Something is happening in the UK legal industry. The number of UK law firms doing residential conveyancing fell by nearly 900 between 2021 and 2026, yet transaction volumes grew in both of the last two years. The market is consolidating, with volume players taking share and smaller firms being squeezed out. But the speed of change and the nature of what is squeezing the middle is all new.

The pressure is already here

The structural challenges facing smaller UK law firms are not hypothetical. They are present, documented and accelerating across every practice area.

Legal aid spending in England and Wales has been cut by 28% in real terms since 2012, a reduction of £728 million. The Ministry of Justice's own provider survey found that four in ten civil legal aid providers intend to actively leave the sector within five years. More than half already make a loss or break even on legal aid work. In parts of the South West and East Midlands, legal aid "deserts" (areas with no local provider) are already a documented reality.

In personal injury, the whiplash portal reforms removed an entire category of routine, predictable work from smaller firms almost overnight. The Law Society has formally stated that the reforms impacted access to justice. The employment tribunal system is carrying a backlog of 515,000 open cases, with the average wait from submission to hearing now running at twelve to eighteen months, and the Employment Rights Act 2025 is expected to push claims volumes significantly higher still.

Anti-money laundering compliance, which applies across conveyancing, commercial property and increasingly probate, now runs to more than 1,500 pages of guidance putting a huge administrative burden on to small and mid sized firms without a dedicated compliance function. And the SRA are not interested in excuses - issuing 73 fines totalling nearly £1 million for AML breaches in 2024/25 alone. A third of firms inspected were found non-compliant. For a small firm without dedicated compliance resource, this is not an inconvenience. It is an existential risk.

None of these pressures is necessarily new in isolation but it sets out a backdrop of uncertainty, exactly at the moment when a bigger new disruption is arriving: Legal AI.

Where the market is moving

The UK's biggest players are certainly not waiting to see how AI plays out. PwC's most recent Law Firms Survey found that the top 100 firms expect AI to reduce chargeable hours by 16%. While this sounds alarming, it is being planned for, investing in AI infrastructure that makes their cost base lower while their output quality stays high or improves.

The enterprise AI platforms serving this market, Harvey, Legora, Luminance and others, are well-funded, well-marketed and, for the firms that can afford them, genuinely capable. Harvey alone has raised hundreds of millions of dollars. These tools handle document review, contract analysis, legal research and drafting assistance at scale.

For a firm with fifty fee earners, the economics of a £50,000 annual platform subscription are manageable. For a firm with eight, they are not. The result is a growing capability gap between large and small firms that compounds over time. The firms using these tools are getting faster and cheaper at the work that currently takes smaller firms the most time. The gap between what large firms can offer and what smaller firms can offer at what price point and turnaround, is widening every month.

The limits of AI and where your USP wins

Here is where the received narrative about AI and legal services needs some scrutiny.

The dominant AI platforms are trained on public data: published case law, legislation, academic commentary, publicly available contracts. They are broad. They are impressive on many tasks. But they have a fundamental limitation that is rarely discussed openly: they cannot know what has never been made public.

The reality is that more than 90% of litigation settles before trial. The position statements, the mediation transcripts, the email chains where both sides quietly signal what they will and will not accept, the instinctive read a litigator develops over years of watching cases resolve. None of this is in any current training dataset. It lives in the heads of experienced practitioners and in the closed files of firms that have been doing this work for decades.

The same is true across practice areas. The clinical negligence solicitor who knows which medical experts actually help cases settle, versus which ones antagonise the other side. The employment lawyer who can read from an ET3 exactly how the respondent's solicitor will play the hearing. The conveyancer who has handled enough complex chains to know which signals mean a transaction is about to fall apart. This knowledge is not generic. It is not in a public dataset. It is proprietary, and it is the most valuable thing a law firm has.

We think the promise of Legal AI products like Harvey is also overplayed. It's interesting to note that even Microsoft Copilot which is a generalist chatbot, performs nearly as well as Harvey in independent extraction tests. Harvey doesn't train its own models - they use Anthropic's models. At a software level, the tools themselves are not the differentiator. The knowledge that sits behind them is.

The value shift

What is affecting legal firms more profoundly is the value shift that AI creates.

For most small law firms, a significant proportion of revenue comes from billing for time spent on administration, document production and process. Drafting a report on title, preparing court bundles and forms forms, writing estate accounts and so on. These tasks are time-consuming, require qualified staff and form a substantial part of the annual invoice.

AI does all of these tasks better, faster and cheaper than any fee earner and these tasks are well within the remit of any frontier model, whether legal specific or not. LLMs are great at routine, formulaic, word-based tasks and so eat this sort of work whole.

The firms that simply use AI to do the admin faster, bill the client the same and pocket the difference will find that competitive pressure erodes that margin within a few years as the capability becomes ubiquitous and clients start to notice. The firms that understand what is actually happening will use the shift differently.

We think this will have a profound change in terms of what clients are prepared to pay for. In a world where hours are no longer a metric for effort, they will need to pay for good outcomes - and the specialist knowledge and human skill that produces them.

As AI makes generic process work commoditised, it is the skill and experience of seasoned legal experts that creates outsized value.

Where to go now

We think that the firms positioning well for this transition are doing one thing first: capturing and systematising their specialist knowledge.

The knowledge that sits in the heads of their senior fee earners, in their closed files and in the patterns of how their matters actually resolve is what translates into their competitive moat. Making it accessible, consistent and deployable is what turns it into a durable business asset.

We also think this will lead to emergent and more customer-focused legal service offerings. We envisage multiple tiers:

The first tier is automated: routine matters handled end-to-end with AI, with a lawyer reviewing and signing. Lower cost, higher volume. Standard wills, pro forma contracts, compliance documentation, straightforward conveyancing.

The second is agentic: the client self-serves with AI guidance, and a lawyer stays in the loop for the decisions that require genuine judgement. Mid-tier pricing, broader reach, accessible to clients who want guidance but not full representation. Lease queries, employment issues, dispute triage.

The third is white glove: human-led from start to finish, premium by choice rather than by default. Complex litigation, high-value transactions, sensitive family matters. The clients who want it pay more for it, not because it is the only option, but because it is the right one for them.

The infrastructure to build it exists now, at a price point that is accessible to firms well below the Magic Circle. What blocks smaller firms is not the technology, it is the strategic clarity about what they are building and why, combined with the implementation support to build it without disrupting a practice that has ongoing client commitments and finite management bandwidth.

The opportunity window

Markets do not telegraph their inflection points in advance. The 900 conveyancing firms that have exited since 2021 did not mostly close because of a single dramatic event. They thinned out gradually, then suddenly, as the economics stopped working.

The firms that come through the current transition in good shape will not necessarily be the largest or the most technically sophisticated. They will be the ones that worked out early what their real value is, built the infrastructure to deliver it consistently, and created a service model their clients can choose from rather than a single offering at a single price point.

The first step is not buying a platform. It is understanding what you know that AI does not.

Work with Aperic

Aperic works with law firms of all sizes to build the knowledge infrastructure and service architecture to compete in this environment. We design the systems, specify the technology and build the roadmap, without the enterprise price tag. If this resonates, we'd like to talk.