In partnership with

The SRA has proposed writing AI into the Statement of Solicitor Competence, the standard every solicitor must keep meeting after qualification, and the proposals affect supervisors as well as the people using the tools. That is the lead this week, and it arrives in the same week that an Upper Tribunal judge dismissed an appeal whose grounds, drafted by a paralegal using ChatGPT, were "completely infected" by AI, and invited a wasted costs application against the firm. Elsewhere, the new CoCounsel Legal is now available in the UK, Co-op Legal Services is defending its use of AI to score staff calls with bereaved clients, and a conveyancing survey suggests most firms' AI use is still down to individuals.

AI in Practice

The SRA wants checking AI output written into the standard every solicitor must meet

On 8 October the SRA opened an eight-week consultation on the Statement of Solicitor Competence (which sets out what every solicitor must be able to do, and which qualified solicitors are obliged to keep meeting when they assess their continuing competence needs) and on the Functioning Legal Knowledge tested in the SQE. The current statement is about ten years old. The SRA has decided against a standalone technology competence, saying it is not necessary, and proposes instead to add technology and AI to competences which already exist. As reported by Legal Futures, the duty to maintain competence would take into account "developments in technology, including AI", the research competence would add checking "the outputs from technology including AI, by reference to primary resources", and the drafting competence would refer to the "appropriate use" of technology. A new section on supervision makes the point that "competence rests not only with solicitors using technology and AI in legal practice" but also with those who supervise others and oversee their work. The consultation closes at 11.59pm on 3 December 2026, and the SRA expects to publish final versions in April 2027, with changes taking effect from September 2027.

Much of this restates what the SRA's August warning notice and a run of judgments have already said, and this author doubts that anyone who has followed the cases will find the substance surprising. What changes is who has the obligation. A warning notice states what the regulator expects, whereas the competence statement is the benchmark against which a solicitor's continuing competence is measured. Once checking AI output against primary sources is written into it, "I did not know the tool could invent a case" becomes a much weaker explanation, and a supervising partner who never looked at a trainee's AI-assisted research will be judged against that express standard rather than a general one.

For a small firm the practical effect falls on records and supervision more than on technology. If the proposals go through as drafted, a solicitor's annual reflection on training needs ought to cover the AI tools they actually use and how they check them, and a supervisor ought to be able to show how they oversee AI use by the people they supervise, including paralegals and other non-authorised staff (which is exactly where the Sarker case below went wrong). Legal Futures reports that the new supervision section was prompted by the Mazur ruling, and it is the part this author will read most closely in Annex 1, because it is where the SRA will settle how much a supervisor must actually see. These are proposals only, and the wording may change before April.

Takeaways

  • Act: Read the proposed statement in Annex 1 and decide by mid-November whether to respond. In the meantime, list who in the firm uses AI for research or drafting, who supervises each of them, and how their output is currently checked against primary sources.

  • Watch: The SRA's response in April 2027, and whether the final supervision wording expects supervisors to understand the tools their teams use, which would be a heavier obligation than checking the output.

  • Risk: A supervisor who cannot show how a junior's AI-assisted work was checked, at a point when the competence statement says in terms that they should be able to.

Read: SRA consultation, the SRA announcement, and the reporting at Legal Futures

On your radar

  • A paralegal's ChatGPT grounds of appeal end in a wasted costs invitation: In an immigration appeal reported by the Gazette on 8 October, Upper Tribunal Judge Keith found that the grounds of appeal and skeleton argument, drafted by a paralegal at Sarker Solicitors in east London using ChatGPT, were "completely infected" by AI and "seriously misleading", with invented cases, invented quotations from those cases and invented passages from the First-tier Tribunal's own decision. The firm had been warned in the permission decision that the judge had been misquoted, yet still filed a skeleton argument built on the same grounds. The judge said the supervising solicitor had not taken on managerial responsibility for the paralegal, that supervision "did not begin to mitigate the risk of filing seriously misleading submissions", and that the failings were caused by the firm's misuse of AI rather than by the technology. The appeal was dismissed, the Home Office was invited to apply for wasted costs within 28 days, and two people involved have reported themselves to the SRA. Why it matters for UK lawyers: this is the exact supervision failure the SRA's proposed competence wording is aimed at. The earlier warning made it worse, because once a court has flagged a problem every later document will be read for whether anyone acted on it. If a court or an opponent queries an authority or quotation in one of your documents, treat it as a trigger to re-check the whole document against the sources before anything further is filed. (Law Society Gazette)

  • The new CoCounsel Legal is now available in the UK: Thomson Reuters announced on 14 October that the next generation of CoCounsel Legal, the agentic version launched in the US in August, "is available now in the United Kingdom", with Australia, Canada and other markets expected by the end of October. It comes with a Multinational Profile which, according to the announcement, brings legal content from more than 100 countries and territories into one research workflow and flags issues which may need local advice. The release lists research reports built on Westlaw and Practical Law, matter workspaces, bulk document review in tables and integrations which vary by country, but it does not say which UK sources sit behind the research reports or where UK customers' data is processed. Why it matters for UK lawyers: this was covered here last week as on its way, and it has arrived sooner than the "later this year" Thomson Reuters gave in August. This author has had a demonstration of the new version and will report once a proper trial is done. If your firm is offered the upgrade, ask in writing where UK data is processed and which English sources the research reports draw on, and test it on authorities you already know well before signing new terms. (Thomson Reuters)

  • Co-op Legal Services is using AI to score staff calls with bereaved clients: The Guardian reported on 8 October that Co-op Legal Services uses a system built on an OpenAI model to analyse calls between staff and clients about probate, wills and estates, assessing more than 50 aspects of each call and giving managers pass or fail scores, and one member of staff described it as "oppressive and dystopian". CLS told Legal Futures that the tool is not a decision-maker, that it supports "coaching and development for colleagues", and that it is confident the system meets SRA requirements, adding that it has received no formal complaints or grievances. Neither report says whether clients are told that their calls are analysed by AI. Why it matters for UK lawyers: many volume practices already record calls, and adding AI analysis changes the processing for staff (whose monitoring needs a lawful basis, transparency and, in most cases, a data protection impact assessment) and for clients, whose confidential and often sensitive conversations are passed to a third party model. If your firm records calls and is considering AI analysis of them, check that the DPIA, the staff monitoring policy and the client privacy notice all cover it before it goes live. (Legal Futures)

  • Most conveyancing firms say their AI use is still individual experimentation: InfoTrack's 2026 Digital Conveyancing Maturity Index, unveiled at its own summit and reported on 12 October, found that 69% of firms described their AI adoption as individual experimentation, 18% had rolled it out across a team and 14% said it was embedded as standard practice. This is vendor research published as sponsored content, and the sample size is not given. Why it matters for UK lawyers: AI use which depends on individuals is the hardest kind to supervise, and it is the position the SRA's proposed competence wording would make most uncomfortable. If your conveyancing team falls into the first group, ask each fee earner which tools they use and for what, and record the answers before deciding which uses the firm will support. (Legal Futures)

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.

The AI notetaker that gets the hard words right

Your AI is only as good as what you feed it. Feed it a transcript with your product names, acronyms, and numbers spelled wrong, and you spend the afternoon hand-fixing the output.

Wispr Flow Notetaker uses your dictionary and calendar, so product names, acronyms, numbers, and uncommon names come out spelled right. It pops up when your meeting starts and captures Zoom, Google Meet, Teams, and Slack huddles with one click, no bot joining the call.

Then your meetings show up inside Claude or ChatGPT through the built-in connector. No copy paste. Try Notetaker free on Mac and Windows.

For Review

Why law firms are getting AI supervision wrong (Patrick Herald, Legal Futures)

A blog from 9 October, published as associate content sponsored by Access Legal, arguing that verification checklists and e-learning treat AI errors as occasional glitches, when a correct citation and an invented one come from the same process and look equally convincing. Its practical value is the point that supervision cannot be discharged by a policy document alone, and that firms should look at the deadlines and workloads which push people towards shortcuts; worth reading before you draft any response to the SRA's supervision proposals.

Read or listen: Legal Futures

Use of AI in preparing court documents (Civil Justice Council)

Not this week's news: the CJC's interim report and consultation closed in April, and its final report is expected later in 2026. Respondents broadly agreed that statements of case and skeleton arguments need no new AI-specific rule, leaving witness statements as the main open question. Read it alongside the SRA proposals, because between them they will set what a litigator must check, and what they may have to declare, when AI has been used.

Practice Prompt

Try the below prompt to review how AI is used and supervised in your team against the SRA's proposed competence wording, before the consultation closes on 3 December. 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 a supervising solicitor, COLP or practice manager at a law firm in England and Wales who wants to review how AI is used and supervised in one team, measured against the SRA's October 2026 proposals to add technology and AI to the Statement of Solicitor Competence. As publicly reported, those proposals would require solicitors to maintain competence taking into account developments in technology including AI, to check the outputs of technology including AI by reference to primary resources, to make appropriate use of technology in drafting, and would add a supervision section making clear that competence rests with those who supervise others as well as with those who use the tools. Your task is to produce a gap review and an action list. You are not assessing whether any tool is good, and you are not concluding that the team is compliant.

Context to apply:

- The team and its work: {e.g. a four-person private client team doing wills, LPAs and probate, or a litigation team of two solicitors, a trainee and a paralegal}
- Who does what, by role and qualification: {list each role, whether they are a solicitor, trainee or non-authorised, and who supervises them}
- The AI tools in use, approved or not: {e.g. firm-licensed research tool, Microsoft Copilot, a practice management system's AI features, personal ChatGPT accounts; say "unknown" if nobody has asked}
- What each tool is used for: {e.g. legal research, first drafts of letters, summarising documents, drafting grounds or witness statements}
- How AI output is currently checked, and by whom: {describe, or say "no set process"}
- How supervision currently works: {e.g. file reviews, sign-off before anything leaves the firm, weekly meetings}
- The firm's AI policy and training: {summarise, or say "none in writing"}
- Any incident or near miss so far: {describe without names, or say "none known"}

Produce the review under these headings:

1. Current use
Set out, role by role, which tools are used for which tasks, and mark separately what has been confirmed and what has been assumed.

2. Checking against primary sources
For each task where AI output could reach a client, a court or an opponent, state how the output is checked against primary sources (legislation, judgments, rules, the client's own documents), who does the check, and whether any record of it is kept. Mark each as adequate, unclear or missing, with a one-line reason.

3. Supervision
For each person supervised, describe how their supervisor would know which tasks involved AI, what was checked and by whom. Identify any route by which AI-assisted work could be filed or sent with no qualified person having reviewed it, paying particular attention to work done by paralegals and other non-authorised staff.

4. Competence and training
Identify, for each solicitor, what they would need to know about the tools they use or supervise in order to meet the proposed wording, and suggest specific items for their next continuing competence reflection.

5. Escalation
Set out what should happen when a court, an opponent or a client queries an authority, quotation or fact in a document prepared with AI, including who must be told, and a rule that the whole document is re-checked before anything further is filed.

6. Consultation points
List any points where the proposed wording, as summarised above, would be difficult for a firm of this size to meet in practice, and the evidence the firm could give if it decides to respond by 3 December 2026.

7. Action list
Give no more than eight actions, each with an owner and a date, in order of priority.

8. What this review cannot tell you
Flag every point where the output rests on an assumption about the team, the tools, the proposals or the law, and say what the firm would need to establish to replace it with a fact.

Constraints:

- {Add firm-specific constraints, for example a client who has prohibited AI use, an insurer requirement, a recent file review finding, or a limited training budget.}
- Apply the law of England and Wales and the SRA Standards and Regulations. Treat the SRA's proposals as proposals only, and where you are unsure what the consultation paper or Annex 1 says, say so and tell the user to check the document itself.
- Do not invent rules, regulatory guidance, case law, product features or consultation wording. If you cannot verify something, mark it as unverified.
- Do not conclude that the team or the firm is compliant, competent or adequately supervised. The purpose is to surface gaps and open questions.
- Write in plain English and keep each finding short enough to act on.
- This is a planning aid, not legal or regulatory advice. The firm, its managers and its COLP remain responsible for supervision and for any use of AI in client work.

How did we do?

Hit reply and tell me whether your firm plans to respond to the SRA's consultation, and what you would want the supervision wording to say.

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!

Level Up Newsletter

Level Up Newsletter

Staying Ahead with AI

Staying Ahead with AI

Step by step on how to use the latest in AI and how it ranks against what you're already using!

Superhuman AI

Superhuman AI

Keep up with the latest AI news, trends, and tools in just 3 minutes a day. Join 1,500,000+ professionals.