The people who can see its risks most clearly aren't in the room.
AI adoption in law firms is being driven by leadership and technology teams. The people who manage risk, compliance and finance are being consulted last. That is not just a cultural problem. It is an operational one.
A few weeks ago, I was speaking to a managing partner at a large firm about AI adoption. She was worried about how to communicate the changes ahead, especially to support staff.
"The challenge," she said, "is that they're the people most likely to lose their jobs."
I understood the concern. But it revealed something more dangerous than she intended. She was already thinking about these people as recipients of change, not participants in it. And that instinct, however well-meaning, is exactly how firms are walking into avoidable problems.
The governance gap
Most law firms now have some version of an AI strategy. Many have steering committees, task forces, innovation leads, and working groups. What they rarely have is a single partner willing to be personally accountable for the outcome.
Without accountability, language drifts quickly from “me” to "we." Difficult questions about who absorbs short-term disruption, who carries the early reputational risk, and who owns the consequences if something goes wrong remain unresolved. The result when partnership governance meets a decision that doesn't fit neatly into consensus: working groups that can't decide, steering committees that steer nothing, and strategy documents that everyone endorses but nobody owns.
Into that vacuum, AI decisions get made anyway. Tools get piloted. Workflows get redesigned. And the people who actually understand how those processes work – the legal cashiers, COFAs, compliance officers, finance managers and practice directors – are consulted afterwards, if at all.
What this actually costs
This is an operational risk with measurable consequences because the people excluded from these conversations are precisely the ones who can see where AI creates compliance exposure. They know that automated billing tools need to be reconciled against SRA accounts rules. They know that AI-assisted document handling can create client money risks if controls haven't kept pace. They know that faster is not always safer, and that in a regulated environment, the cost of getting something wrong is rarely confined to one team.
They understand, in ways that don't always make it into strategy presentations, that faster billing doesn't automatically mean faster collection. Automating a step in the process doesn't resolve the judgement calls that sit behind it. Speed without proper oversight doesn't eliminate the problem. It just moves it downstream.
These are not abstract concerns. They are the operational realities that legal finance and compliance professionals navigate every day. When AI is layered on top of existing workflows without involving the people who understand those workflows, firms don't just risk errors. They risk making the same mistakes faster.
Why it keeps happening
Partnership governance evolved to protect equality: consensus decisions, shared risk, measured change. It worked well when partners were doing broadly similar work at broadly similar risk. That’s no longer true. A corporate practice facing AI-disrupted transactional work is not in the same position as a family law team. An early adopter carries reputational and client risk that a cautious partner doesn't face yet. But the governance structure treats them as though they are similarly positioned.
Change moves at the speed of the most cautious voice. And the people who could most usefully intervene – the operational professionals who can distinguish genuine progress from badly governed haste – are not in the room where decisions are being made.
What ILFM members should do about it
The right response is not to wait to be told what AI means for your role. It is to step into the conversation from a position of professional strength. That means being willing to ask three questions, not as a technologist, but as someone whose professional responsibility is to spot where risk sits and where controls need to remain, even if they evolve:
- Where could AI remove repetitive work that genuinely adds little value, and how would we verify that the output meets current, manual standards?
- Where does AI introduce new compliance exposure (around client money, confidentiality, accounts rules, or reporting accuracy) that hasn't yet been designed for?
- And where must human judgement remain explicit, visible and accountable, not as a default, but as a deliberate governance choice?
These are not obstructive questions. They are the questions that protect firms from implementing AI in ways that look efficient on paper and create serious problems in practice.
The firms that navigate this period best will not be those with the loudest AI strategy. They will be the ones that combine speed with judgement, experimentation with proper controls, and the confidence to involve the people who understand how the firm actually operates.
The people who can see most clearly where AI will go wrong in law firms are the ones managing compliance, finance, and operational risk. In most firms, they are the last ones invited into the conversation. That needs to change. And in many firms, the people who can change it are reading this article.
Greg Orme is an award-winning author, top team facilitator and leadership advisor. His book The Human Edge won Business Book of the Year 2020. He works with senior leaders across professional services, helping organisations navigate change, build leadership capability, and respond to AI and digital disruption.
Comments