Will AI disrupt legal this time or is it just more hype?
For years we've been hearing that AI will "disrupt" legal. The reality is that it hasn't, yet. "Disruptive innovation" has been misused and misunderstood. I’m writing to reclaim the phrase and share my view on why now, with the latest advances in AI, disruption may actually be coming.
Defining Disruptive Innovation
Harvard Business School's Clayton Christensen introduced the theory of disruptive innovation in the 1990s. The theory explains what happens when certain kinds of new products challenge established incumbents, and the behavior of incumbents in the face of such competition.
Initially, a disruptive product might be considered inferior to existing offerings and only address a small, low-end, or underserved segment of the market. The disruptor will typically leverage a different business model, a unique approach to the market, or, importantly for legal, a technology innovation that provides it with a cheaper cost structure.?
Incumbents, often focused on improving existing products for their most profitable customers, may overlook or underestimate the potential of the disruptor until it’s too late. Or, they may feel justified in not responding because they consider their product superior on traditional dimensions and their profit margins are higher.?
When the disruptor’s approach is based on technology innovation, like software, what happens next is key: the new product has a fast rate of improvement. The product gets better and better while maintaining its lower cost structure. As it improves, it appeals to a broader customer base and gains market share. In this growth, it can displace incumbents or force them to adapt to the new paradigm.?
Considering the current wave of AI as a disruptive technology/product and applying this theory to the legal industry presents an interesting theoretical case study.
Part 1: The Legal Industry Landscape
In an expansive study of the legal landscape, Professor Bill Henderson of Indiana University observes that the legal market can be divided into two segments, one serving individuals (People Law) and the other serving corporations (Organizational Law).?
While these two segments have different economic drivers, Henderson argues that they share the core problem of lagging legal productivity, which increases the price of legal services relative to other goods and services.
Corporate customers, under stress from the growing complexity of a highly regulated and interconnected economy, have coped by expanding their in-house teams and insourcing more work. We’ve also seen the rise of alternative legal service providers. Both of these changes come at the expense of law firms.
The People Law segment, by comparison, is characterized by a decline in paying customers and shrinking lawyer income. Increasingly, these customers’ legal needs can’t be profitably served by lawyers using a one-to-one model.
The traditional practice of law, and the regulation of the legal profession, is premised on this one-to-one relationship between lawyer and client. With only so many hours in the day, and only so many lawyers available, there is a hard cap to the amount of legal service that can be provided one-to-one. In contrast, legal service provided by software is one-to-many – a virtually limitless number of customers can be served by software with no marginal cost.
The solution, Henderson argues, is regulatory reform that enables lawyers to work closely with professionals from other disciplines, like technology, to drive down costs and increase productivity.?
Stanford Law School's David Engstrom picks up this argument but also posits that while regulatory reform is important, there are additional barriers. In large law firms, organizational and cultural barriers (e.g. the billable hour) have stymied innovation more so than regulation. And in People Law, the barrier to improvement has been a lack of truly capable technology.?
Part 2: AI as Legal Disruptor
There are a vast number of people and small businesses today who do not get legal advice because it is prohibitively expensive for them, even at a rate that Organizations would consider cheap (e.g. $300/hour). Their legal needs are varied, ranging from dispute resolution to personal, family, and transactional matters.?
The latest wave of generative AI has many known flaws, but it is also impressively capable across diverse topics. And it’s very cheap. Consider ChatGPT for the sake of a current example, but it is by no means the only option or the end-point for where this is going. Today, ChatGPT will answer questions such as:
It can answer all of those questions in seconds, at any time of day, and users can interact with it in plain English to probe more deeply and customize more.
ChatGPT is surely already providing many people with legal advice. And while that advice is probably less accurate, less nuanced, and inferior in many ways to a traditional lawyer, it is also free – and it’s better than the alternative, which is nothing.
领英推荐
It is here, in this combination of product features and cost and market, that massive disruption in legal feels most ripe with potential. The offering: "LawGPT" leverages new technology, has a software business model and is affordable, convenient, and "mostly" accurate, and serves a part of the market underserved by incumbents.
[Important to note that the regulatory landscape is not favorably disposed to this. There are likely to be regulatory roadblocks and/or litigation barriers a successful challenger would need to overcome.]
While the product today is inferior in accuracy to lawyers, especially in certain legal tasks, it will improve. As it improves, it will maintain its advantages: its low cost, speed, and convenience. Its increasing capabilities will make it appealing for a wider range of legal tasks, and to customers who have higher demands for accuracy and quality. Expansion within People Law to more and more use cases is predictable.?
At the same time, there are expansion opportunities into the legal work of Organizations, where there are a variety of tasks – not to necessarily be conflated with full jobs – that are susceptible to automation or strong tech enablement of lawyers.?
LawGPT may ultimately come to create and serve a vastly larger legal services market than the one that exists today.?
Part 3: Disruption in Organizational Law?
If we continue this study into potential impacts in Organizational Law, it helps to look more narrowly at specific tasks/services. As contracting is one that we at LegalOn are experts in, I want to focus there: the work of in-house legal teams reviewing contracts.
Today, legal teams perform contract reviews manually, with basically no technology enablement besides Microsoft Word. This is the “incumbent” service. The “customer” of this service is the business, which values reaching agreement quickly, cost-effectively, and with minimized risks. The manual contract review method, while thorough, can be slow, expensive, and prone to human error, especially with high volumes or complex contracts. Companies are rarely satisfied with this status quo.
Do all companies have the same balance of speed, cost, and risk preference? Certainly not. Some are interested in faster or lower-cost reviews, even if there is a trade-off in quality (or for some contracts but not others). Others are interested in improving accuracy even if it costs more. The market is not monolithic, and there is space for new technology to find a fit that works for some customers, even if not for all, yet.
AI tools for contract review, like LegalOn, hold the promise of increasing speed, accuracy, consistency, and scalability. While they can't today offer the full and nuanced understanding of a seasoned attorney, they can review limitless numbers of contracts quickly, tirelessly, identify potential issues, and – in the case of LegalOn – offer supportive guidance that improves human review.
These abilities, and this powerful “AI + You” partnership, are no longer theoretical, they are proven in practice. For example, teams using LegalOn report average time savings of 40%, improvement in quality and consistency (90% report quality improvements), and the ability to scale a contract review process with fewer people.?
With such substantial improvements available today from AI, it has become impossible for me to imagine a future in which people are not regularly using AI to help them review contracts.
As the technology behind these AI tools improves, their capabilities expand. What starts as a tool for initial reviews or simpler contracts will evolve to handle more complex contracts and tasks, thus appealing to more customers. In another post, I mapped 5 levels of legal AI autonomy to describe where we are today and to help identify the yet-to-be-developed capabilities of the future.
Traditionalists might initially dismiss AI tools as inferior or inadequate for the nuanced world of contract review. But by the time they recognize the efficiency and potential accuracy of AI, early adopters will have significantly reduced operational costs and turnaround times, doing more and better quality work at lower cost.
84% of respondents to a recent survey were enthusiastic or cautiously optimistic about the use of AI to assist with contract review.
Part 4: In-House Take Action
For in-house lawyers and leaders, now is a good time to begin considering with renewed vigor the role of technology.
Regardless of your team’s size or stage, whether you’re growing or shrinking, technology should be a first-class priority. It definitely is for functions like sales and marketing.
Here are some of the questions we believe are the right ones to be asking:
Thanks for reading. For more content like this, follow me and follow LegalOn Technologies .
Daniel Lewis So why don’t we set up a FORUM (debate or kaffeeklatche) on why that HASN’T happened (in theory) and FAILSTORMING what could prevent it from happening now. UP FOR DEBATE: -What/whither/how practice has changed over the last decade+ -Whether “disruptions” likely to be all good or bad.
CEO of LegalOn
1 年Would love to hear from Richard Tromans, William Henderson, Jason Barnwell
Helping GTM teams automate outbound with AI
1 年Accessibility is another factor which plays a huge role. Adoption is relatively higher now because of it. We have already seen other knowledge workers experiencing significant productivity increases with just the standard GPT-4. Although in its current state, AI is not best suited for all kinds of tasks, here's a great piece on it- https://www.oneusefulthing.org/p/centaurs-and-cyborgs-on-the-jagged. Organizational changes might still take time but individuals are already reaping the benefits. I am with you, it is different this time around. Thank you for the great read.
Contracts Queen, Inventor of GlobalNDA & Zoey, Avocado, Fashionista & CLM Maximizer. Connecting people. Making law accessible. Humanizing contracts. #LegalSparkle
1 年Not disrupt it, but wiggle it. It will be a slow shimmy, but things will shake up and fall out and everyone will have fun either way.