AI adoption in law: What it will take to make it stick

According to our global research of over 3000 business and technical decision makers, 85% of organisations are now using AI, yet only 17% have operationalised it. That gap tells a familiar story — but also a hopeful one.

AI adoption is not linear

In our recent webinar, The 17% Problem: Why most AI investments aren't delivering and what to do about it, four leaders from across the legal industry gathered to discuss why that gap exists.

One of the most resonant comments during the session came from Lorie Almon, Chair and Managing Partner at Seyfarth Shaw, who described organisations as being “all of the categories at once,” when it comes to AI adoption success. That observation will ring true for anyone working in the legal sector.

AI adoption is not uniform, and it’s unrealistic to expect teams to reach total operationalisation at the same pace. Legal teams experience hot spots: individuals or groups pushing ahead while others move more cautiously. This is the natural shape of adoption in complex professional organisations.

Adoption is iterative, uneven, and ongoing. The organisations that make progress are those that accept this reality and invest in bringing people with them over time — through clear narratives, accessible support, and visible examples of success to learn from and build upon.

Behaviour change is the real work

When AI initiatives stall, it often reflects a broader sector level challenge: the legal profession has limited muscle memory for sustained behaviour change at scale.

It is unsurprising that organisations experience fear, confusion, change fatigue, and uneven engagement with AI. They are predictable responses in a profession historically optimised for certainty, risk management, and precedent rather than experimentation and iteration.

Amanda Raad, Co-lead, Anti-Corruption and International Risk at Ropes & Gray, described AI as a “living, breathing creation” during our webinar, capturing the shift. What the sector is learning is that capability compounds over time. Lawyers need permission to learn with AI. That means acknowledging that value is not immediate, and that experimentation and repeated application are critical.

Reena SenGupta, Executive Director at RSGi, described the current situation as the “frozen middle”. AI is in use, but its application remains tactical. As Alex Smith, Senior Director, Product Management at iManage, observed, this often happens because early adoption occurs at a task level, not a process level. Meaningful value emerges when teams rethink the underlying process itself. Without that shift, adoption plateaus and AI use remains limited.

Behaviour change accelerates when lawyers are supported to reimagine how they deliver their expertise with structured training, time to explore, peer learning, and shared examples of what “good” looks like when deploying AI. 

Information architecture

Alex emphasised the criticality of Information Architecture (IA) for successful AI adoption. The real issue is not simply that poor IA leads to inconsistent AI outcomes (although that’s undeniably true). The deeper problem is missed opportunity. IA is not a back office concern — it is a direct enabler of lawyer value and more differentiated client service.

When information is fragmented, organisations can’t realise what they know. Lawyers lack visibility into the full breadth of their collective experience, and that fragmentation directly weakens the impact of AI.

As Alex noted, applying AI often forces legal teams to confront an uncomfortable question: what can we see in our data — and what can’t we see at all? This has led organisations to reflect on how their information became so fragmented in the first place: likely a combination of light governance, busy lawyers prioritising immediacy over structure, little short term pressure to fix what “mostly works,” and complexity created by growth and scale.

Reena’s observation to “step back” before leaping forward is critical here. Make an intentional choice about where to start. As Alex observed, there are almost always parts of an organisation where information is already better structured. Leveraging what is already working creates momentum and avoids the unrealistic expectation that the entire information estate must be structured before value can be realised.

A unique moment for the profession

The tools gaining traction today engage directly with language, meaning, context, and interpretation. Law is an industry built on words, and AI is meeting lawyers where they operate.

Lorie’s observation that AI capability will soon move from the level of a third year associate to a senior partner highlights why this moment matters so deeply. Rather than diminishing the role of lawyers, AI sharpens it. Human value increasingly concentrates around judgment, empathy, relationship building, strategic advice, thought leadership, and the ability to translate complexity into insight clients can act on.

AI is not just a productivity tool. It is an enabler of a far more client centric model of legal service delivery, forcing the sector to revisit a long standing question: where do lawyers add value?

Hyper focus on clients

Throughout the discussion, one conclusion kept resurfacing: AI adoption succeeds when anchored relentlessly in client value. 

When lawyers can clearly see how AI helps them deliver better advice, greater consistency, or richer insight that clients genuinely care about, adoption stops being a technology or user problem and becomes a professional one.

Ultimately, AI will not transform law because it is impressive technology. It will transform the profession because it enables legal teams to deliver more of what clients value. The organisations that recognise this are not just chasing adoption metrics; they are redesigning legal services around the client and letting AI do what it does best in service of that goal.

An ongoing investment 

AI adoption is a leadership, culture, and design challenge. The firms and legal teams that make progress will be those that treat adoption as an ongoing investment — in people, in information, and in an honest reckoning with where lawyer expertise is irreplaceable and where AI can genuinely augment it. 

Want to learn more? Watch the on-demand webinar here. And be sure to check out our new AI Confidence™ assessment, which benchmarks your organization's maturity on the curve separating AI experimentation from AI impact.