AI for law firms: what UK solicitors can automate in 2026

Intake and conflict preparation are the safest first builds, because the AI gathers facts and a solicitor still makes every decision. A typical intake build reads each new enquiry from the website form or the inbox, pulls out the parties, the matter type and any deadline, and opens a draft matter in the case management system, for example Clio, LEAP or Actionstep.

The same build runs each party name through the firm's conflict search and returns every hit with the matter it came from. The SRA Code says "You do not act if there is an own interest conflict or a significant risk of such a conflict" (paragraph 6.1), so the decision to act stays with a solicitor.

AML and ID checks

The SRA's 2023 Risk Outlook report lists parts of anti-money laundering checks among the routine compliance work AI can automate. In practice that means requesting ID and source of funds documents, chasing the client until each one arrives, filing it to the matter and pre-filling the risk assessment from the ID provider's result.

The firm's AML duties come from the Money Laundering Regulations 2017. The sector guidance is the Legal Sector Affinity Group (LSAG) guidance, and the SRA says HM Treasury approved its latest revision with effect from 23 April 2025. The risk rating, any enhanced due diligence and every suspicion stay with people, and a suspicion goes to the firm's MLRO.

Client updates and time entries are the most repetitive writing in a firm, and both suit a draft-then-approve pattern.

Matter updates to clients

The build drafts a status email at each milestone, or every week, from the matter notes, the last correspondence and the next step. The fee earner reads it and approves it with one click.

Time recording

The build drafts time entries from the day's sent emails, documents edited and calls logged, in the firm's narrative style and units. The fee earner confirms each entry before it reaches a bill.

No published study gives the hours UK firms lose to unrecorded time, so estimate it from your own figures. A worked example with stated assumptions follows.

1. Assume 12 fee earners each fail to record 12 minutes (two 6-minute units) of billable work a day, over 220 working days.
2. That is 12 × 0.2 hours × 220 = 528 hours a year.
3. At an assumed charge-out rate of £180 an hour, 528 hours is £95,040 of work that never reaches a bill.
4. If draft entries recover half of it, the firm bills £47,520 more a year.

Replace each assumption with figures from the firm's own time-recording reports.

Client data should only go into AI tools whose terms keep it confidential and out of model training. The SRA's August 2026 notice warns that with public AI tools "legal professional privilege may be permanently waived and unable to be recovered". It tells firms to check that client data is not used to train models "except where explicitly authorised", and says free and paid tools can both carry risk, depending on the provider's terms. Paragraph 6.3 of the SRA Code still says "You keep the affairs of current and former clients confidential".

For UK GDPR, the ICO's "Guidance on AI and data protection" covers data protection impact assessments, lawfulness and transparency. The ICO says the guidance is under review after the Data (Use and Access) Act, so check the page before each new build.

What to keep human

• Deciding whether to act, and every conflict decision
• Advice to the client
• Anything filed at court, with every citation checked against an authoritative source
• AML risk ratings and every suspicion
• Sending anything to the client or the other side
• Final bills

Yarigai is a done-for-you AI agency in London. It builds AI agents, document reading and filing, inbox triage and internal tools inside a client's own accounts, which for a law firm means its email, document store and case management system. The firm owns the code, the data and the accounts from day one.

A project starts with a free 30-minute scope call on Google Meet, and the firm keeps the notes whether it hires Yarigai or not. A blueprint of about a week, credited to the build, sets out a written plan, a clickable prototype, the monthly running cost, a fixed price and the payback maths. The build takes 2 to 6 weeks in the firm's accounts with a demo of working software every week. A single workflow, such as intake, can go live in about two weeks.

A person approves anything that sends money, emails a customer or deletes data. AI and hosting bills are paid by the firm at cost, and the monthly running cost is on the quote. Yarigai's guarantee is one sentence: "If the build isn't profitable for you, we won't build it."

Yarigai has no published law firm case study. Its public proof is the software it built and runs, described on the about page, and its document and operations work is on the services pages.