooking the emergence of a paradigm shift that may well breathe new life into the profession.

At the recent Legalweek Conference – the legal industry equivalent of the Consumer Electronics Show – the exhibit hall in Midtown Manhattan was overflowing with AI-powered chatbots claiming to speed up document analysis and drafting.  Without a doubt, the application of generative AI to these notoriously tedious workflows is changing how law is practiced today.

But in the R&D labs of Silicon Valley – and increasingly, in user-facing applications – a new AI superpower is emerging that will change how law is practiced tomorrow.

The capacity for large language models (“LLMs”) to mimic human reasoning has deeper implications for the practice of law than merely accelerating due diligence and document generation.  Namely: LLMs are exceptional at deriving insights from vast quantities of unstructured information.

Though less attention-grabbing than ChatGPT’s poetry or DALL-E’s art, the automated extraction of nuanced concepts from complex documents heralds the transformation of law into a data-driven profession.

In transactional practice, the primary role of an attorney is that of a dealmaker, where success is measured by the ability to close deals quickly and efficiently on the most favorable terms for their clients.

Each negotiated term represents a “deal point”: a contractual variable that the parties have agreed upon, often through hard-won battles.  If one could extract and aggregate deal points from a large set of transactions and run analytics on this data, a world of possibilities emerges.

The problem is that deal data is often buried in complex provisions, in some cases requiring hours of attorney labor to extract relevant information from a single document.  As a result, troves of valuable insights collect dust in the depths of law firm databases.

Enter LLM-augmented data extraction.  We can now teach machines to emulate how an attorney would review a contract and extract its deal points.

In our research at Centari, the results have exceeded our expectations.  Not only can the machine accurately extract easily interpreted data like dollar amounts, it can also capture abstract information requiring a deeper understanding of the law, like the relationship between a party’s representations and its liability.

For an industry that has historically relied on intuition to drive decision making, this is a watershed moment.  Imagine going into a negotiation with total command of the chessboard, knowing your competitor’s past positions, visualizing market trends, and pinpointing where you can push the envelope for your client.

In the data science world, the power of data to change the rules of the game is sometimes referred to as “moneyball” – an allusion to the book and movie depicting the 2002 season of the Oakland A’s baseball team.  Moneyball follows general manager Billy Beane, fresh off a postseason defeat by the Yankees, as he develops unconventional statistical models to recruit undervalued players and build a competitive team on a limited budget.  The story illustrates a paradigm shift in how ballplayers are picked – a shift away from subjective criteria in favor of data-driven insights.

Consider the implications of such a strategy for the competitive business of law.  At the negotiation table, information is power.  In a battle over a merger agreement’s materiality threshold, which attorney wins the day: the one who cites their personal experience negotiating similar provisions over the course of their career, or the one who has data at their fingertips enabling them to hammer home the reasonableness of their client’s stance?

This is moneyball for the legal industry, where underdogs can outmaneuver star players by harnessing the full force of their firm’s latent knowledge.  And so the law firm data arms race begins.

In one version of the future, the revenue gulf between large and midsize firms will widen as large firms capitalize on their massive repositories of private data.  However, a bureaucratic culture could inhibit some of the largest players from seizing this opportunity.  This presents another version of the future where smaller firms outpace their rivals across the dual vectors of service efficiency and effectiveness.

Yet another possibility is that clients grab the wheel themselves.  Through a combination of public deal data and pooled repositories of anonymized private data, businesses might one day build enough transactional intelligence to disintermediate their attorneys.  Shakespeare is smirking.

The legal industry today is standing on the shore, watching the water recede.  Sure, there will be firms that cling to age-old assumptions and find themselves overwhelmed by the impending wave.  But increasingly, we are witnessing firms of all sizes recognize the signs and seek higher ground.

For these innovative firms, the future of the legal profession is not one of obsolescence but of opportunity, where the synthesis of human judgment and LLM-augmented intelligence opens a new frontier of data-driven dealmaking.