A disruptor's framing of the AI squeeze on Irish legal fees says merge with a giant or spend like Kirkland or get caught in the middle. The threat is real. The only-two-options answer is wrong for most Irish firms.
The Irish Independent ran a piece this week on AI forcing a reckoning in Irish legal pricing. The headline argument, made by the head of a newly merged transatlantic firm setting up in Dublin, is that clients now see what AI does in their own businesses and are demanding the same efficiency in their legal bills — and that this will break the business model of firms caught between elite and multi-jurisdictional.
The diagnosis is largely right. The prescription deserves a second look.
What the article gets right
Three things in it are simply true, and Irish firms that pretend otherwise are in trouble.
Clients are repricing legal work. Once a general counsel has watched AI compress a task in their own operation, the firm charging six hours for something that now takes one has a credibility problem. Value-based and outcome-based pricing stops being a partner-retreat talking point and becomes a client demand. That pressure is here, and it accelerates.
The cost of the technology is real and rising. The frontier legal AI platforms are not being sold at what they cost to run. When that changes — and US firms committing nine-figure sums to build their own tooling tells you the direction — the price of "keeping up" by procurement alone goes up sharply.
And the firms most exposed are the ones in the middle: too big to be nimble, not elite enough to command a premium, not multi-jurisdictional enough to follow clients abroad. That is an honest read of a real squeeze.
What it leaves out
Where the argument turns convenient is the menu of responses. As told, there are essentially two: merge into a large international platform, or out-spend the field on technology. Both, conveniently, are the options that favour large international platforms and the AI vendors selling the most expensive tools. The firm making the argument is itself a newly merged $1.5bn platform. That does not make it wrong. It does make it partial.
The unstated premise is that AI advantage is bought — that the firm with the biggest balance sheet and the priciest platform wins. For the very top end of bet-the-company M&A and complex cross-border work, there is something to that. For most of the Irish legal market, it is not how the efficiency that clients can actually see gets produced.
The work a client notices — faster turnaround, fewer hours on routine matters, responsiveness that used to require another hire — rarely comes from a frontier reasoning model. It comes from automating the repetitive, structured work that sits around legal work: intake, conflict checks, document assembly, deadline triggering, client communications, the chase-and-reconcile of compliance cycles. None of that needs a six-figure platform licence. It needs someone to find where the hours actually go and build the fix inside the systems the firm already runs.
The arms-race framing is the expensive way to lose
Here is the trap. A mid-tier firm reads that it must either merge or spend, concludes it cannot out-spend Kirkland, and either panics into a defensive merger or freezes. Both responses cede the actual ground.
We have done this work in an Irish firm. A debt-recovery practice was about to add headcount to absorb a 33 per cent jump in volume from a new contract. The audit found that most of the work driving the hire did not require a solicitor at all — court-area lookups, conflict checks, batch postal runs, figure re-verification, automated chasing. We built the automation inside the firm's existing Microsoft 365 environment. The headcount never had to happen. No frontier platform was involved. The client-visible result — capacity to take on a third more work without raising fees — is exactly what the pricing pressure demands.
That is the path the two-option framing erases: targeted, independent AI adoption that captures most of the value without a capital arms race. It is available to a twelve-partner firm, not only a fourteen-hundred-lawyer one.
What this means for an Irish firm deciding what to do
The honest position is not "ignore the disruption." It is real, and the firms that do nothing will be repriced by their own clients. The position is that the response is a strategy question, not a balance-sheet one.
Three things are worth doing before signing up to anyone's platform — or anyone's merger.
First, find where the hours actually go. Most firms have never audited it properly. The cost is not where partners assume, and the automatable share is usually larger than expected. The audit, not the tool, is the starting point.
Second, separate the work that needs frontier AI from the work that needs a workflow. Far more of the client-visible efficiency sits in the second category, where the cost of capture is a fraction of a platform subscription. Decide build, buy, or configure on the merits — not on a vendor's roadmap.
Third, get the governance right in parallel, because the regulator and the Law Society are not waiting for the market to settle. Confidentiality, privilege, and the EU AI Act apply whether the AI was bought from Harvey or built in-house. We developed two of the practitioner tools on the Law Society of Ireland's own AI resources; the obligations are specific and they are already live.
The disruption is coming, and the article is right that some Irish firms will not survive it. But the ones that do will not all be the biggest. They will be the ones that worked out, precisely and independently, where AI earns its place in their practice — and built that, instead of buying the most expensive answer to a question they had not yet diagnosed.
Acuity AI works with Irish law firms on exactly this: AI governance for law firms, legal process automation, and AI training — independent, senior-led, with no platform to sell. See the work, or talk to us.