This audience searches within hours of an arrest, a charge or a police contact. The site has to help immediately without exploiting the panic that brought them there.
The first screen has one job
Somebody searching after an arrest needs four things established immediately: that the firm covers this jurisdiction, that it handles this type of matter, how to reach a human right now, and what not to put in writing. Everything else on the page is secondary.
That last point is routinely missing. Criminal defence sites invite detailed case facts through unsecured web forms, which creates a record the client did not intend and the firm did not want.
What must not appear
No promise of dismissal, acquittal, bond outcome or charge reduction. No published allegations or identifiable client narratives without exceptional consent and review. No case-result carousel implying a typical outcome.
Demonstrate experience factually instead: courts served, years admitted, prior roles where relevant, certified specialisations where the state permits the term, publications, teaching, speaking. Factual credentials are both permissible and more persuasive to a family member doing the research, which is frequently who is actually on the site.
Where AI is and is not appropriate
The credible product claim in this category is narrow: technology can help capture initial contact, route urgent enquiries and ensure nothing sits unanswered overnight. It should never appear to provide an assessment of a criminal matter to a member of the public.This is not a theoretical caution. Concern within the profession about AI and the unauthorised practice of law is rising sharply, and an interface that looks like it is evaluating someone’s charges is exactly the shape of the problem
The intake process
Quick routing, conflict check, secure note capture, and follow-up by a lawyer or trained staff. Define what happens to a call at 2am, because in this practice area that is a normal arrival time rather than an edge case.
| In Thomson Reuters’ 2026 research, the share of lawyers describing AI as a major threat in relation to the unauthorised practice of law rose to 50%, from 36% the year before. Separately, a Stanford RegLab study published in the Journal of Empirical Legal Studies found leading legal research AI tools producing hallucinated output between 17% and 33% of the time.Thomson Reuters Institute, 2026 AI in Professional Services Report. Magesh et al., Journal of Empirical Legal Studies, Stanford RegLab. |
W
The line to put on the page
“Speak to our intake team now. Please do not submit case details through this form.” It costs you nothing and it tells a sophisticated reader the firm knows what it is doing.
Review your after-hours intake
Most criminal defence firms discover their published availability and their actual availability differ.
Sources
- Thomson Reuters Institute, 2026 AI in Professional Services Report.
- Magesh et al., Journal of Empirical Legal Studies, Stanford RegLab study on legal AI hallucination rates.
