I have been reading a run of surveys this week, and the finding which ought to matter to most firms is that clients have noticed AI and want to know what it means for their bill. That is the lead story. Elsewhere, a High Court judge has told expert witnesses to start disclosing AI use now, a listed firm has built an AI tool for source-of-funds enquiries, and the ICO's deadline on agentic AI arrives next week.
AI in Practice
Clients are asking why AI has not cut your fees, and many firms have no answer ready
BigHand's Law Firm Talent and Resourcing Report 2026, published this week and based on a survey of more than 800 law firm leaders in the UK and North America, found that 46% say clients have questioned why fees have not fallen despite the firm's use of AI, and 42% have received requests for alternative fee arrangements linked to AI efficiencies. Only 1% said clients had not raised AI's effect on pricing or service delivery at all, while just 23% regularly monitor matter-level profitability at management level (Legal Futures). Some caution is needed: BigHand sells resourcing software, the findings appear as the company's own associate content rather than independent reporting, the UK and US figures are combined, and an earlier BigHand report in July gave very similar numbers. This author would view it as a direction of travel rather than suggesting one should rely on the figures given.
That direction is consistent with other research published the same week. Barclays' Business Prosperity Index, reported on 11 November, found that 32% of the 125 legal services firms surveyed say they have redesigned their business model and workflows to make full use of AI (against 16% across business and professional services generally), and 93% plan to increase AI investment over the next year, although Barclays itself describes the legal figures as indicative because of the small sample (Solicitor News). Firms are spending and restructuring, and clients can see it.
For hourly work, the Law Society's practice note of 21 October already answers much of the question: firms may bill only for time actually and properly spent, not for the time a task would have taken without AI, so the saving reaches the client by default, provided time recording reflects what was really done (Law Society). Fixed and capped fees are harder. A price set before the firm adopted AI, which the client agreed and still considers fair, is not improper simply because the work now takes less time, but it is where the question will be asked at the next panel review, tender or renewal, and "we have invested heavily in AI" will not satisfy a general counsel who is using the same tools in-house.
This author's view is that firms able to explain, in two or three sentences, where AI is used on a given type of matter, what it actually saves and where that saving goes (a lower price, a faster turnaround, or more senior time on the difficult points) will be in a better position than firms which leave the client to guess. The figures need to be measured on real files, though, because promising savings which have not been tested is its own risk.
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 voice-to-text app that built its own speech model
Most dictation apps are a wrapper around someone else's speech model. Wispr Flow trained its own.
It's called Canto, and it was built for the way people actually talk to AI: thinking out loud, changing direction mid-sentence, half-whispering in an open office.
Wispr tested it on 10 hours of real dictations from more than 2,300 speakers. Canto had the lowest word error rate of every model tested, including models from Google, OpenAI, AssemblyAI and Deepgram. That means your prompts to ChatGPT, Claude and Cursor come through the way you said them.
Talk naturally in any app and Wispr Flow types clean, formatted text, 4x faster than typing.
On your radar
A High Court judge has told experts to disclose AI use now, before any rule requires it: Speaking at the Bond Solon Expert Witness Conference, Mrs Justice Joanna Smith (who co-chaired the Civil Justice Council working group whose final report led this newsletter last week) told experts to disclose AI use which "may directly or even indirectly generate or influence the content of your evidence", warned that undisclosed use could lead to "searching questions in cross-examination", and said it "makes good practical sense" for an expert to tell the instructing solicitor about intended use first. She described the "erosion of independent judgment, the outsourcing of your opinions to an AI agent" as the main risk to expert integrity, as reported by the Gazette on 7 November. Why it matters for UK lawyers: last week's report was a set of recommendations which still need rule changes, whereas this is a senior judge telling experts what the court expects today, and the expert's first point of contact is you. Add an AI question to your letters of instruction if you have not already, and ask experts instructed on live matters whether, and how, they have used AI in their reports. (Courts and Tribunals Judiciary, Law Society Gazette)
A listed firm has built an AI tool to run source-of-funds enquiries with clients: Keystone Law has launched Origin, which sends clients (and third-party funders) a link guiding them through questions about where their money came from, asks follow-up questions and requests supporting evidence (for example a marriage certificate where funds are held in a spouse's name), then compiles a report for the handling lawyer. Keystone says its compliance team still judges whether the source of funds is "complete and fit for purpose". The coverage does not say which model the tool uses, where client data is held, or whether it will be offered to other firms, and there is no independent assessment of how well it works. Why it matters for UK lawyers: source of funds is one of the most time-consuming parts of AML compliance for conveyancing and private client teams, and Keystone is not the only provider offering AI help with it. Before adopting any such tool, make sure your firm-wide risk assessment and matter risk assessments record who makes the final judgement, and ask the vendor where clients' financial documents are stored and processed. (Keystone Law, Legal IT Insider)
Legora is pitching AI to UK courts and government legal teams: Legora announced on 11 November that Lucy Frazer KC (a former Cabinet minister and twice Solicitor General, who has advised the company since July) is taking an expanded part-time role leading its work with governments, regulators, local authorities and courts, starting in the UK and EMEA, and it points to the record Crown Court backlog of 80,829 cases at the end of June 2026 as the kind of problem it wants to help with. No UK government contract has been announced. Its chief executive says "every answer can be checked, and a lawyer signs off on the result", which is a claim to test rather than accept. Why it matters for UK lawyers: local authority and government legal teams can expect more approaches of this kind, with procurement rules and freedom of information requests on top of the questions private firms ask. If you work in a public body's legal team, make sure any pilot goes through procurement and a DPIA before it touches live casework, and ask for the checking and sign-off claims to be demonstrated on your own material. (Artificial Lawyer, Legal IT Insider)
Reminder: the ICO's call for evidence on agentic AI closes on 20 November: Covered here on 30 October, and the deadline is now a week away. The responses will inform the ICO's guidance on agentic AI and the statutory code of practice on AI and automated decision-making which it must produce under the Data (Use and Access) Act 2025. Why it matters for UK lawyers: agents which act across email, documents and case management systems handle client personal data in ways a chatbot does not, and the code will set the standard firms are measured against. If your firm is piloting agents and has a view, submit it by 20 November, and check that the DPIA for any pilot covers what the agent can access and do on its own. (ICO)
For Review
Is Big Law 'AI Gaslighting' Us? (Richard Tromans, Artificial Lawyer)
A short opinion piece from 12 November which asks what large firms' expensive AI announcements actually deliver to clients, and suggests firms show tangible evidence under four headings: lower cost, speed, quality of work and reduced risk. It is written from a large-firm, largely US perspective, but its suggestion that fixed-fee and no-win, no-fee practices may be where clients see the benefit first is worth reading alongside this week's lead, and the four headings are a reasonable template for the AI paragraph in your next pitch.
Read or listen: Artificial Lawyer
'We now ask the AI first': new research on AI in in-house legal (Ben Wheway, Legal Business)
A report from 11 November on the 2026 General Counsel Legal Tech Pulse, produced with Thomson Reuters (a vendor) from more than 250 senior in-house lawyers in the UK and Europe, which found that 82% are implementing generative AI and that close to 60% now regard it as reliable "completely" or "a lot", up from 40% last year. The anecdotes matter more than the percentages for law firms: one head of legal says AI is "more than enough" in 90% of cases, so the team no longer calls a law firm first, and another says the company has stopped using external counsel for minor litigation. Read it before your next client review meeting to understand which work is most likely to stay in-house.
Read or listen: Legal Business
Practice Prompt
Try the below prompt to review how your firm's use of AI on a particular type of matter is reflected in its pricing, and to draft a short, accurate explanation you could give a client who asks. 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 managing partner, head of department or finance director at a law firm in England and Wales. Clients have started to ask how the firm's use of AI affects what they pay. Your task is to review one type of matter, set out how AI is used on it and how that is reflected in price or service, and draft a short explanation the firm could give a client. You are not deciding the firm's prices, and you are not concluding that the firm's billing practice is compliant.
Context to apply:
- The type of matter: {e.g. residential conveyancing purchases, commercial lease renewals, employment tribunal claims for respondents}
- How the work is priced: {e.g. fixed fee of £X, capped fee, hourly rates by grade, conditional fee agreement}
- When the current price or rates were last set, and whether that was before or after the firm began using AI on this work: {give dates, or say "unknown"}
- The AI tools used on this work and the tasks each is used for: {e.g. firm-licensed drafting tool for first-draft reports, transcription tool for client meetings; say "unknown" if not confirmed}
- Time data, if any, comparing files before and after AI was introduced: {summarise, or say "none measured"}
- Who checks AI output on this work, and how the check is recorded: {describe, or say "not formalised"}
- What the client care letter and terms of business currently say about AI and charging: {quote or summarise, or say "nothing"}
- What clients have actually asked, if anything: {e.g. a panel review question on AI efficiencies, a request for a lower fixed fee}
Produce the review under these headings:
1. Where AI is used
List each stage of the matter and state whether AI is used, might be used or is not used. Mark separately what has been confirmed and what has been assumed.
2. Effect on time and cost
For each stage where AI is used, state what time data exists and what it shows. Where no measured data exists, say so plainly and do not estimate a saving. Identify the new time AI creates (operating the tool, checking and correcting output) as well as the time it removes.
3. Billing position
For hourly work, explain how time should be recorded and described so that only time actually and properly spent is billed, consistent with the Law Society's October 2026 practice note as publicly reported (time operating the tool and reviewing, verifying and revising its output may be charged where the client care letter allows; any separate technology charge must be explained and agreed in advance and must not duplicate other fees). For fixed or capped fees, set out whether the price was set before AI was introduced and what questions a client is likely to ask about it.
4. Where the benefit goes
Set out, honestly, whether the benefit of AI on this work currently goes to the client (lower price, faster turnaround, more senior attention) or to the firm (margin), or a mix, and what evidence supports that view.
5. Draft client explanation
Draft a plain-English explanation of no more than 150 words which the firm could give a client who asks how AI affects their fees on this type of matter. It must not promise savings that have not been measured, must not overstate what the tools do, and must make clear that a named solicitor remains responsible for the work.
6. Options
Give no more than four options for the firm to consider (for example, leaving the price unchanged with a clear explanation, adjusting the fixed fee, offering a faster-turnaround option, or measuring time for three months before deciding), with one line on the risk of each.
7. What this review cannot tell you
Flag every point resting on an assumption about the firm's tools, time data, client terms, the Law Society's guidance 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 panel agreement with fixed rates, an insurer requirement, a pending tender, or a client who has prohibited AI use.}
- Apply the law of England and Wales and the SRA Standards and Regulations. Treat the Law Society practice note as guidance from a representative body, not regulation, and where you are unsure what it says, tell the user to check the note itself.
- Do not invent time savings, prices, product features, regulatory guidance, case law or statutory provisions. If information is missing, say what is missing.
- Do not conclude that the firm's billing is compliant or that its prices are fair. The purpose is to surface the position and the open questions.
- Write in plain English and keep the whole review short enough to read in five minutes.
- This is a planning aid, not legal, regulatory or financial advice. The firm and its managers remain responsible for its pricing, its billing and what it tells clients.How did we do?
Hit reply and tell me whether a client has asked you yet what AI means for their bill, and what you said.
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!




