AI for solicitors and law firms

Client confidentiality and privilege

Confidentiality is where most firms’ worries start, and rightly. In its August 2026 warning notice on the misuse of AI, the SRA raised concern about “confidential client information being entered into public AI tools, risking potential breaches of confidentiality”. It says client information should only go into AI systems “where appropriate contractual, technical and organisational safeguards are in place”.

In practice that means three things. Use AI tools whose terms keep your data out of model training and under your control. Limit what each tool can see to what the task needs. And for the most sensitive matters, use private AI that runs on hardware or cloud accounts your firm controls, so nothing goes to an outside AI provider at all.

Hallucinations and checking the work

AI tools can produce convincing text that is simply wrong. The SRA notice warns that they “can produce ‘hallucinations’, generating fictitious cases, references or seemingly factual assertions”, and the courts have already dealt with cases where invented citations reached them.

We build tools that make checking easy rather than optional. Summaries link back to the page they came from, research answers show their sources, and nothing goes to a client or a court without a solicitor reviewing it. As the SRA puts it, “You remain responsible for the advice and outputs you provide to clients as a legal professional.”

UK GDPR and the SRA

Using AI doesn’t change your obligations; it adds another system they apply to. Under UK GDPR that usually means a lawful basis, an updated privacy notice and a data protection impact assessment for higher-risk uses. On the regulatory side, the SRA expects you to stay accountable for work done through AI tools and to supervise it effectively.

We handle the technical side: which tools, what settings, what access and what logs. We don’t give legal or regulatory advice, and your COLP and data protection lead stay in charge of those decisions. The SRA’s notice and the Law Society’s guide, Generative AI: the essentials, are good places to start.

An AI policy for your firm

The SRA notice says it is “incumbent on all managers and employees of SRA regulated firms to understand their obligations”. A short, practical AI policy is the simplest way to make that happen.

We help firms write one that people will actually follow: which tools are approved, what can and can’t go into them, how outputs are checked, who supervises, and how to report a mistake. We then back it with training, so fee earners and support staff understand the reasons as well as the rules.

Private AI for law firms

For many firms, private AI is the most comfortable place to start. A ChatGPT-style assistant runs on your own hardware or in your own cloud, staff sign in with their work accounts, and it answers questions from your precedents and know-how without client data leaving your control.

It handles drafting, summarising and transcription well. For the hardest research tasks, we’ll be honest about where a specialist legal research product or a cloud model does better, and help you use those safely. How private AI works.

Where AI helps: drafting, review, conveyancing, dictation and intake

Most firms start with one of five jobs. Drafting from precedents saves fee earners from a blank page. First-pass review of contracts and bundles gets to the key clauses faster. Conveyancing chasing keeps files moving without someone spending their morning on follow-up emails. Private dictation and transcription replaces outsourced typing. And intake triage gets new enquiries to the right person with the details already captured.

We usually suggest a short pilot on one of these, measured against how the work is done today, before rolling anything out across the firm.