Legal AI Still in 'Caveat Emptor' Phase

In the early 2000s when Google was first trying to introduce local businesses to AdWords, it held in-person “Google U” seminars around the country. Google speakers at these events likened paid search to “directional advertising” in the yellow pages. In those days, online marketplaces, directories and ad platforms often sized their revenue opportunity by looking to the then $15 billion print directory industry.  

One of the top yellow pages revenue categories was legal. More than twenty years ago you’d find dozens of small law firms spending hundreds of thousands of dollars annually buying full page ads in multiple books to cover a large metropolitan market such as Los Angeles. Lawyers eventually moved away from print and entered the digital marketing melee. Now it’s one of the most competitive local advertising segments online. 

$400+ Billion in Revenue

There are a range of estimates for the value of the legal industry in the United States and globally. Various analysts put it somewhere above $400 billion in the US and more than $900 billion globally in 2024. The US figure includes revenue generated by law firms, in-house legal departments, legal consulting, legal technology and various other services. While Dialog's initial small business survey included some law firms, the legal respondent pool was not large enough to break it out as a stand-alone category (we rolled it into a presentation also featuring Business & Professional Services and Healthcare).

But a relatively new "State of AI in Legal" report from Ironclad offers a look at AI adoption across the legal industry in the US. The survey results capture data from 400 in-house legal teams and 400 law firms. The findings were published in Q1 this year, so we can probably assume that adoption has increased somewhat since then.

AI Adoption and Trust

Despite general concerns about accuracy and "hallucination," the survey found AI is increasingly trusted and being used by legal professionals. As the graphic indicates, 74% of those surveyed are already using AI tools and 92% of them say AI has improved their work.

Nearly all corporate, in-house attorneys surveyed (96%) said their companies allow AI use. That compares with 74% of law firms. But when it comes to actual adoption, "84% of in-house respondents [said] they use AI tools for legal work, versus 60% at law firms." One reason AI adoption may be greater among in-house teams is that time is a cost for them, but it's typically a source of revenue for law firms (i.e., billable hours).

The picture isn't entirely one of unchecked enthusiasm, however. There's still a fair amount of ambivalence. Roughly 50% of respondents indicated a moderate level of trust, while 21% had a high level of trust in AI. About 29% were neutral or untrusting. Law firms are more concerned with information accuracy and security issues than in-house teams. As a consequence, use and adoption trails.

The in-house respondents are more positive overall. Again this may be partly attributable to "business model" differences around time and efficiency. There may be some ambivalence among law firms about AI tools reducing billable hours.

The report speculates that law firm resistance may also be partly about fear of getting sued for malpractice based on some AI error or mistake. And they're also concerned about how AI could affect client confidentiality and privilege issues.

AI Legal Use Cases

Notwithstanding these concerns, across both respondent populations, AI appears to be improving their work. When asked, "Has using AI tools for your job improved or worsened how you work?," 32% indicated it has significantly improved their work and another 60% said their work had been somewhat improved.

The chart below highlights the specific areas where respondents believe AI is/would be most trustworthy and could be deployed (the question is conditional but the answers probably reflect current usage).

AI excels (or could excel), according to the report, at performing repetitive tasks, such as tagging metadata in contracts, flagging risky clauses, and certain types of contract analysis. Tagging metadata in contracts for later search and review, the top use case, could help significantly streamline document management.

Beyond contract-related tasks, AI is being used for research: 50% of respondents indicated that AI helps them do better, faster, and more in-depth research. More broadly, AI is saving time and helping lawyers "offload mundane tasks," which is a general selling proposition and use case for AI across industries.

The report also speculates that AI might be a remedy for lawyer burnout, by helping reduce "task overload." (The report offers an overly upbeat assessment of attorney job satisfaction.) The survey asked, “In which of the following areas do you believe AI could help you the most?” The chart below captures the responses. Looking at those responses, we can imagine a SaaS-based “virtual legal assistant” handling many of these.

Market Size and Growth

The global legal SaaS market is substantial and growing. It was valued at $8.5B in 2023 with analysts suggesting it could reach $14.5B by 2028. Much of the growth will be driven by SaaS tools that imbed AI and new AI-centric applications for lawyers. Some of the companies now operating in the segment include:

• Ironclad. Focused on contract lifecycle management (CLM). Based in San Francisco, the company has raised more than $334 million and was valued at $3.2 billion in its latest funding round in 2022.
• Evisort. Specializes in contract management and analytics. Based in San Mateo, California, it raised $155.6 million. Its most recent valuation is more than $500 million.
• Luminance. Performs document review and legal analysis. Based in London, UK, Luminance has raised over $100 million and is valued at roughly $200 million as of its last funding round.
• Casetext. Acquired by Thomson Reuters, it specializes in legal research and drafting. Based in San Francisco, Casetext raised $64 million before being acquired for $650 million in 2023.
• LawGeex. Specializes in contract review automation and is based in Tel Aviv. The company has raised around $45 million.
• Blue J Legal. Specializes in predictive legal analysis. Based in Toronto, Canada, it has raised approximately $10 million.
• LegalLytics. Specializes in legal marketing analytics and is based in Tucson, Arizona. Funding information isn't publicly available.

The 'AI-Native' Law Firm

The emergence of what we might call an "AI-native" law firms is inevitable. These firms, whether large or small, will deeply integrate AI into daily operations and client interactions. Done right, AI will make these firms more efficient and competitive than those choosing to avoid or stay away from AI solutions – though caution is not unwarranted in some legal contexts.

Applying AI solutions end to end – from client acquisition on the front end to collecting professional fees on the back end – will take time and discipline and won’t happen overnight. Applying AI to each aspect of a law firm's activities will require careful orchestration and coordination. And it probably take will take a "different kind of lawyer" to fully seize the opportunity. 

These AI-native law firms should experience considerable efficiency gains by training AI to handle mundane tasks, basic research, client onboarding, managing some client interactions and traditional back-office functions, such as billing and collection. If these efficiencies can be realized, the AI-native law firm can potentially improve work-life balance for lawyers and paralegals and enable them to focus more on clients and case management. (This is always the promise of technology – automate mundane tasks, leaving more time for "what matters" – which often doesn't materialize.)

You can think of these AI-native firms as insurgents in the $435 billion legal industry. Insurgents prey on the margins that incumbents extract from clients and customers through legacy models. While we wouldn’t expect these AI-native insurgents to gain substantial market share in the near term, they are likely to steadily nibble away. Later, those nibbles could turn into full-fledged meals. We're likely to see market share changes in 3 - 5 years.

The caveat is that the legal profession in the US is highly stratified, though intensely competitive, which is an "x variable" in this equation. Regardless, law firms will likely feel pricing pressure from leaner competitors, using AI, at every level of the profession.

Large and Small Firms Will Benefit

For large firms, AI should offer a path to additional scale. By integrating AI into their operations, large firms can handle more complex cases that require extensive document analysis and research. By shifting mundane tasks from junior associates and paralegals, they could potentially generate more high-value billable hours spent on substantive legal work. 

For smaller practices, AI can also be a game-changer and probably more so. It will allow those firms to better compete with larger rivals by automating time-consuming tasks such as research (not without risk), contract analysis and document review, as mentioned. This means, hypothetically, that small firms could take on more cases, have faster turnaround times, and potentially compete on fees more aggressively while maintaining profitability.

Since small practices often operate with limited staff, AI can serve as a force multiplier. By automating tasks, AI allows these firms to handle a larger caseload without hiring more lawyers, paralegals and admins. This leads to reduced costs and allows small firms to be more agile and responsive.

Key Questions Remain

AI is poised to play a pivotal role in legal services, as it will or already is in other industries. Its primary appeal today is boosting efficiency and productivity, and supporting more informed decision-making. However, in an ethical-rules based industry such as law, successful integration of AI will depend on addressing the previously mentioned data accuracy and privacy concerns.

The rollout of AI will not be painless and likely come with a heavy dose of anxiety and tension for many in the profession, especially older practitioners. How will the benefits from the new efficiencies be distributed? Will the law firms pass them on to clients in the form of reduced fees? Will these new efficiencies mean fewer associates hired? How will billable hours be impacted?

With an assist from ChatGPT, here are some of the key questions surrounding the current and future AI rollout in legal:

1. Ethical Considerations
   • How do we ensure that AI is used ethically in legal practice?
   • What safeguards need to be in place to prevent biases in AI-driven legal decisions?
   • How can client confidentiality be maintained when using AI tools, especially those that involve cloud-based processing?
2. Accuracy and Reliability
   • How accurate are AI tools in performing tasks such as legal research, contract analysis, and case predictions?
   • What protocols should be established to verify and cross-check AI-generated outcomes?
   • To what extent should human oversight be maintained to ensure the reliability of AI-assisted legal work?
3. Impact on Legal Jobs
   • How will AI affect employment within the legal industry, particularly for junior associates and paralegals?
   • What new skills will legal professionals need to develop to work effectively alongside AI tools?
   • Will AI create new roles within law firms, such as AI specialists or legal technologists?
4. Regulation and Compliance
   • What regulatory frameworks need to be established to govern the use of AI in the legal sector?
   • How will AI use in legal practice comply with existing laws and regulations, such as those concerning data privacy and security?
   • What role will government and professional bodies play in regulating AI technologies in law?
5. Client Interaction and Trust
   • How will clients perceive and trust AI-driven legal services?
   • What transparency measures should be adopted to explain AI's role in legal decision-making to clients?
   • Will clients demand human oversight or validation for AI-generated legal advice?
6. Access to Justice
   • Can AI help increase access to legal services for underserved populations?
   • What are the risks of creating a "digital divide" where only those with resources can access advanced AI-powered legal services?
   • How can AI be leveraged to provide affordable and equitable legal services?
7. Intellectual Property and Ownership
   • Who owns the intellectual property generated by AI tools, such as automated contract drafts or legal research summaries?
   • What are the implications of using AI-generated content in legal arguments and briefs?
   • How will intellectual property laws evolve to address AI contributions in the legal industry?
8. Liability and Accountability
   • Who is liable if an AI tool provides incorrect legal advice or analysis?
   • How will accountability be determined in cases where AI errors lead to legal malpractice?
   • What legal precedents and frameworks will be established to address AI-related disputes in the legal context?
9. Data Security and Privacy
   • How will AI tools handle sensitive legal data while ensuring compliance with privacy laws?
   • What security measures need to be implemented to protect against data breaches involving AI systems?
   • How will the storage and processing of legal data by AI systems be regulated?
10. Future of Legal Education and Training

• How will legal education need to adapt to prepare future lawyers for a profession increasingly influenced by AI?
• What role will AI play in continuing legal education and professional development?
• How can law schools incorporate AI training into their curricula to better equip students for the evolving legal landscape?

These questions highlight the complexity (and opportunity) of bringing AI into the legal industry, as well as the need for careful consideration and planning to ensure AI is used effectively, ethically and in a way that benefits both lawyers and their clients.