Episode 7. Law Society AI practice note, what solicitors change.

0:08 On Wednesday the Law Society published fifty eight pages on AI, and one sentence in them reaches every firm that bills by the hour.

0:16 If you charge on time, you should not bill for the time a task would have taken had AI not been used.

0:23 The hour the machine saved belongs to the client, and the profession's own body has now said so in writing.

0:31 Good morning, and welcome to The Legal AI Brief, the Monday round of the AI news that matters to smaller law firms.

0:37 I am Saqib Khan. My legal career began in large commercial law firms and includes fourteen years as an owner of a law firm and four years as a fee-paid county court judge.

0:50 I run Adliora, an AI consultancy that helps small and mid-sized firms adopt AI safely, profitably and within the rules.

1:00 Every Monday I take the developments from the past week, separate the noise from the parts that touch your practice, and set out in plain terms what a sensible firm does about them.

1:11 Let us get into the week.

1:13 Four stories this week, and one of them dominates. On Wednesday the twenty first of October the Law Society published its practice note, Using artificial intelligence in legal practice.

1:25 It runs to fifty eight pages and sets out how the duties you already carry apply when your firm uses AI.

1:33 That is the main story. On the same day Legora, the legal AI company, announced that it had raised five hundred and fifty million dollars at a valuation of eight point eight billion.

1:44 It says its annual revenue went from one million dollars to one hundred million in under eighteen months, and that in-house legal teams made up more than forty per cent of its new customers in the last quarter.

1:57 Money on that scale pays for sales teams, so hold on to what this episode says about choosing a tool.

2:04 On Tuesday Womble Bond Dickinson, a firm of thirteen hundred lawyers with offices across the UK and the US, launched a unit it calls AI Labs.

2:14 It builds tools for structured, repeatable work such as regulatory monitoring and contract review.

2:20 And last Monday Law News reported an Opinium poll for Uswitch of two thousand UK adults, carried out in July.

2:30 Thirty nine per cent said they would consider asking AI rather than a solicitor about a legal question, and the Law Society note speaks to that, as you will hear.

2:41 Start with what the note is and what it is not. The Law Society represents solicitors and does not regulate them, so a practice note creates no rule, and the note says in terms that it neither promotes nor discourages the use of AI.

2:57 Its value lies in carrying the general duties the SRA enforces into the documents a firm already produces, from the client care letter to the bill and the contract with the supplier.

3:09 When a regulator, an insurer or a judge later asks what a reasonable firm did in the autumn of 2026, this is the document they will reach for.

3:19 The Society's headline was that AI is not a legal entity and that responsibility remains human.

3:26 Its president, Brett Dixon, put it this way: professional responsibility cannot be transferred to any AI system.

3:33 You heard the same from the SRA in its warning notice in August.

3:37 The note goes further on four matters that touch a small firm directly.

3:41 The first is what you tell clients. The note expects solicitors to give clients appropriate information about their use of AI.

3:50 It suggests that your client care letter or terms of business say how far the firm uses AI, which systems it uses, for which tasks, and how client information is protected.

4:02 If a client asks how AI was used on their matter, you should be able to give a clear and accurate answer, covering whether their information went into the system and whether a solicitor reviewed what came out.

4:15 Consent is a separate question, and the note draws the line with care.

4:20 Using AI inside your own secure systems, where client data does not train the model, is not likely of itself to need informed consent.

4:30 Transparency covers it. Using client information in a way that puts confidentiality at risk, training an AI system on it among them, is likely to need the client's specific consent.

4:42 The note runs the other way too. You can ask clients to tell you when material they hand over came from a chatbot, and check it before you rely on it.

4:53 If four in ten adults would consider asking a machine before they ask you, some of what lands on your desk was written by one.

5:01 The second is the bill. On time-based fees, you should not bill for the time a task would have taken had AI not been used.

5:09 Time spent operating the system, reviewing and verifying its output, revising it and applying your judgement can be charged where your client care letter allows it.

5:18 A separate technology charge has to be explained and agreed in advance, and it must not recover through a fee what your hourly rate already covers.

5:28 For fixed and capped fees the note asks for a clear scope and an overall fee that stays justified.

5:35 The third is checking the work, and it is the passage I would read twice.

5:40 The note says that an AI system cannot reliably validate the accuracy of its results on its own, and that relying on AI to check AI output will not satisfy your professional obligations if the output turns out to be wrong.

5:55 It names automation bias, the habit of trusting a confident machine, and suggests you review AI output with the scepticism you would bring to a trainee's draft.

6:06 It adds that oversight grows more important as systems act with more autonomy, and it names AI agents.

6:14 A tool that drafts a paragraph is one thing. A tool that works through a whole file and proposes the next step needs a supervisor who knows it is there.

6:23 The fourth is choosing the tool, and here the note says something smaller firms need to hear.

6:28 The risks of free chatbots are well known, but paid tools need scrutiny as well.

6:33 Before you adopt one, the Society expects you to consider how you access it and on what terms of service, how your data is stored and shared, where the provider operates, how likely it is to produce misleading results, and whether it holds the information it needs for the relevant jurisdiction, which for you means the law of England and Wales.

6:53 You can put those five questions to any supplier in writing this week, and a supplier who cannot answer them has told you something.

7:00 Two further points deserve a mention. On privilege the note advises caution and points to an Upper Tribunal decision this year which said that uploading confidential documents to a public AI tool placed them in the public domain.

7:14 And on negligence, the Society says that neither it nor the SRA has issued guidance on when a solicitor should use AI, and that it knows of no positive duty to do so.

7:25 Artificial Lawyer accepted the point about responsibility, then put a question against the line that AI-generated work should receive appropriate human oversight: define appropriate.

7:36 It also argued that some junior work can now go to a tool, with the right checks.

7:41 The first is a fair criticism. The note tells you that oversight must match the risk of the task, and it does not tell you how many pages a partner has to read.

7:57 A word about what I do. Adliora advises small and mid-sized law firms on adopting AI without falling foul of the rules that govern you. Where to begin, which tools earn their place, what your duties on supervision and confidentiality mean in practice, and how to write a policy your people will follow. If this subject sits on your desk and nobody in the firm owns it, that is the problem I am paid to solve.

8:31 For a firm of five to fifty people, the note reaches three documents and one habit.

8:37 Your client care letter comes first. Most of the client care letters I have read this year say nothing about AI, and after Wednesday that silence reads as a decision.

8:47 A client who later learns that a tool drafted their lease, and who was never told, now has the Law Society's own note to point to.

8:56 The remedy is a short paragraph, written for the systems your firm uses rather than the ones a vendor once promised.

9:04 Your time recording is the second. Suppose a fee earner records two hours on a review that took twenty minutes with a tool and forty minutes of checking.

9:14 On the note's reasoning the client pays for the hour, not the two.

9:19 Ask your team how they record AI-assisted work today, and expect to find that nobody has decided.

9:26 Firms on fixed fees escape the arithmetic but not the question, because the overall fee still has to stay justified.

9:35 The third is your supplier file, which in the small firms I have seen holds an invoice and a set of terms that nobody read.

9:43 The five questions in the note give you a file worth keeping, and they are the questions an insurer or the SRA is likely to ask after something has gone wrong.

9:54 The habit is supervision. A draft arrives looking finished, the formatting is clean, the citations sit in the right form, and the reader relaxes.

10:03 That is automation bias in an ordinary office, and the SRA's warning notice in August and a run of tribunal decisions this year show where it ends.

10:14 The note's answer is to treat machine output as a trainee's draft.

10:18 In a small firm that only works if the supervisor knows which parts of the draft the tool produced, which brings you back to your AI policy and whether it requires anyone to say so.

10:30 Then there are the juniors. Mr Dixon warned of a catastrophic gap of future legal skills and knowledge if the profession leans too hard on these tools.

10:40 A firm that hands first drafts to a tool has to decide how a trainee now learns to draft, and in a firm of twelve that decision falls on one training principal with no graduate scheme behind them.

10:55 Here is the order in which I would take it. Have one partner read the note this week.

11:00 Fifty eight pages is an afternoon's work. Then draft the client care paragraph, and keep it to what is true for your firm.

11:09 Name the AI tools you use and the work they help with, say that a solicitor reviews the output, and say that client data does not train the provider's model if that is the case.

11:22 Ask the client to tell you if anything they send was produced with AI.

11:27 If you cannot honestly write the sentence about training, you have found your first problem.

11:33 Next, send the note's five questions to each AI provider you pay, and keep the answers on the supplier file.

11:41 Settle the time recording rule with your team, in writing. Fee earners record the time spent operating the tool and checking its work, and never the time the task would once have taken.

11:54 Finally, put the date of the note in your AI policy, so the policy shows the firm read it.

12:01 Here is the one instruction to take from today. Before the end of next week, add a paragraph on AI to your client care letter and terms of business, and have a partner sign it off against the Law Society note.

12:16 The bench note this week comes from the regulator rather than a court.

12:21 On Thursday the SRA confirmed that its rules separating the compliance officer roles from the person who holds the power in a firm will not come in as planned in January.

12:32 It paused them in September after feedback that they carried unintended consequences for smaller firms and sole practices, and it promises firms a lead-in time before any change.

12:46 In the same statement it said it is looking at proactive supervision of firms it identifies as higher risk, seeking assurance that their governance arrangements are in place and that their compliance functions can operate independently.

13:01 Read it as a judge reads it. A pause is not an acquittal. Stephen Larcombe of the Conveyancing Task Force called it a pause, not a resolution, and he is right.

13:14 The statement does not mention AI. When the SRA asks a higher-risk firm to show its governance, though, the question of who owns the firm's use of AI sits inside the answer, and the compliance officer is the person who will be asked.

13:31 A firm that uses the pause to write down its AI decisions will have something to show.

13:44 And a word about my technological venture, so you know where I stand. Bracton is a case management system for firms in England and Wales with agentic AI designed in from the ground up, rather than bolted onto software written twenty years ago. Overnight it reads every live matter, so the morning list is waiting with the action drafted and the research carrying a source you can open. A person signs everything off. Bracton legal dot co dot uk.

14:15 That is the week. If you take one point from it, take this. A person in your firm has to own what these tools produce, and that will not change whatever the tools cost or how capable they become.

14:30 If you want an honest view of where your firm stands, there is a two-minute readiness check on the Adliora site, and you are welcome to send a short message and talk it through with me directly.

14:42 I am Saqib Khan, this has been The Legal AI Brief, and I will be back next Monday with the week that was.

14:50 Thank you for listening.

14:51 The transcript of this episode, with a timestamp on every line, sits beside the two minute readiness check at adliora dot com. Have a great week.