AI Is Buying Itself Into Legal Tech. The GTM Problem Is Still Human

The adoption numbers look extraordinary.

Corporate legal AI adoption more than doubled in a single year — from 23% to 52% according to the ACC and Everlaw's GenAI Survey of 657 in-house legal professionals across 30 countries. Thomson Reuters found active generative AI use climbing from 14% to 26% year over year. Every major investor concluded the same thing: the legal AI market had arrived.

The GTM problem had not.

What the Adoption Data Doesn't Show

Here is what the 52% adoption figure does not reveal:

• How many of those legal departments are running three tools simultaneously, with no intent to standardize on any of them
• How many pilots were launched by legal ops directors who still need to survive a seven-person procurement committee before a contract can be signed
• How many of that 52% are at the evaluation stage rather than the production stage — actively assessing, not actively using

A Law360 analysis in mid-2025 found that a significant portion of revenue currently booked as ARR may not be truly recurring. Law firms are piloting multiple tools simultaneously — in many cases without any intention of committing to all of them long-term.

The Paradox

The higher the adoption headline, the more sophisticated and skeptical the buyers become.

64% of in-house teams now expect to depend less on outside counsel because of AI capabilities. That is not an adoption barrier. It is a signal that buyers are developing real opinions about what AI should and should not do in legal work. They have run the pilots. They have seen the hallucinations. The buyers who were impressed by the technology in 2023 are evaluating it on outcomes in 2025. That is a harder conversation.

What the Winning Companies Did Differently

The companies that converted investment into durable traction in 2025 did not win because their AI was better.

Harvey reached $100 million in ARR by solving the trust problem before scaling the marketing problem. Its reference account strategy — Allen & Overy, Paul Weiss, PwC before moving downstream — was a GTM decision, not a product decision. The AI was table stakes. The trust infrastructure was the moat.

EvenUp went deep on personal injury and reached a $1 billion valuation by understanding the workflows, the billing structures, and the specific failure modes of personal injury practices at scale. Domain understanding shortened cycles. The AI was the delivery mechanism. The trust was what opened the door.

The pattern is consistent: vertical depth + reference accounts + procurement-ready infrastructure — not AI feature differentiation.

The GTM Problem Remains Human

Legal buyers do not make decisions based on capability demonstrations. They decide:

1. After trusted recommendations from peers who used the product in their actual workflows
2. After security review confirming the data handling meets client confidentiality standards
3. After a budget process involving people who weren't in the demo, and need the case made without feature references

"The industry will need to continue to demonstrate strong exits to substantiate the amount of early-stage capital already invested in the space." — Gordon Crenshaw, The LegalTech Fund

The AI is buying itself into every legal organization. The problem that remains is stubbornly human: trust, credibility, reference accounts, and the organizational coherence to support a buyer through a process that no amount of model performance can accelerate on its own.

If you find this content insightful, consider subscribing to our LinkedIn Newsletter and following our company page for our latest articles!