For many small and medium-sized enterprises, legal work is something to be squeezed in between sales calls and payroll runs. A supplier goes quiet, a customer disputes an invoice, or a contract needs tightening, and the fastest route to a first draft is increasingly a search bar or a chatbot.
The behaviour sits inside a wider shift. The Office for National Statistics (ONS) says 25% of UK businesses reported using some form of artificial intelligence in late December 2025, up from 9% when the question was introduced in September 2023. The British Chambers of Commerce (BCC), working with Intuit, found 35% of SMEs said they were actively using AI in 2025, up from 25% in 2024.
Demand for legal help, meanwhile, remains at a consistent level. In its small business legal needs survey, the Legal Services Board estimated that 31% of small businesses in England and Wales experienced a legal issue in 2021, yet only 25% used professional help as their main adviser. Using an OECD model, the regulator also estimated that 12% of those with a legal issue had an unmet legal need.
Satish Jakhu, Managing Director and Head of Dispute Resolution & Litigation at RLK Solicitors, says his team is increasingly dealing with disputes where the early steps were shaped by AI tools or a do-it-yourself approach.
“On the surface, AI-generated advice looks polished and commercially sensible,” he explains. “The problem is rarely that it’s obviously wrong. It’s that it’s almost right. It sounds confident and appears to deal with the issues, but it’s fundamentally misaligned with how the business operates or how disputes actually unfold.
“In live litigation, what matters is not whether an argument can be made, but whether it’s the right argument to run. AI will often throw out every conceivable point without differentiating between what’s decisive and what’s noise. Poorly prioritised arguments can dilute credibility and actively weaken a case.”
He also argues that time lost early is rarely recovered. “Delay is one of the most common risks associated with DIY and AI-led approaches. Businesses assume a problem will “blow over” or worry that involving lawyers too early will escalate matters. In reality, early legal input usually reduces cost because it preserves options, protects evidence and maintains leverage.
“By the time we’re instructed, evidence may be incomplete, communications may already be unhelpful and commercial relationships may be damaged beyond repair. At that stage, it’s often about containing damage rather than shaping the outcome.”
The cost of being wrong —
For SMEs, these issues collide with a court system that is busy and procedural. The Ministry of Justice’s Civil Justice Statistics Quarterly shows County Court claims rising to 490,000 in January to March 2025. The mean time from claim to trial was 49.8 weeks for small claims and 74.7 weeks for multi, intermediate, and fast track cases.
Jurisdiction is another blind spot. “Businesses often don’t realise the jurisdictional gap until an opponent, insurer or court points it out- by which time the damage is done,” Jakhu says. Many widely used tools are trained on broad datasets and do not reliably distinguish between legal systems, which can turn a plausible-sounding approach into a procedural misstep.
Regulators and professional bodies have been sharpening their own guidance as the technology spreads. The Law Society notes there are currently no statutory obligations on generative AI companies to audit outputs for factual accuracy, and advises against feeding confidential information into free, online tools where users lack control over how data is processed. A YouGov poll commissioned by The Law Society in January 2025 found that 77% of UK adults said they would not trust AI like ChatGPT to provide legal advice.
In June 2025, Dame Victoria Sharp, President of the King’s Bench Division, warned that lawyers who cite fake cases generated by AI could face sanctions, underlining the need to verify outputs even in professional settings.
Where AI fits, and where it doesn’t —
Jakhu is careful to draw a line between using AI as a tool and treating it as a substitute.
“Legal judgment is about outcomes, not just wording. It’s about enforceability, risk appetite, leverage and commercial reality. AI can draft words, but it cannot judge how those words will perform in a live dispute.”
A Thomson Reuters Institute report published in 2025 found 26% of legal organisations were already using generative AI, up from 14% the prior year, with heavy use in document review, legal research, and drafting. The direction of travel is towards augmentation, with human accountability sitting over the output.
For SMEs, AI can be a starting point for drafting and issue-spotting, but the moment a matter becomes contentious — or jurisdiction, evidence, or deadlines start to matter — speed alone can become a false economy.
You must be logged in to post a comment.