Episode 1. Clients are asking AI first, what law firms should do.

0:08 Three thousand two hundred and eighty seven people in this country were asked what they did when a legal problem landed on them.

0:15 One in six had put it to an AI chatbot. Among the youngest group, those aged eighteen to twenty four, it was better than one in four.

0:24 Almost none of them started with a solicitor.

0:27 Good morning, and welcome to The Legal AI Brief, the Monday round of the AI news that matters to smaller law firms.

0:35 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.

0:47 I run Adliora, an AI consultancy that helps small and mid-sized firms adopt AI safely, profitably and within the rules.

0:58 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.

1:10 Let us get into the week. Four things happened worth your attention.

1:15 On the first of September the legal charity JUSTICE and the Administrative Fairness Lab published research showing that a sixth of people facing a legal problem now consult AI about it, and that the youngest respondents did so at more than twice the rate of those in their late fifties and early sixties.

1:35 On the fourth of September the Solicitors Disciplinary Tribunal struck a registered foreign lawyer from the register after he put citations before the regulator that did not support what he had cited them for, in his own disciplinary proceedings, and then wrote an email of apology that carried further false material.

1:55 The tribunal says it is the first time a lawyer's use of AI in legal proceedings has been litigated before it.

2:03 Applications for the government's advisory AI Growth Lab in legal services close at one minute to midnight on the twenty seventh of September.

2:12 The sandbox puts firms and lawtech businesses in front of the SRA, the Legal Services Board, the Council for Licensed Conveyancers and the Information Commissioner's Office together, and it grants nobody an exemption from a single existing obligation.

2:30 And in Europe, the AI Office in Brussels and the national market surveillance authorities begin their first scheduled wave of compliance inspections this month, with recruitment screening tools, credit assessment systems and clinical triage first in the queue.

2:48 The main story is the first of those, because it reaches your fee income rather than your compliance file.

2:54 The research comes from JUSTICE, the law reform charity, working with the Administrative Fairness Lab.

3:00 They surveyed three thousand two hundred and eighty seven people across the United Kingdom, with the fieldwork running from the end of last year into January.

3:09 The headline is that around one in six people who ran into a legal problem consulted an AI chatbot about it.

3:16 Broken down by age, twenty six per cent of those aged eighteen to twenty four had done so, against ten per cent of those aged fifty five to sixty four.

3:26 Read that carefully, because the number that matters is not the twenty six per cent.

3:31 It is the shape of the curve. The generation that will bring you their first house purchase, their first employment dispute and their first family matter over the next decade has already formed the habit of asking a machine first.

3:47 The generation currently sitting in your client base has not.

3:51 Nothing about that gap suggests it closes in your favour on its own.

3:56 What were people asking for? Four things came up repeatedly.

4:00 They asked the chatbot to explain legal jargon, which is to say the language in the letters they had been sent, some of which came from firms like yours.

4:08 They asked it to check advice they had already been given by a lawyer.

4:13 They asked it to draft complaints and emails. And they went to it for reassurance, which is a plain way of saying they were frightened and wanted somebody to talk to at eleven o'clock at night.

4:25 Sit with the second of those for a moment. People are taking your advice and running it past a chatbot before they act on it.

4:32 Not because they distrust you in particular, but because checking a thing has become free and instant, and people check.

4:40 Your written advice is now read by a second reader you never see, one that has no file, no context and no duty to anybody.

4:49 If your letter of advice is hedged, dense, or silent about why you reached the view you did, the machine will fill that space with something confident and possibly wrong, and your client will believe the confident thing.

5:03 The researchers were careful about what they found, and the care is worth repeating.

5:08 Only six per cent of the people using these tools relied on the tool alone.

5:12 The average user was combining it with search engines, government websites, legal advice centres, family and friends.

5:20 So this is not a picture of people replacing lawyers with chatbots.

5:24 It is a picture of the chatbot becoming the first stop and the constant companion in a journey that still, for many people, ends somewhere else.

5:34 That is the opening, and it is also the risk. The risk is in a finding that ought to trouble anybody who advises the public.

5:41 The research indicates that designing a chatbot to be friendly and warm makes it more likely to reinforce a false belief the user brings to it, and that the people who sound most vulnerable draw the least accurate answers.

5:55 Turn that round and it says the person with the weakest position, least able to afford to be wrong, gets the worst information delivered in the kindest voice.

6:04 Ellen Lefley, deputy legal director at JUSTICE, put the broader point plainly, saying that the question is not whether AI becomes part of people's legal journeys, because it already is.

6:16 Dr Jed Meers of the Administrative Fairness Lab made the same observation about scale, that people are turning to these tools across all kinds of legal problems and in significant numbers already.

6:28 The subject matter tells you where this bites hardest. The problems people brought to chatbots cluster in housing, employment, debt, consumer disputes and family, which is to say largely in the areas that legal aid no longer reaches.

6:43 These are not people choosing a machine over a solicitor. Many of them are people who had no realistic route to a solicitor, and who now have something that answers.

6:54 Whether what it tells them is right is a separate question, and one they are in no position to judge.

7:06 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.

7:41 Three consequences follow for a firm of your size, and none of them is about buying software.

7:47 The first is that your first conversation with a new client has changed without your agreement.

7:52 The client arriving at your door has been briefed. Sometimes that briefing is sound and your job is to confirm it and move on, which is faster and cheaper for both of you.

8:03 Sometimes it is wrong in a way that has already shaped what the client thinks the case is worth, what the time limits are, or whether they have a claim at all.

8:12 Either way you are now correcting or confirming a prior view rather than forming the first one, and if you do not know what that prior view is, you will spend the first meeting talking past it.

8:25 The second is that the enquiries reaching you have changed in character.

8:29 If a sixth of people are getting a serviceable answer to the simple question from a free tool, the simple question stops arriving.

8:37 What arrives instead is the harder problem, the one the chatbot did not resolve, the one where the client sensed the answer was thin.

8:46 That is better work at better margins, and it is also work that requires more of your judgement per hour and rewards a fee model built on outcome rather than on the routine document.

8:58 The third is the one nobody enjoys. Some of the people asking chatbots about housing possession, an employment dismissal or a debt they cannot pay would have been your client at a modest fee, and will now act on a machine's answer and lose.

9:12 You are not responsible for that. You are, though, in a market where the alternative to your advice is no longer nothing.

9:19 The comparison a client makes is between your fee and a free answer that sounds assured, and the way you win that comparison is by being clear about what you are selling, which is somebody who carries the responsibility for being right.

9:34 Start with your own letters. Take the last five pieces of written advice that left your firm and read them as a nervous client would at midnight.

9:43 If the reasoning is buried, if the risk is hedged into invisibility, if a lay reader would come away unsure what you recommend, then you have handed the chatbot the job of explaining you to your own client.

9:56 Write so that the second reader has nothing to add. Second, put a line in your client care information saying that clients are welcome to check what they are told, and that if anything they read elsewhere contradicts your advice they should come back to you rather than act on it.

10:13 That costs you nothing and it converts a silent second opinion into a phone call you can answer.

10:19 Third, decide now how you handle the client who brings you a machine-drafted letter, particulars of claim or complaint.

10:27 Somebody in your firm should be able to say what you charge to review it, what you will not put your name to, and where you draw the line between fixing a document and taking on the case.

10:37 This arrives on your desk unannounced, and a position invented in the moment is usually the wrong one.

10:44 Here is the one instruction to take from today. This week, ask every new client a single question at the first meeting.

10:52 Before you came to us, did you ask a chatbot about this, and what did it tell you.

10:58 Write down what they say. In a month you will know more about how your own market has moved than any survey can tell you.

11:07 The tribunal decision of the fourth of September deserves a judge's reading rather than a headline.

11:13 Reporting names the lawyer as Abhishek Kumar, a registered foreign lawyer facing the tribunal over a 2024 conviction.

11:21 He acted for himself. The written answer he sent the SRA contained quotations and citations that did not support the propositions he advanced.

11:31 When the regulator's counsel spotted it, he emailed to explain and apologise, and that email had been drafted with AI too, and carried more false material.

11:42 He told the tribunal he had not meant to mislead anybody and did not have the expertise to verify what the tool produced.

11:51 The tribunal gave that the answer it had to. A regulated individual is obliged to meet the standards of the profession, and an inability to check is not a defence to putting unchecked material before a decision-maker.

12:04 The tribunal found a high degree of culpability and struck him from the register.

12:10 What makes the case instructive is not the first mistake but the second.

12:15 The apology repeated the error, which tells you the process had never changed.

12:21 Sanction follows the pattern, not the slip.

12:30 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.

13:02 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.

13:16 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.

13:28 I am Saqib Khan. This has been The Legal AI Brief, and I will be back next Monday with the week that was.

13:36 Thank you for listening.

13:37 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.