The accountancy firm AI visibility checklist

This list is ordered, not alphabetical — the first section is the one our own sweep data says does the most work, and the last is the technical foundation everything else assumes is already in place. Every item traces back to a specific finding from this territory's own sweep or its regulatory research, not a generic checklist with "AI" added.

Third-party review presence — do this first

• A real, current, actively-managed Trustpilot presence. It was the single most-cited domain in our sweep, ahead of every directory and most firms' own sites.
• Review volume, not just a handful of legacy reviews — the engines in our sweep leaned on review count as a trust signal more than any other single factor.
• Reviews kept honest under the Digital Markets, Competition and Consumers Act 2024 — no fake or incentivised-without-disclosure reviews.

Content that names a specific capability

• Every service page states, in plain crawlable text, exactly what the team handles and for whom — not a general reputation statement.
• Named specialisms are backed by something checkable: a stated fee, a named client type, a specific practice area.
• No page relies on "trusted" or "experienced" language as its main claim — an engine can't verify sentiment, only specifics.

ICAEW/ACCA and statutory compliance

• ICAEW or ACCA membership details stated in plain text, not buried in a PDF or an image.
• PII position disclosed — clients are legally entitled to this under the Provision of Services Regulations 2009, regardless of AI search.
• Complaints procedure disclosed in plain text, per the same regulations.
• No unsubstantiated superlatives ("best accountant in [city]") anywhere on the site — ICAEW Code of Ethics Section 250 and the ASA's CAP Code both apply.
• Don't invent a price-transparency policy the profession doesn't require — state your actual pricing position honestly instead.
• AI-drafted marketing copy goes through the same named-partner sign-off as anything else before it publishes.

Measurement

• You can currently answer: how often is the firm named across the questions its buyers actually ask?
• You can currently answer: what share of those mentions become a specific recommendation?
• Both of the above are tracked per engine, never as one blended figure — Gemini converted zero of 184 mentions in our sweep, which a blended average would hide entirely.

Technical foundation

• robots.txt explicitly allows the AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended and the rest) rather than staying silent on them.
• Organization and Person schema present on the homepage and key profile pages, with real sameAs links out to Companies House and equivalent.
• Every named accountant has a visible byline and credentials, not an anonymous "our team" page.

“There's a difference between showing up and being recommended. AI cannot recommend what it doesn't know.”

None of this is exotic. It's the same discipline as any other client-facing marketing, applied to a new audience of readers that happens to be AI engines rather than people — which is exactly why the firms already doing it well in our sweep weren't running some secret AI strategy. They were being specific, current, and independently verifiable, and the engines noticed.