Most of the pressure law firms feel on GenAI arrives from the outside. This series has argued that the response has to be built from the inside.
The GenAI strategy sequencing conversation most law firms need to have is not technically complex.
It is politically expensive.
By now, most firms have approved tools, launched pilots, written policies, and invested meaningfully in AI capability. On paper, this looks like progress. In practice, many leadership teams feel something less settled: the investments don't quite add up to a clear direction.
That unease is not a technology problem. It is an AI governance problem.
Over the previous blogs, I argued that most firms are running three strategies simultaneously — Trust First, Capital Discipline First, and Competitive Positioning First — and that capital allocation has usually already sequenced them, even without an explicit decision. Strategy sequencing is inevitable. The only question is whether it is deliberate.
Why the Conversation Hasn't Happened
Strategic sequencing forces exposure.
It exposes that some investments were reactive — driven by client guidelines, lateral competition, or vendor momentum rather than deliberate choice. It exposes that external messaging may be ahead of internal capability. It exposes that certain partners have shaped adoption more than any formal strategy has.
Most uncomfortably, it exposes that capital allocation has already answered the strategic question — just without anyone explicitly agreeing to the answer.
In a partnership structure, no constituency is irrational. Risk leaders are right to protect the firm's reputation. Finance leaders are right to question growing technology spend. Practice leaders are right to worry about market perception. Sequencing inevitably privileges one pressure first — which means someone feels temporarily constrained. Avoiding the conversation preserves internal equilibrium. It also preserves ambiguity.
In strong markets, ambiguity is affordable. In tighter ones — where clients question AI-related billing and demand visible efficiency gains — ambiguity becomes expensive to sustain. That is where most firms now find themselves.
Step One: Build the Inventory Before the Meeting
The sequencing conversation fails if it starts from opinion. It works when it starts from facts.
Before any leadership discussion, someone — usually the COO or CIO — needs to produce a one-page inventory of every GenAI-related commitment made in the past 24 months: licenses, pilots, governance initiatives, training programs, infrastructure upgrades, internal time spent, and what the firm has said publicly in pitches, on its website, and in recruiting conversations.
Most firms are surprised by what surfaces. The total investment is almost always larger than anyone remembered. The gap between external communication and internal reality is almost always wider than anyone acknowledged. And the pattern — which categories grew fastest, which got protected, which stalled — almost always points somewhere.
What makes this harder than it sounds: most firms don't have this information in one place. That coordination effort is itself diagnostic. If it's difficult to assemble a clear picture of what's been spent, that difficulty is part of what the conversation needs to address.
Step Two: Name the Pattern, Not the Problem
Once the inventory exists, the conversation has one job: not to debate which strategy is correct, but to name which one the spending reveals.
"What should our GenAI strategy be?" invites advocacy and entrenches positions. A better question is: "What does our investment pattern tell us we have been optimizing for?" That reframes the conversation from prescriptive to diagnostic. It makes trade-offs visible without requiring anyone to concede their priority was wrong.
One pattern worth naming explicitly: some firms haven't chosen their sequencing at all — it was chosen for them. A major client sent updated outside counsel guidelines. A procurement team asked pointed questions about AI efficiency. A lateral hiring competition forced a public commitment. In each case, the firm ended up in Competitive Positioning not because leadership made a proactive bet, but because an external force required a response. A reactively-sequenced firm has less flexibility than one that chose deliberately. Naming that honestly is what makes it possible to regain control of the direction.
What makes this harder than it sounds: the people in this conversation have competing incentives to interpret the inventory favorably. Someone needs to hold the frame — to keep the conversation diagnostic rather than defensive. In a partnership, that role is harder to play than it looks.
Step Three: Agree on Three Things, Not Everything
Full alignment across a large partnership is neither achievable nor necessary. What is necessary — and achievable — is agreement on three things that cannot remain local decisions.
The governance floor. What is permissible firm-wide, regardless of which tools a group uses or how advanced their adoption is? One page, not twenty.
The client narrative. What can any partner say when a client asks how the firm's AI investments benefit them? It needs to be consistent, honest, and defensible across every practice group — not because every group is doing the same things, but because the firm speaks with one voice on what those things are in service of.
The investment logic. On what basis will the firm expand, pause, or redirect GenAI spending going forward? Clear enough that when the next tool request arrives, there is a shared basis for the decision rather than another negotiation from scratch.
Everything else can flex. The matrix of different adoption speeds across practice groups is not the problem. The problem is the matrix without a frame.
What makes this harder than it sounds: this requires the managing partner to hold a line on what is non-negotiable at the firm level — in a governance structure where very little formally is. The pushback is legitimate and needs to be heard. But it needs to be held against three structural realities: client conversations are firm-wide, cost structures are shared, and reputational exposure doesn't stay local.
The Variable Governance Can't Control: People
Even with clarity on sequencing, one variable resists formal control.
In every firm I observe, GenAI adoption is being shaped less by strategy and more by a small number of individuals — the partners who are genuinely curious and become internal champions, and the partners who are genuinely skeptical and become quiet blockers. Both are often more influential than any governance decision.
This plays out most visibly in the litigation versus transactions divide. In some firms, transaction is leading — high-volume drafting and structured processes make AI gains immediately visible and defensible. In others, litigation is leading — partners handling complex matters have found AI genuinely useful for research and preparation, and their early adoption has pulled the firm forward. Leadership in both cases tends to attribute the difference to the nature of the work. What's usually closer to the truth is that the people who made it work found a way, and the people who haven't yet haven't found — or been given — the conditions to do so.
The practical implication: the sequencing conversation needs to include not just which strategy governs first, but who is being asked to carry it. A strategy that asks influential GenAI champions to slow down will stall. A strategy that asks reluctant skeptics to go public before they're convinced will fail differently but equally. Knowing who your champions are — and who your credible skeptics are — is as important as knowing which strategy is leading.
In a partnership, people don't implement strategy. They are strategy.
The Harder Truth
Naming what the spending reveals requires acknowledging that some of it wasn't intentional. Agreeing on a client narrative requires admitting the current one may be inconsistent. Holding the governance floor requires spending political capital on something that produces no immediate revenue. And recognizing that people shape strategy as much as formal decisions do requires giving up some of the comfortable fiction that direction flows purely top-down.
None of that is impossible. Firms do it.
But it rarely happens until someone decides that the cost of continued ambiguity exceeds the cost of clarity. That decision cannot be delegated to a committee, outsourced to a vendor, or solved by another pilot.
Governance begins when leadership decides that ambiguity is no longer affordable.
Everything else — frameworks, policies, pilots — follows from that decision.