Emerging risk · AI & PII

PI insurance for AI-using professionals — how insurers view AI-assisted advice

Reviewed by Matthew Bartlett, Director, Apex Insurance Brokers Limited · Published 14 July 2026

AI tools are now embedded across UK professional services — drafting, research, document review, coding, tax preparation, financial modelling, advisory reasoning. This page sets out how UK PI insurers currently treat AI-assisted work, where coverage gaps sit, and what disclosure is expected at renewal.

How PI insurers currently think about AI-assisted advice

As of mid-2026, most PI insurers do not exclude AI-assisted work as a class — the professional using the AI carries the same civil liability whether the work is AI-drafted or hand-drafted. The insurer cares less about the tool and more about the supervision.

But: some insurers are beginning to ask specific questions at renewal about AI use. Some emerging wordings introduce sub-limits or excess loadings for AI-driven claims. The market is not settled.

The four scenarios where AI creates PI exposure

Unchecked AI-generated advice. Professional uses AI to draft advice, delivers it without review. AI hallucinates a case citation, misstates a rule, or gives factually wrong guidance. Client acts on it. Standard professional negligence claim, PI responds.
AI as sole decision-maker. Firm uses AI to make client-affecting decisions without human oversight (automated underwriting, automated conveyancing document generation, algorithmic advice). Insurer scrutiny sharpens.
AI training on client data. Confidentiality breach if AI provider trains on client-identifiable data without consent. Not a classic PI claim, but adjacent to cyber and data-protection cover.
AI-assisted work misattributed. Professional signs off on AI-generated work as their own without appropriate supervision. Complaint or claim challenges the ‘professional judgment’ basis of the advice.

Disclosure at renewal — what to expect

Some insurers now ask specifically about AI tool usage in the proposal form. Answer honestly and fully.
Where AI is used, describe: (a) which tools, (b) which work types, (c) supervision protocol, (d) whether AI-generated output is reviewed by a qualified professional before delivery.
Where the firm uses AI for client-affecting decisions without human oversight, expect underwriter follow-up.
Do not describe AI use in vague terms. Underwriters read this and price accordingly — specificity helps.

Coverage gaps to watch for

Definition of ‘professional service’. Some wordings define this narrowly; ensure AI-assisted advice falls inside the definition.
Delegated authority. Where AI is used as an implicit delegation, the insurer wants to see the professional retaining responsibility. Wordings vary.
Sub-limits on automated decision-making. Some newer wordings sub-limit or exclude AI-driven decisions. Read specifically.
Data breach vs PI overlap. AI-related confidentiality failures may sit inside a cyber policy or a PI policy; sometimes both, sometimes neither. Confirm at inception.

Professional-body expectations

SRA (solicitors) has published guidance on AI use — supervision, competence, confidentiality, transparency to client. Compliance with SRA guidance supports the fair-presentation position at PI renewal.
ARB (architects) Standard 8 — PII adequate to the practice. AI-assisted design work is professional-liability-relevant.
ICAEW (accountants) and other DPB bodies have issued guidance on AI in advisory and audit contexts.
FCA Consumer Duty applies to AI-assisted retail investment advice. AI does not override the four outcomes.
RICS (surveyors) has begun to address AI-assisted valuation.

Practical steps for firms using AI

Document your AI use policy — which tools, which work types, who supervises.
Maintain human-in-the-loop review for anything client-facing.
Keep an AI-tool inventory with vendor SOC 2 status where available.
Address AI use in your Consumer Duty framework (for FCA-authorised firms).
Disclose AI use at PI renewal in the specific form insurers request.
Track AI-related notifications or complaints separately for annual review.

Frequently asked

Do PI insurers currently exclude AI-assisted work?

Most standard UK PI wordings do not exclude AI-assisted work as a class. Some newer wordings introduce sub-limits or specific conditions. Read the wording; ask the broker to confirm.

Do I have to disclose AI use at renewal?

If the proposal form asks, yes — fair-presentation duty requires it. Where the form does not ask specifically but AI is a material feature of your practice, prudent disclosure is still recommended.

If AI hallucinates and my client acts on it, does PI cover me?

Under a standard PI wording, yes — the professional's civil liability for the advice is what triggers the policy, and it doesn't matter whether the advice was AI-drafted. But: if the firm delivered AI-generated advice without professional review, the insurer may argue supervision failure. Human-in-the-loop matters.

Does using AI reduce or increase my PI premium?

Generally neutral at the moment. Where AI is used with strong supervision, some insurers see it as a risk-reduction factor (fewer manual errors). Where AI is used without supervision, some insurers price up. The market is not settled.

Is AI-generated confidentiality breach a PI claim or a cyber claim?

Depends on the wording. Confidentiality breach linked to professional advice may fall under PI; breach of data-handling contracts or notification duties typically falls under cyber. Some firms carry both and address the overlap explicitly.

Do professional-body rules override PI wording on AI?

The professional-body rules (SRA, ARB, ICAEW etc) set the standard of care and the fitness-and-propriety requirements. PI responds to civil liability that flows from breach of that standard. Compliance with professional-body AI guidance supports the PI position.

What if my staff use AI tools I haven't authorised?

Shadow AI use is a growing risk area. From a PI perspective, the firm's professional responsibility for the output is unchanged. From a supervisory perspective, the firm needs a documented AI-use policy and monitoring. Discuss at PI renewal.

Are there specialist PI insurers for AI-heavy professional practices?

The market is still developing. As of mid-2026, no UK insurer publicly markets itself as an AI-PI specialist. Some Lloyd's syndicates have appetite for AI-adjacent risks. A specialist broker with wholesale access tests where the market sits for a given firm.

What might your PI premium look like?

A guideline range built from the premiums insurers have actually quoted on risks we handle. Pick your profession and enter a few details — it updates instantly.

Profession

Annual fee income — gross, last 12 months

Limit of indemnity

Any claims history?

Any higher-risk work? — e.g. large or overseas projects

Guideline range — this is not a quote

Choose your profession and enter your fee income to see a guideline range.

Get an actual quote →

How these figures are produced

This guide is built from Apex's own market data: the premiums insurers have actually quoted and charged on professional indemnity risks we have handled. Each night that data is aggregated into anonymised rate bands by profession, fee income and limit of indemnity. No client information is published — a band only appears where it contains at least five separate records, and unusually high premiums are excluded so a single atypical risk cannot distort the guide.

The range shown spans the typical spread of recent market outcomes for similar risks. Individual quotes can fall outside it in either direction. Figures exclude insurance premium tax at 12%.

This calculator is not a quote and is not an offer of insurance or advice. Your actual premium depends on full underwriting of your business, including your activities, claims record and insurer appetite at the time.