Six months ago, the client was not interested. The firm had proposed using AI in part of an IP matter, and the answer was no.
Six months later, the same client came back with a different request: reduce the fees.
The technology had moved. The client had moved too.
That same tension was noticable throughout Iltacon. Client expectations around AI are moving quickly. A firm can be moving faster than it ever has before and still find that client expectations are moving even faster.
Law firms have good reasons to be careful with AI
There is nothing wrong with diligence. Law firms deal with confidentiality, conflicts, client commitments, professional responsibility and security, and new technology has to fit inside all of that. So when a firm takes time to assess an AI tool or a new way of working, there are very good reasons for it.
However, client expectations are not standing still while that happens. Clients are learning what AI can do in their own organizations, seeing work happen faster elsewhere, and changing how they think about cost and productivity.
That shift doesn't announce itself. It shows up quietly, for example, in a fee conversation six months after the firm thought the question was closed.
What the firm actually did with that AI opening
When that IP client came back, the firm didn't respond by discounting the existing process. Instead, it went back to the workflow itself. It mapped where the attorneys' time was actually going and looked for the point where the most value was leaking out.
Carey Jordan of Vorys described what happened next during an excellent ILTACON session, AI-Powered Practice Group Transformation Journey.
The answer turned out to be drafting, specifically what Carey called "draft spin": the cycle of an associate producing a draft, a senior attorney reworking it, and the associate not quite absorbing why.
It was expensive and slow. And, importantly, nothing on the market addressed it directly, because it isn't really a drafting problem. It's a judgment-transfer problem.
So Vorys built something new through its collaboration with Stanford Law School's liftlab. Working with the researchers, the firm tried to externalize what "good" looked like in the heads of its most experienced attorneys — not simply a set of rules, but the patterns of judgment a senior attorney applies almost instinctively.
The method that worked best wasn't asking attorneys to describe their reasoning. Instead, they were shown pairs of drafts and asked which one was better, using those preference rankings to reconstruct the judgment behind their choices. According to the presentation, that approach turned out to be roughly seven times more effective than direct interviews at capturing something usable.
The result was a tool that captured judgment. It reviewed associate drafts the way that specific senior attorney would, flagging what to change and why, with the reasoning visible rather than just a redline.
Associates could learn from it instead of simply receiving corrected copy. It didn't just speed up drafting; it shortened the distance between a junior associate's instincts and a senior attorney's. It addressed the root cause of "draft spin", the part that was costly for the firm and its clients.
AI improvement has to show up in the client experience
The ILTACON presentation largely ended there. But for me, that left an important question: What has changed for the client?
After all, the original trigger for the work was not a request to build a better drafting tool. The client wanted the fees to come down.
As AI allows a firm to improve how work gets done, clients will reasonably expect to see some of the benefit. That doesn't automatically mean lower fees — it could mean faster turnaround, better consistency, earlier insight, more proactive advice, a different pricing model, or a service that was previously too expensive to offer at all.
Client expectations are shaped by outcomes, not internal effort
A firm may spend months evaluating tools, working through security questions, developing policies, and training people. That work matters, but none of it is visible to the client.
The client experiences the service — whether the answer arrived faster, whether it was better, whether the firm caught something before being asked to.
A firm can therefore make enormous internal progress on AI without the client seeing much difference at all.
That gap between internal effort and external experience is exactly where client patience can run out. It is also why legal AI adoption and a sophisticated law firm AI strategy do not automatically translate into client value.
Why being proactive about AI matters
Waiting for the client to ask for lower fees is one way to discover that expectations have changed. It is probably not the best one.
A stronger position is to go to the client first: What do you need? And when that is difficult to answer, the firm can bring ideas to the conversation: here is what we would like to change about how we do this work, here is where we think it could improve the outcome, and here is what we're still testing.
That conversation is about more than technology. It's about the relationship.
It turns AI from something the firm is implementing internally into something that can strengthen the relationship externally, on the firm's timeline instead of the client's.
It also gives the firm a better chance to understand client expectations before they show up as pressure on fees, turnaround times or service levels.
The challenge is not simply to move faster on AI
Law firms should not abandon diligence, and not every new capability needs to become client-facing immediately.
The real challenge is getting better at separating the decisions that genuinely require caution from the ones where the firm can safely learn by doing: run the pilot, test it against real work, measure the result, stop what doesn't work, expand what does, and keep asking what changed for the client along the way.
A firm may be adopting AI quickly. It may be training hundreds of people. It may have strong governance, good tools and an increasingly sophisticated innovation function. All of that matters.
But ultimately, client expectations will be shaped by what those investments change about the service they receive.
The client should be able to answer a much simpler question:
What has become better?
That is where AI investment starts to become client value — and where progress becomes more important than speed.
Rise Above AI Chaos: A Business Fable About Leading Organizations Through the AI Revolution