What the SRA's June 2026 Supervision Guidance Means for Your Conveyancing Practice

Let me be direct about what this article is and is not. This is not a legal opinion on the SRA guidance. UtterConnection is not an SRA-regulated law firm, and nothing here constitutes regulated compliance advice. What this is, is a plain-English reading of the SRA's Effective supervision guidance as updated on 12 June 2026, separating what it requires from what we recommend on top, based on 22 years running conveyancing operations. Where I am inferring rather than quoting, I say so.

You should read the SRA's guidance in full and take qualified regulatory advice if you are uncertain about your specific obligations. What you should not do is read the headline and assume it does not apply to your firm because you have not "officially" adopted AI.

One obligation, and two things you have to decide for yourself

The guidance is written the way the SRA writes everything: outcomes-focused. It sets a standard and does not tell you how to meet it. So there is one hard obligation in it, and then a large amount of latitude that you have to fill in yourself. Firms get into trouble in the latitude, not the obligation.

1. The obligation: a named human keeps ultimate responsibility

This is the one must in the section: an authorised individual retains ultimate responsibility for any legal service delivered with AI assistance. Not the firm in the abstract, not the vendor, not the tool. A person. The SRA's Misuse of AI warning notice of 17 August 2026 says the same thing more bluntly: “AI has no separate legal personality.”

The practical problem for a conveyancing firm is that you cannot assign responsibility for something you cannot see. Your case management system, your email platform and your drafting tools are shipping AI features by default, and staff use free tools on their own initiative. None of that is mentioned in the guidance. We are telling you it matters because we have watched it play out, not because the SRA said so.

2. Our recommendation, not a requirement: write down why

Be clear about the status of what follows. The SRA does not require a written AI rationale, an AI policy, an AI register or a decision log. None of those words appear in the June guidance or in the August warning notice. Anyone telling you otherwise is selling you something, including anyone who tells you that and then sells you a governance engagement.

What the guidance does require is that you can show effective supervision and a responsible named individual. A short written rationale is simply the cheapest way we know of to be able to show it. It is evidence, not compliance. If you can evidence the same thing another way, do that instead.

3. Where the guidance stops and we start: making review visible

The guidance says outputs should be subject to “appropriate human review, scrutiny and professional judgement.” It does not define appropriate, and it does not say the review has to be recorded. That is our addition, and we will defend it, but we are not going to dress it up as a rule.

The argument for recording it is simple. A review that happens in someone's head is indistinguishable, on the file, from a review that never happened. If a matter goes wrong two years later, the question will be whether anyone checked, and the honest answer will be that nobody can remember. UK and R (on the application of Munir) v SSHD [2026] UKUT 81 (IAC) is the cautionary version: at [38] the tribunal held that a supervisor who fails to catch a junior's hallucinated citations is “likely to be more culpable” than a lawyer who fails to catch their own. The supervising solicitor there was referred to the SRA.

What the Guidance Does Not Require

It is worth being equally precise about what the guidance does not mandate, because a good deal of the commentary in the sector, including some aimed at selling you governance, is overstating the burden.

• The SRA does not prohibit the use of AI tools in conveyancing. The guidance is about governance, not prohibition.
• The SRA does not require firms to use specific AI governance software or technology platforms. Your documentation obligations can be met through properly structured Word documents, SharePoint records, or case management custom fields, as long as they meet the evidential standard.
• The SRA does not require firms to notify them of every AI tool in use. What it requires is that the firm could produce that information on request.

The CLC Position

The CLC did not publish anything in June 2026. Its AI and Technology Principles, eleven of them covering risk of harm, security, data use and privacy, impact assessment, capability and explainability, came earlier and are expressly non-mandatory. The CLC calls them an evolving document. What binds CLC firms is the CLC Code of Conduct, its AML requirements and UK GDPR. That said, CLC-regulated firms should not assume the SRA material is irrelevant to them: the supervision and accountability thinking is substantively the same, and it is where the sector is heading.

What PI Underwriters Are Asking

Separately from the regulator, insurers have started asking. Howden's 1 April 2026 solicitors' PII renewal review reports underwriters focusing on how AI is being applied across practice areas, what governance policies and procedures are in place, and whether staff have been trained. Two caveats we would rather state than have you discover: Howden describes a soft market with record capacity, not a squeeze, and it notes that full proposal forms are typically required only every three years, with short-form declarations increasingly accepted. So this is a question you should expect, not a cliff edge. The three things being asked are:

• Does the firm have a documented AI usage policy?
• Has the COLP signed off on the use of AI in client-facing work?
• Does the firm have a process for verifying AI outputs before they are sent to clients or third parties?

A firm that cannot answer yes to all three, with evidence on file, faces underwriter uncertainty at renewal. That uncertainty translates into premium increases, coverage restrictions, or, in the most exposed cases, coverage questions on claims arising from AI-related incidents.

The Speed Dimension

Everything above is framed as a risk issue, because that is how the SRA frames it. But there is a second dimension that the guidance creates space for, which is not being discussed enough in the sector.

A firm with proper AI governance in place, documented oversight, written rationales, human verification gates, is a firm that can use AI confidently and at scale. The governance is not the cost of using AI. The governance is what makes AI use commercially viable and governance-secured.

Consider a paralegal facing a leasehold management pack. Where the firm has decided which tool is approved, for what, and who owns the output, they use it and the review is on the file. Where it has not, they either avoid the tool because nobody has told them it is allowed, or they use it anyway and nothing is recorded. Same person, same tool, opposite risk profile. The governance is not the obstacle to speed. The absence of it is. We do not sell the tool, and we are not going to quote you a time saving we have not measured.

The firms that move fastest on AI adoption in 2026 and 2027 will not be the firms that ignore the SRA guidance. They will be the firms that implement it quickly and use it as the framework within which they build genuine operational advantage.

What to Do This Week

If you have not yet conducted a structured AI audit of your firm's technology environment, this is the immediate priority. Not because anyone is about to inspect you, but because you cannot assign responsibility for something you cannot see.

The audit does not need to take weeks. A structured session, a focused call built on an anonymous staff questionnaire run beforehand, covering your CMS configuration, your email and Microsoft 365 environment, your staff's actual tool usage (which requires psychological safety to surface honestly), and your existing policy documentation, will give you the information you need to act.

Start with the free AI Reality Check on the UtterConnection homepage. Six questions, and it returns a RED, AMBER or GREEN rating against the supervision and confidentiality expectations set out above. It is a self-assessment prompt, not a diagnostic and not a compliance verdict, but it will tell you whether you are operating with awareness of your exposure or without it.

Regulatory Notice: This article is published by UtterConnection Conveyancing Limited, an independent operations and technology management consultancy. It does not constitute legal advice or regulated compliance advice. All references to SRA and CLC guidance are for informational purposes only (including UtterConnection's operational assessment based on the SRA's Effective supervision guidance of 12 June 2026, its Misuse of AI warning notice of 17 August 2026, and its Risk Outlook report on AI in the legal market of 20 November 2023). You should take advice from a qualified legal professional or compliance specialist before making decisions about your firm's regulatory obligations.