Three groups of people are currently using AI in and around fitness-to-practise work: registrants facing a case, students training to do this job, and us.
I have quite different things to say to each. What follows isn’t a warning about AI in general — we’ve built most of Sequentus with it — but the uses vary enormously in how well they’re going, and it’s worth being specific about which is which.
Registrants
Almost every client has run their case through AI before they speak to us, hoping for a definitive answer. They very rarely get one.
AI draws on an enormous pool of general information, while every case turns on its own facts. Outcomes vary wildly — not least depending on the quality of representation at the hearing, or the absence of any. A confident general answer about what happens in cases like yours is not an answer about your case.
We’ve had clients attempt to instruct us using AI, and that usually creates more problems than it solves. Hallucinations are common, and a good deal of our early work then goes into dissuading a client from something they’ve been told with total confidence by a machine.
But the most damaging use by far is reflections.
Insight is decisive in fitness-to-practise cases, and a great many clients now ask AI to write their reflective statements. It is the worst possible approach.
A badly written reflection that reads as genuinely held is worth a thousand times more than a polished one that is clearly composed by AI. Panels sit on these cases day in, day out — they can spot it as immediately as I can. Sometimes it’s the formatting that gives it away. Far more often it’s simply elegant words with very little meaning behind them.
And this is an adversarial system. When a registrant is cross-examined, it becomes very plain very quickly whether their reflections are genuinely held or whether they are saying what they believe needs to be said. A reflection has to survive being asked about. Nothing generated for you will.
Students
Students tend to assume AI use is obvious and acceptable, on the basis that it’s where the profession is heading.
My first caution is prosaic: unless you’re working on a corporate subscription, running case material through a consumer chatbot may well be a breach of confidentiality. That alone should stop the practice in its tracks.
The second is the same point I make to clients. AI has broad and often well-curated knowledge, but it knows very little of the practical and professional considerations that a lawyer brings to bear as professional judgement.
The third is more fundamental. Lawyers are thinkers and analysts. If you aren’t willing to develop those skills, and would rather delegate them, the law may not be the right career. My advice to students is to try to outpace AI and understand where the future opportunities are, rather than take the quick win of getting a task done without much hassle. And if you never find your own way of doing the work, life becomes considerably less enjoyable when you’re carrying the daily stress of the job — and remain professionally responsible for work you didn’t really do.
There’s a heavy emphasis on professional ethics in both barrister and solicitor training. But much like our healthcare clients, professional standards don’t seem to register as readily as they might, until something goes wrong.
Us
Given all of that, it would be inconsistent not to be clear about our own use.
We run an entirely bespoke technology stack, built with AI on enterprise-grade secure data infrastructure. That includes our own case management system, our student and training platforms, and the automation that connects them — workflows, document handling, and matter progression that in most firms of our size are still done by hand, if they’re done at all.
None of it is off-the-shelf, and none of it is a compromise. It is built precisely around how regulatory defence work actually runs, which is something no generic practice management product has ever managed. Building it this way has put us some distance ahead of firms many times our size, and that isn’t an accident of enthusiasm — it’s a deliberate strategic choice about where a small organisation can compete.
Client work is a different matter entirely. I want our people spending their time thinking and analysing, and almost all of our client work is the product of somebody sitting at a desk doing exactly that. What AI does is take the mundane parts off their plate so that more of their attention goes where it should.
There are exceptions where it has earned its place directly. Data analysis through our corporate accounts has produced genuinely valuable insight in areas like the assessment of clinical data, or comparison across different documentary sources. But it is humans who bring that together and find ways of using it for the client’s benefit.
AI will inevitably play a large part in the future of this profession, and for our size we’re well ahead of the curve. My optimism is tempered by my own obligations as a practising lawyer, and by knowing that the buck always stops with me.
We’re currently working on systems that will let registrants use AI to help progress their own case and build their understanding of the fitness-to-practise process. It’s an emergent idea, and it isn’t designed to phase out lawyers. It’s designed to make the process more accessible at a much lower cost — which has been the point from the beginning.
The short version
Use it for the work that doesn’t require judgement. Don’t use it for the work that is judgement — and above all, don’t use it to tell a panel what you think.
Sequentus Legal Services represents healthcare registrants before the GMC, GPhC, NMC and other regulators, on fixed fees.
