AI in Conveyancing: What the SRA Expects of Firms That Use It

The Solicitors Regulation Authority (SRA) has no rules written specifically for artificial intelligence (AI). A firm that uses AI is held to the same SRA Principles and Codes of Conduct as before, and the SRA's warning notice of 17 August 2026 says that using AI "does not diminish or transfer your professional responsibilities". For conveyancers that means checking outputs, protecting client confidentiality, supervising the people and tools involved, and assessing money laundering risk before a new tool goes live.

I work at Lex Sentinel, which builds AI tools for conveyancers, and nothing below depends on a product. This is general information for firms regulated by the SRA in England and Wales and is not legal advice. I will update it when the SRA or the Information Commissioner's Office (ICO) publishes something new. Our guide to the firm-level anti-money laundering (AML) framework covers the money laundering side of the same question.

Does the SRA regulate the use of AI in conveyancing?

The SRA does not regulate the use of AI in conveyancing directly. It regulates outcomes and applies its existing rules, and its compliance tips on technology say firms may use any technology they think appropriate, subject to the Principles and their other obligations. The page also says the Compliance Officer for Legal Practice (COLP) is responsible for regulatory compliance when new technology is introduced, that firms should be clear who clients can turn to if the technology fails, and that "it should always be made clear to clients where they are interfacing with AI", which most naturally covers client-facing tools. The tips are non-binding. The body of that page is dated 9 February 2026, though its metadata gives an earlier date, so the live page is the safer reference.

Nothing I found sets a technical standard for AI, approves particular tools or prohibits a particular use in conveyancing.

What does the SRA's warning notice on AI say?

The SRA's warning notice treats AI misuse as a conduct risk for every regulated firm and individual. The notice, published on 17 August 2026, deals mainly with two things: fabricated or inaccurate outputs, and client confidential information going into AI tools. It says client data should go into a tool only where there are appropriate contractual, technical and organisational safeguards, and that public tools risk waiving privilege. Supervisors remain accountable for the work of junior staff who use them.

The notice cites Principles 1, 2, 4, 5 and 7 and a list of paragraphs in the Codes of Conduct. I haven't checked each paragraph against its text on the SRA's rules site, so I don't summarise them here. In a press release the same day, the SRA reported 42 reports of potential AI misuse between July 2025 and July 2026, with investigations into citations, supervision and confidentiality. None of that is specific to conveyancing, and I found no published disciplinary outcome yet.

What do the courts expect when a lawyer uses AI?

A lawyer is expected to check anything an AI tool gives them before relying on it. In R (Ayinde) v London Borough of Haringey and Al-Haroun v Qatar National Bank ([2025] EWHC 1383 (Admin), 6 June 2025), the Divisional Court said lawyers have a professional duty to check the accuracy of research done with AI tools. It said existing guidance was not enough on its own and that leaders and regulators must take practical steps to make sure practitioners comply. In Al-Haroun, 18 of the 45 citations did not exist and the solicitor was referred to the SRA.

Both cases arose in litigation, but the principle carries over to any document a conveyancer signs: an AI-drafted enquiry, report on title or summary of a lease is your work product once you send it.

How does supervision apply to AI-assisted work?

AI-assisted work is supervised in the same way as any other delegated work. The SRA's effective supervision guidance, updated on 12 June 2026, expects appropriate human review, scrutiny and professional judgement of AI-assisted outputs, and leaves ultimate responsibility with the authorised individual. The SRA's 2023 Risk Outlook report on AI adds two points that still read as sensible practice: firms should not rely on an AI system to check its own accuracy, and responsibility can't be passed to an IT team or a supplier.

In practice, a fee earner should be able to say who reviewed an AI-assisted output, what they checked and what they changed. If the file can't show that, the supervision point is weak.

What do the money laundering rules add?

The money laundering rules add a specific duty to assess the risk before a firm adopts the tool. Regulation 19(4)(c) of the Money Laundering Regulations 2017 requires policies that ensure appropriate measures are taken, in preparation for and during the adoption of new technology, to assess and if necessary mitigate any money laundering or terrorist financing risk. In its guidance on policies, controls and procedures, published on 22 October 2026, the SRA says 40% of the policies it reviewed over the last three years were missing or incomplete on this, the most commonly missing area. It also says, in the context of consultants and template policies, that you cannot pass the regulatory risk of non-compliance on to a third party, and the point applies equally to suppliers.

That guidance does not mention AI by name. It reaches AI because AI is new technology. If you use a tool for identity checks or screening, adopting it is also a reason to update your policies.

What does data protection law require?

UK data protection law requires a data protection impact assessment (DPIA) where processing is likely to result in a high risk, and adds safeguards where a significant decision about a person is solely automated. The ICO treats innovative technology such as AI as a high-risk indicator when it is combined with another criterion, such as scoring or sensitive data. If in doubt, do one, and record your reasons if you decide not to. Its page on when to do a DPIA, dated 13 October 2025, says it is under review because of the Data (Use and Access) Act 2025.

That Act introduced new Articles 22A to 22D on automated decision-making, in force from 5 February 2026 according to section 80 of the Act. A decision is solely automated where there is no meaningful human involvement. Article 22B restricts such decisions where they are based on special category data. For significant decisions the safeguards include information about the decision, the chance to make representations, human intervention and a right to contest it. The ICO consulted on draft guidance on automated decision-making between 31 March and 29 May 2026. I haven't confirmed whether final guidance has been published, so the draft is the most I can point to.

A conveyancing example: a tool that scores a client's risk and routes the file with no one looking at the score is the kind of arrangement these provisions are aimed at. A tool whose score a named person reviews and can override is a different position, though the review has to be real.

What should a firm do before it adopts an AI tool?

Before adopting an AI tool, a firm should do the six things below. None is a rule in those words; this is how I would approach it.

1. Name the tool, what it is used for and who owns it, and record that in your firm-wide risk assessment and policies.
2. Assess the money laundering risk under regulation 19(4)(c) before go-live, and again when the tool changes.
3. Complete a DPIA where the processing is likely to be high risk, or record why not, and check where the data is processed and whether the supplier trains models on it.
4. Decide who reviews each type of output, what they check, and how that is recorded on the file.
5. Tell clients when AI is used on their matter, in line with the SRA's compliance tips, and offer a non-technology route where a client doesn't want it. Where a significant decision is solely automated, the data protection safeguards also include human intervention and a right to contest it.
6. Test the tool on synthetic files before it touches a real one, and keep the results.

About the author

Appan Pathmanathan is a solicitor and Chief Operating Officer of Lex Sentinel, which builds AI tools for UK conveyancers. He has worked in UK legal practice since 2001, including litigation involving money laundering and fraud, and conveyancing, and has been managing partner of a law firm. He delivers AML and compliance training. Connect on LinkedIn.

Sources

Checked on 25 October 2026.

1. SRA, Misuse of AI warning notice (17 August 2026)
2. SRA, Responsible use of AI press release (17 August 2026)
3. SRA, Compliance tips for solicitors
4. SRA, Effective supervision guidance
5. SRA, Risk Outlook report: artificial intelligence in the legal market (20 November 2023)
6. SRA, Anti-money laundering policies, controls and procedures
7. Money Laundering Regulations 2017, regulation 19
8. [R (Ayinde) v London Borough of Haringey and Al-Haroun v Qatar National Bank [2025] EWHC 1383 (Admin)](https://www.judiciary.uk/wp-content/uploads/2025/06/Ayinde-v-London-Borough-of-Haringey-and-Al-Haroun-v-Qatar-National-Bank.pdf)
9. Data (Use and Access) Act 2025, section 80
10. ICO, When do we need to do a DPIA?
11. ICO, Consultation on draft guidance on automated decision-making including profiling (2026)
12. Law Society, Generative AI: the essentials

Frequently Asked Questions

Can a conveyancing firm use AI under SRA rules?

Yes, a conveyancing firm can use AI under SRA rules. The SRA has no AI-specific rules and says firms may use any technology they think appropriate, subject to the Principles and their other obligations. The firm stays responsible for the output.

Do we have to tell clients we use AI?

No numbered rule requires it, but the SRA's non-binding compliance tips say it should always be made clear to clients where they are interfacing with AI. The Law Society advises agreeing with clients whether and how AI is used. Record your approach in your client care information.

Can we put client data into a public AI tool?

The SRA's warning notice says client data should go into a tool only where there are appropriate contractual, technical and organisational safeguards, and warns that public tools risk waiving privilege. The Law Society says generally not to feed confidential information into generative AI tools, and for free tools, none at all.

Does a firm need a DPIA before using AI on client files?

A firm needs a DPIA where the processing is likely to result in a high risk. The ICO treats AI as a high-risk indicator when it is combined with another criterion, such as scoring or sensitive data. If in doubt, do one, and record your reasons if you decide not to.

Who is responsible if an AI tool gets something wrong?

The firm and the authorised individual are responsible if an AI tool gets something wrong. The SRA says the use of AI does not transfer professional responsibilities, and that responsibility can't be passed to an IT team or a supplier.