Login Get started

AI for lawyers, UK

AI for lawyers in the UK that cites the judgment.

Research grounded in BAILII, Find Case Law and legislation.gov.uk with every authority linked and its treatment checked. Drafting as track changes in Word. One question asked across a whole disclosure bundle. Agents that watch limitation dates. All inside a practice management system built for solicitors, with the controls a COLP would expect.

Get started
Book a demo

What it does

The four jobs UK lawyers actually hand to AI.

Not “AI in the legal sector” in the abstract. The reading, drafting, checking and chasing that fill a fee earner’s week, done inside the file so the output does not need copying anywhere.

Research, England and Wales first

Plain-English questions answered from retrieved judgments and statutes as amended, each authority linked with neutral citation and subsequent treatment. Scottish and Northern Irish sources where BAILII holds them.

Drafting in Word

Redlines as native track changes, findings as comments, OSCOLA-formatted citations, and drafting from the firm’s own precedents rather than a generic form. See the Word add-in.

Disclosure and due diligence

Governing law, break clauses, change of control and limitation across an entire bundle, answered as a grid with every cell citing its page. See AI contract review.

Agents on the caseload

Limitation dates, CPR deadlines and unanswered client emails flagged in the matter by background agents, with approval required before anything is sent. See workflows and agents.

Regulation

What a UK firm has to be able to say about its AI.

Competence and supervision

Who checked it, and can you show that?

The SRA’s position is that a solicitor remains responsible for work produced with AI, and the High Court has been explicit about the consequences of filing invented authorities. ModulawAI answers from retrieved sources with linked citations, so checking is a click rather than a search, and every AI output is logged with the fee earner who reviewed it. That log is your supervision evidence.

Confidentiality and UK GDPR

Where does client data go, and who trains on it?

Nowhere it should not, and nobody. Client data is encrypted in transit and at rest, held in AWS data centres, never used to train models, and covered by a data processing agreement with a published subprocessor list. Access is role-based per matter. Read the DPA and subprocessors before the first upload; a COLP should not have to ask.

Client money and time

Does the AI’s work get billed?

Because the AI runs inside the matter, the time spent reviewing its output is captured in six-minute units against the file like any other work, and the research it produced stays attached to the matter for the file review. Practice management, client account handling and billing are covered on legal practice management software UK.

Practice areas

Where UK firms start.

  • Litigation: research with treatment checked, chronology from the bundle, disclosure review as a grid, and limitation monitoring.
  • Corporate and commercial: due diligence extraction across a target’s contracts, first drafts from the firm’s precedents, and redlines in Word.
  • Employment: settlement agreement drafting, tribunal deadline tracking, and research across Employment Appeal Tribunal decisions.
  • Private client: will and LPA drafting from templates, and estate documents generated from matter data.
  • Conveyancing and property: lease portfolio review for break clauses and rent review dates, and standard-form generation. See solicitor practice management.

The record

What the SRA has said, and what happened in court.

The leading English authority is the joined judgment in Ayinde v London Borough of Haringey and Al-Haroun v Qatar National Bank [2025] EWHC 1383 (Admin), handed down by the Divisional Court on 6 June 2025 under the Hamid jurisdiction. In the first case, grounds for judicial review in a homelessness matter misstated section 188(3) of the Housing Act 1996 and cited five authorities that did not exist, among them a fabricated El Gendi v Camden LBC. Counsel accepted she may have relied inadvertently on AI-generated results returned through a search engine. The court made wasted costs orders of £2,000 plus VAT against each of the two lawyers involved, £5,000 of the law centre’s fees were written off, and both were ordered to report themselves to their professional regulators.

The regulator followed. The SRA’s warning notice on the misuse of AI, published 17 August 2026, is short and unambiguous. It states that solicitors remain accountable for all work and advice delivered to clients whether or not AI has been used in its preparation; that reliance on an output of AI would not be a suitable defence; that supervisors must effectively supervise work done for clients and ensure the individuals they manage are competent; and that client information should only be entered into AI systems where appropriate contractual, technical and organisational safeguards are in place to protect confidentiality. It closes by warning that failing to have proper regard to the notice puts a solicitor at risk of disciplinary action. The SRA has separately reported dozens of AI-related conduct reports in the year to July 2026, covering inaccurate legal citations, supervision and confidentiality.

The Law Society’s position is complementary rather than contradictory. Its generative AI guidance for solicitors stresses human oversight as an ongoing requirement rather than a one-off sign-off, and its 2026 technology procurement guide warns that unsafe adoption exposes firms to risk, setting out steps for identifying need, briefing suppliers, negotiating contracts and evaluating tools after deployment. None of this prohibits AI. All of it assumes the firm can evidence what the tool did and who checked it.

How that maps onto a compliant deployment

  • Verification has to be cheap or it will not happen. Retrieval-grounded answers with a link on every citation turn checking into a click. A tool that summarises without sources shifts the entire verification burden onto the fee earner, who is the person under time pressure.
  • Supervision has to be evidenced. Every AI action in ModulawAI is logged with inputs, outputs and the fee earner who reviewed it. That log is what a COLP produces when asked how the firm supervises AI-assisted work.
  • Confidentiality has to be contractual. Read the data processing agreement and the subprocessor list before the first client file is uploaded, and require the same of any other vendor. Data is encrypted in transit and at rest, held in AWS, and never used to train models.
  • Client transparency is a firm decision. The SRA’s Risk Outlook suggests firms consider telling clients when AI is used on their matter and how. That belongs in the client care letter, and no software can write it for you.

Sources

FAQ

Questions, addressed.

Is it safe for a UK solicitor to use AI for legal research?

Yes, if the tool retrieves the authority before answering and links to it, and the solicitor checks the citation. That is the standard the courts and the SRA expect. It is not safe to file anything produced by a general chatbot that cannot show its sources.

Which UK sources does it use?

Judgments from BAILII and The National Archives’ Find Case Law, legislation as amended from legislation.gov.uk, and EU material from EUR-Lex where retained law is engaged. Coverage is strongest for England and Wales.

Can we use it alongside Westlaw or Lexis?

Yes. Many firms keep a subscription database for editorial commentary and use ModulawAI for question-led research, drafting and review inside the matter. There is also a scoped REST API and an MCP server for connecting other tools.

How is it priced for a UK firm?

Per seat with usage-based credits, in pounds. See pricing.

Does it work for barristers and in-house teams?

Yes. Barristers use the research and drafting without the practice management modules; in-house teams use matter tracking, contract review and the precedent library. Permissions and modules are configured per workspace.

Related

Keep reading.

Bring a point you have already researched.

Book a demo and check every citation we return against your own note.

Get started
Book a demo