Do I need a big content programme to see results?

Not to start. The strongest lever in our own data — a real, active Trustpilot presence — is a review-generation exercise, not a content-volume exercise. A handful of specific, well-evidenced practice-area pages did more work in our sweep than a large volume of generic content would.

How do I know which engine matters most for my firm?

Run the same buyer questions your firm actually gets asked across ChatGPT, Gemini and Claude separately, and look at the recommendation rate on each — not just whether the firm is mentioned. Our own sweep found the same firm behaving completely differently engine to engine, so there's no substitute for checking directly.

Should I focus on reviews or directories first?

Reviews, based on our own sweep — third-party review volume outweighed directory presence for accountants specifically, which is the opposite order to what we found sweeping solicitors, where a professional directory led. Check both, but don't assume the order that worked in one professional-services sector transfers unchanged to another.

//.research

How to make an accountancy firm visible in AI search

Four things moved the needle in our own sweep, in this order: build a real Trustpilot presence, name a specific practice capability an engine can quote, check every engine separately, and don't expect your own website to do the deciding.

Steen Stones · Reviewed 5 Aug 2026

Build a real, current Trustpilot presence first. Everything else on this page matters less than that one line, because uk.trustpilot.com was the single most-cited domain in our sweep — ahead of every directory, every review platform, and most firms' own sites. If a firm has no active review base there, it is starting the rest of this work from behind.

1. Third-party review volume

uk.trustpilot.com collected 11 citations in our sweep, more than any other single domain. That's a genuine sector difference from what we found sweeping solicitors and R&D tax advisers, where a professional directory or a firm's own site led instead. For accountants specifically, real client reviews on an independent platform appear to carry more weight with the engines than anything else we measured.

11citations for uk.trustpilot.com in our sweep — the single most-cited domain, ahead of every directory

Avouch multi-engine market sweep, 5 August 2026 — UK accountants, 50 buyer questions across ChatGPT, Gemini and Claude

2. Content that names a specific capability, not a general reputation

Every recommendation quote in our sweep ties the pick to something checkable and specific: "specialises in contractors and freelancers, includes FreeAgent, uses qualified accountants and currently charges £119 + VAT monthly", "my recommendation for wealthy individuals, inheritance tax, property or international affairs", "specialises in charities and social-purpose organisations". None of them say "this is a well-regarded firm". Practice-area pages need to state, in plain text an engine can lift directly, exactly what the team does and for whom.

3. Check every engine separately, never a blended average

In our sweep, Crunch converted every single one of its Claude mentions into a recommendation and 50% of its ChatGPT mentions, while converting zero of its Gemini mentions — because Gemini converted nothing for anyone, across all 187 firms tallied. Average those three numbers and you get a figure that describes none of the three engines. Track recommendation rate per engine, not as one combined score, or real progress on ChatGPT and Claude will be invisible inside a number Gemini's flat zero is dragging down.

4. Don't expect your own website to do the deciding

A firm's own site still gets cited in this sector — crunch.co.uk was among the top three most-cited domains in our sweep. But across Avouch's wider 1,828-case research, a firm's own website was never once the sole source behind a recommendation. It gets a firm into the conversation; the third-party layer sitting alongside it — chiefly reviews, in this sector — decides what the engine does with that mention.

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

What not to spend the budget on first

A full website rebuild before the review base exists. A site with no clear content or no visible credentials is a real problem, but in our sweep the single most-cited source was a third-party review platform, not any firm's own domain. Fix the review base and the specific practice-area claims first; the website work compounds on top of that foundation rather than substituting for it.

//.questions

Common questions

Do I need a big content programme to see results?: Not to start. The strongest lever in our own data — a real, active Trustpilot presence — is a review-generation exercise, not a content-volume exercise. A handful of specific, well-evidenced practice-area pages did more work in our sweep than a large volume of generic content would.
How do I know which engine matters most for my firm?: Run the same buyer questions your firm actually gets asked across ChatGPT, Gemini and Claude separately, and look at the recommendation rate on each — not just whether the firm is mentioned. Our own sweep found the same firm behaving completely differently engine to engine, so there's no substitute for checking directly.
Should I focus on reviews or directories first?: Reviews, based on our own sweep — third-party review volume outweighed directory presence for accountants specifically, which is the opposite order to what we found sweeping solicitors, where a professional directory led. Check both, but don't assume the order that worked in one professional-services sector transfers unchanged to another.

AI can't recommend what it doesn't know.

See where you stand

Start free. No commitment.