Every managing partner I talk to has tried the same thing.
The managing partner calls a meeting. The agenda: "Where can we use AI?" Senior associates and practice leads sit around the table. They're asked to identify which parts of the firm's work AI could handle.
The room gets quiet. People offer safe answers. Maybe document formatting. Maybe some light research assistance. Nobody mentions the 40-hour contract review process that an AI tool could reduce to six hours, or the annual audit workflows that now run continuously with minimal human oversight.
They know. They're just not going to volunteer the parts of their own work that might disappear.
This is about incentives, not motivation
Your senior associates are smart people who act in their own rational self-interest. That's what smart people do.
Ask a senior litigation associate to identify which parts of the litigation workflow AI can replace, and you're asking them to produce a list of reasons their role should be smaller. Ask a practice lead to map which client deliverables could be automated, and you're asking them to argue against their own headcount. Ask your best recruiter which sourcing tasks AI handles better, and you're asking them to make themselves less essential.
Nobody does this honestly. The incentive structure makes honest answers professionally dangerous.
The team isn't holding out on you. The incentive structure is.
McKinsey's 2025 State of AI report found that workflow redesign, not tool adoption, was among the strongest contributors to business impact across 31 variables tested. That's what makes it hard: the thing with the most impact is exactly what nobody in the room will volunteer. TSIA, the industry research body for professional services, named this dynamic directly: the cannibalization dilemma. "The more efficient you become, the more revenue you risk losing" under hourly pricing. The firm knows efficiency erodes its own revenue. The team knows efficiency erodes its own headcount. Nobody volunteers to accelerate that process.
The real conversation is a different one
The business model question — what should your firm sell differently because of AI — is a leadership question. It requires looking at your revenue model, pricing structure, delivery model, and client relationships together.
That conversation can't happen in a room full of people whose jobs depend on the current model. It has to happen at the leadership level. But that's not enough on its own. Managing partners and practice leads built the current model. They've succeeded inside it. The incentive problem doesn't vanish when you close the door to senior staff — it just changes shape.
This is why outside facilitation matters for this specific problem. Not because your team lacks intelligence. Because the managing partner can't ask their own partners to price their practices against what AI now does — not without someone else in the room holding that question. Someone who doesn't have a career riding on the answer. The facilitator makes it safe to ask what the partnership structure makes dangerous.
What the outside conversation looks like
It doesn't mean hiring a consultant to tell you which AI tools to buy. Your team can answer that question on their own.
It means putting someone in the room who has no stake in the current business model. Someone who can look at your revenue streams and delivery model, and ask the questions your team can't ask themselves.
Questions like: if AI handles 70% of the process work in this practice area, what's the remaining 30% worth to clients, and can that sustain the current team size? If billing shifts from hourly to fixed-fee or retainer, what happens to partner compensation?
They get more expensive the longer they go unasked.
The cost of waiting
Every quarter you wait, the cost compounds.
Competitors who move first set the terms. Your clients who are pulling work in-house get more comfortable doing it themselves. Client expectations on price keep shifting. And your team, seeing no clear direction from leadership, quietly runs their own math on what AI means for their future at the firm.
The conversation is happening whether you lead it or not. It's happening in hallways, in one-on-ones, and in your best people's private assessments of whether to stay or go.
RSM's 2025 Middle Market AI Survey found that 53% of firms that had already adopted generative AI described themselves as only "somewhat prepared" to do so. A year or more into adoption, more than half still haven't had the strategic conversation that would make them ready. That gap doesn't close on its own. The conditions for it have to be created.
Without someone in the room who has no stake in the answer, you keep getting the safe meeting. Document formatting. Scheduling. Light research assistance. The real conversation never happens.