PwC: AI a long way off from replacing lawyers
With most law firms rapidly adopting AI, new research by PwC has shown that 80% of legal work faces “meaningful barriers” to full automation.
In many key areas – from courtroom advocacy to ‘bet-the company’ moments – clients want to rely on a lawyer they trust.
PwC said: “That is where the human edge lives, and where pricing power will increasingly concentrate.”
The research – The new rules of legal services: five moves to win as AI rewrites value [1] – revealed that law firms “expect AI to unlock efficiency gains equivalent to around 16% of chargeable hours” in 2026. Last year, the figure was 11%.
This includes using AI for tasks such as document drafting and review, legal research and precedent analysis.
PwC estimated that AI could deliver £6bn of the legal work carried out by lawyers in the UK, equivalent to around 10% of the total market.
Researchers observed: “It is unsurprising, then, that 53% of firms report concern about clients using automation to reduce demand, and 73% worry about keeping pace in what many partners now call a ‘tech arms race’.”
Another key insight is the lack of financial gain from the investments made so far into AI. PwC said “only a minority are turning activity into monetisable gains”, unlike in other sectors.
It went on: “Firms that harness AI effectively can differentiate through higher-quality insights, faster and more innovative service delivery, or materially lower cost bases. Firms that do not will struggle to keep up.”
The report stressed that “AI is coming for the production layer of legal work, not the legal profession”.
”
Approximately 80% of the legal market retains meaningful barriers to full automation, including courtroom advocacy, complex negotiation, regulated judgement calls, and the ‘bet-the-company’ matters where clients want a trusted human in the room.”
PwC estimated that around 22% of fees came from work that was “relatively easy to automate”, and clients were sophisticated enough to bring it in-house with AI support, such as “transactional commercial contracting for enterprises” and patent drafting.
There were “moderate barriers” for 39% of work – such as personal injury and clinical negligence, and property work for SMEs – where either “the work is easier to automate but clients are less able to bring it in-house, or clients are sophisticated enough to in-source but the work is harder to automate”.
The other 39% was the work with the highest barriers, where the work “is hard to automate and clients are less able to self-serve”. These areas were often in the consumer and SME segments, such as employment and family work for consumers, and litigation/ADR.
But the report said this did not mean firms should simply retreat to the highest-value work.
“It is that every firm needs a deliberate portfolio view: which practice areas to defend, which to industrialise, which to exit, and where to invest in new tech-enabled propositions.”
PwC predicted that the legal market would continue growing, “despite AI-driven pressure”.
“Firms that are able and willing to invest in AI and transform their delivery model have the potential to meaningfully outperform, increasing margins.
“Efficiencies can be driven at the same time as improving service quality, generating incremental demand from clients that can be serviced with repurposed hours.
“Winners will win faster, and threatened firms will be easy targets for consolidation.”
It described four different AI models: the AI-augmented firm, AI-augmented outsourcing, AI-native firms, and self-serve models.
There were many “AI-augmented” firms across the UK – mainly mid-sized and larger firms – where AI technologies have been added to existing workflows. The work was still led by a lawyer, who used the AI to enhance their productivity.
More smaller firms have chosen the AI-augmented outsourcing model, where they send lower-value work to external partners, who use the same technologies to deliver the work more cost efficiently.
The other models are emerging now “and shorten the value chain materially”.
AI-native firms deliver legal work through largely automated workflows, while self-serve models are where clients use legal AI products such as Harvey and Legora, or general-purpose AI such as Claude, without engaging a law firm at all.
“Each model competes for different parts of the value chain. Most firms will need to deliberately choose where they play,” said PwC.
“Much of the debate in the legal sector has been framed too narrowly. The central question has often been whether AI can technically perform the work of lawyers. That is an incomplete picture.
“The more important question is how AI changes client demand: when clients turn to outside counsel, what they are willing to pay for and which parts of the value chain they are happy to handle themselves.
“The opportunity is not just about doing legal work faster. It is also about having the chance to transform what the firm sells, how it prices it, and who delivers it.”
To drive growth and relevance to clients, PwC highlighted five strategic moves for law firms:
- Work out which parts of the firm’s market can be automated;
- Build the firm’s pricing strategy around the value they deliver to clients;
- Allow AI to operate autonomously between each tier within the traditional law firm pyramid;
- Use AI to run the back office; and
- Have strong AI governance and cyber security so clients trust the firm.
PwC said: “What this looks like in practice depends on the firm’s archetype. AI-augmented incumbents should prioritise portfolio review, pricing reform and back-office industrialisation.
“AI-native or AI-forward challengers should focus on scaling proprietary tooling, productising IP, and building the governance maturity that enterprise clients now demand.
“Outsourcing-led players should double down on cost-to-serve and integration with client systems. The five moves apply to every firm. The order they are taken in does not.”