Sponsored by
This week the story is AI moving from helping with legal work to doing it. A regulated AI law firm has won its first contested court case, a national firm has launched its own contract review tool built on its own know-how, and the Master of the Rolls has again told the profession it cannot afford to sit AI out. The lead is the courtroom result, because it is the clearest sign yet of where AI is genuinely useful and where a human still has to stand up and speak.
AI in Practice
A regulated AI law firm wins its first contested court case
Garfield AI, the first AI-driven law firm authorised and regulated by the SRA, said this week that it has won its first contested trial. Acting for a freelancer who was owed unpaid fees for HR services provided to a hospitality business, Garfield generated the pre-action correspondence and prepared and issued the county court proceedings. At a contested hearing at Wandsworth County Court the claimant recovered £7,000 and the defendant's counterclaim was dismissed. The firm instructed a human barrister for the advocacy, so the AI did not address the court.
The economics are the key point that interests this writer. The firm says the client paid around £400 in Garfield fees to recover £7,000, while the defendant instructed both a solicitor and counsel. For low-value debt recovery, where the cost of acting has long been the practical barrier to enforcing a perfectly good claim, that is a real shift. Garfield says it has now had more than 600 claims started on its platform and has recovered or resolved over £500,000 for users, with claim values from about £30 to £10,000. These are the firm's own figures, and one county court win is not a line of authority, so the claims are simply worth noting, rather than relying upon.
What is on display here is a hybrid model. The AI handles the structured, document-heavy steps (letters before action, particulars, court forms) and a regulated human stays responsible and conducts the advocacy. For any firm with a volume debt recovery or small claims practice, the competitive question has stopped being hypothetical. For everyone else, it is a concrete example of where the technology is strong (repeatable, form-driven process) rather than where vendors say it will be.
Takeaways
Act: If you do volume low-value litigation or debt recovery, look at where a structured, AI-assisted process could undercut your current cost base, and what that means for how you price the work.
Watch: Whether the SRA-regulated AI-firm model spreads beyond debt recovery into other high-volume, low-complexity work.
Risk: The responsibility point stays the same as always. Garfield is regulated and used a human barrister, so the lesson is not "let the AI run the case", it is "AI can run the process under proper supervision".
On your radar
Shoosmiths has built its own AI contract review tool: The national firm has launched Project Apollo, a generative AI contract review tool built over the past year with Microsoft and running in its Azure environment, which the firm owns and controls rather than licensing off the shelf. It is designed to show its reasoning (the firm calls it "auditable reasoning") and to surface Shoosmiths' own playbooks and guidance, and it is pitched as much as a training aid for junior lawyers as a productivity tool, with the firm reporting up to five hours saved on a single contract review. Why it matters for UK lawyers: the build-versus-buy question is now live for mid-size and larger firms, and a tool that records and explains its reasoning is easier to supervise, and to defend, than a black box. If your firm is evaluating contract review tools, add "does it show and log its reasoning" to your selection criteria. (Legal IT Insider)
The Master of the Rolls says ignoring AI is "not an option": Sir Geoffrey Vos told the Expert Witness Institute's annual conference that "if you leave AI behind, you will be left behind", and that ignoring AI is not realistic for anyone in the litigation business. It is a theme he has returned to often (he told London International Disputes Week earlier this month that the profession must "adapt or die"), and these are among his final interventions before he retires as Master of the Rolls on 31 October 2026. This writer set out why he agrees with Vos, and where the argument carries an uncomfortable implication for law firms, in an earlier editorial. Why it matters for UK lawyers: the senior judiciary's direction of travel is clear, but the responsibility for checking AI output stays with the lawyer, not the tool. It is useful framing for any internal conversation about why adoption and verification have to travel together. (Law Society Gazette)
The data protection baseline under your AI tools is shifting: UK data protection law is changing this year, with the provisions of the Data (Use and Access) Act now coming into force and the ICO due to publish a statutory code of practice on AI and automated decision-making over the summer. Why it matters for UK lawyers: any AI tool that processes personal data sits on this moving baseline, and "we have always done it this way" is not a safe assumption. Check that the data protection impact assessment and lawful-basis analysis for any AI tool you use are current, rather than inherited from last year. (ICO)
Ad Break
In order to help cover the running costs of this newsletter, please check out the advert below. In line with my promises from the start, adverts will always be declared.
Most AI newsletters summarize headlines. MavSource is different.
We aggregate updates from all major AI newsletters, podcasts, company news, AI labs, public and private company activity, GitHub projects, funding rounds, earnings calls, and investor letters — hundreds of sources every day. Then we summarize what matters, analyze emerging trends, and add our own founder commentary so you understand why a development may matter — not just what happened.
One 5-minute email, every morning. Built for investors, founders, and operators who want to understand AI as a business, technology, and market trend — not just another news cycle.
The daily email is free. It's also the entry point to a deeper intelligence product covering watchlists, public-company read-throughs, startup trackers, and investor-letter tracking.
For Review
The legal AI scaffold changes everything (Artificial Lawyer)
A study by Legal Nodes took a strong general model (Claude Opus 4.8) and tested it on legal tasks under different set-ups, finding that output quality depended less on the base model than on the "scaffold" around it: context, workflow logic, retrieval, prompting and the use of skills. A useful corrective for any firm choosing tools on model brand alone, and a reason to ask vendors what actually sits around the model.
Read or listen: Artificial Lawyer
How Shoosmiths built its own AI (Microsoft UK Stories)
A longer case study on the thinking behind Project Apollo: why the firm chose to build on its own knowledge and playbooks rather than buy, and how it approached explainability and the training of junior lawyers. Worth reading if you are involved in your own firm's build-versus-buy decision.
Read or listen: Microsoft UK Stories
Practice Prompt
Try the below prompt to structure a letter before action for a low-value debt claim, the kind of high-volume, form-driven work this week's lead story shows AI handling well. Ensure you fill in context and constraints and other aspects marked with {}. Remember to adhere to the Golden Rules and do not upload confidential or privileged information to public tools.
You are assisting an experienced civil litigation solicitor practising in England and
Wales. Your task is to produce a first-draft structure for a letter before action in a
low-value debt claim, as a planning aid only.
Context to apply:
- Creditor: {name; status: individual / sole trader / company}
- Debtor: {name; status: individual / sole trader / company}
- The debt: {amount; what it relates to; the relevant dates}
- Basis of the debt: {contract / invoice / agreement, with the key terms}
- Interest claimed: {basis, for example a contractual rate, or the Late Payment of
Commercial Debts (Interest) Act 1998 where it applies}
- Steps already taken: {prior correspondence, reminders, part-payments}
Produce:
1. A note on which pre-action regime applies (the Pre-Action Protocol for Debt Claims
where the debtor is an individual or sole trader, or the Practice Direction on
Pre-Action Conduct and Protocols otherwise) and what that means for the content and
the timescales.
2. A structured outline of the letter before action, with headings and the points to
cover under each.
3. A list of the documents and figures the solicitor must verify before sending, for
example the contract terms, the sum claimed, the interest calculation and the correct
debtor entity.
4. A short list of the most common ways a letter of this kind goes wrong.
Constraints:
- Apply the law of England and Wales only.
- Do not invent rules, protocols, case names or figures. If you are not certain a rule
applies, say so and leave it for the solicitor to confirm.
- Flag anything that depends on facts that have not been provided.
- This is a planning aid, not legal advice, and the supervising solicitor remains
responsible for the final letter.How did we do?
Hit reply and tell me what you would like covered in future issues or any feedback. We read every email!
Thanks for reading,
Serhan, UK Legal AI Brief
Disclaimer
Guidance and news only. Not legal advice. Always use AI tools safely.
Recommended Newsletters
Below are a few newsletters that I recommend, for various reasons. Check them out!
