The short answer: when someone asks an AI assistant for an accountant — especially a chartered one, or one for audit, year-end or advisory work — the assistant has to decide which firms it can safely vouch for. Unlike "solicitor", the word "accountant" is not a protected title: anyone can use it. That makes the market noisier, and it makes verifiable ICAEW status one of the strongest signals an assistant has to separate genuinely regulated firms from everyone else. So before it names you, it checks whether what your website claims about you lines up with the authoritative record — chiefly the ICAEW register and Companies House. Where they agree, it is confident enough to recommend you. Where they conflict — a trading name that does not match your registered firm, an out-of-date address, "chartered" or "registered auditor" claims the record does not corroborate — it cannot confirm you are the firm the user is looking for, so it hedges or drops you. None of this is visible to you, and there is no notification when it happens.
This is one of the most common and most fixable reasons a capable accountancy firm never appears in AI recommendations. Below is how the check works, the four mismatches that cause exclusion, and what to do about each.
"Accountant" is not protected — which is exactly why AI leans on ICAEW
Here is the distinction that drives the whole thing. In the UK, "accountant" is an unregulated term — there is no law stopping anyone from using it. But "Chartered Accountant" is protected, restricted to members of bodies like ICAEW, ICAS and Chartered Accountants Ireland. Statutory audit, and certain other reserved work, requires registration with a recognised supervisory body, of which ICAEW is the largest.
For an AI assistant, that is the difference between a claim it can stand behind and one it cannot. When a business owner asks for "a chartered accountant in Bristol" or "an ICAEW firm for our year-end", naming a firm that is not actually chartered or registered is a factual error the model is trained to avoid. So it does not simply match the word "accountant" on your website — it looks for confirmation that your firm holds the status it claims. Your relationship to the ICAEW record is therefore a gating signal, applied before the quality of your website content is ever considered.
How AI actually uses the register — entity resolution, not a live lookup
It is worth being precise, because the popular shorthand is misleading. Most AI assistants do not query ICAEW’s database in real time on every request. What they do is build a confident picture of your firm — an entity — from everything they have seen about you across the public web: the ICAEW register, Companies House, your own website, accountancy directories, review sites. The more these sources agree, the more confident the model is that you exist, that you are chartered or registered, and that you do the work you claim. That confidence is what earns the recommendation.
For browsing-enabled answers — Perplexity, Google AI Overviews, and ChatGPT when it searches — register-derived sources can also be retrieved and cited directly. In both cases the principle is the same: consistency between your public information and the regulatory record is the trust signal. Inconsistency is the failure mode.
We call this regulated-entity resolution: the process by which an AI assistant decides whether the firm in front of it is the same regulated firm named in the official record. Firms that resolve cleanly get recommended. Firms that resolve ambiguously get hedged.
The chain looks like this — four authoritative sources feeding a single resolution check, with one binary outcome on the other side:
The four mismatches that get firms excluded
In our analysis of UK regulated firms, the same four discrepancies account for the large majority of avoidable exclusions. Each is a gap between what your website tells an AI assistant and what the ICAEW record tells it.
1. Trading name versus registered name
Your firm trades as "Summit Accountancy" but is registered at Companies House, and listed with ICAEW, as "S A Chartered Accountants Limited." To a human these are obviously the same firm. To an AI assistant building an entity, they can read as two weakly-connected things — neither of which it can confidently recommend. State the registered firm name and your ICAEW firm details plainly on your site, and connect the trading name to them, so the two resolve to one entity.
2. Address and contact inconsistency (NAP)
Name, address and phone number that differ between your website, your Google Business Profile, the ICAEW record and the directories you appear in is the single most common consistency failure. An old office address on one source and a current one on another forces the model to choose which to trust — and uncertainty pushes you down or out. Every public mention of your firm should carry the same address and phone number, matching the record.
3. "Chartered" and "registered auditor" claims the record does not support
This one is sharper for accountants than for any other sector, because the headline claims are themselves regulated. If your site says "chartered accountants" or "registered auditors" but your firm’s ICAEW and audit-register footprint does not corroborate that status, the assistant has no safe basis for repeating the claim — and will not stake a recommendation on it. The same applies to specialisms: "specialist R&D tax" or "ICAEW-accredited probate" needs to be reinforced across your public footprint, not asserted once on a services page.
4. Regulated status that is not machine-readable
Many firms display "ICAEW Chartered Accountants" as a logo in the footer, or "regulated by ICAEW" as an image with no firm number nearby. A person sees it instantly; a machine often cannot extract it. Your ICAEW firm details, in plain text and ideally in structured data, are what let an assistant connect your website to the register with confidence. Without them, you are asking the model to take your chartered status on trust — which, for a recommendation it has to stand behind, it will not do.
Why you cannot see this happening
The hardest part of this problem is that it is silent. When a business owner asks an AI assistant for an accountant and your firm is excluded on a consistency check, there is no impression logged, no bounce, no analytics event — nothing in any dashboard you own. The enquiry simply goes to the two or three firms the assistant could confirm, and you never know the conversation took place. A firm can rank well on Google, have a strong reputation and a full client book, and still be quietly filtered out of AI recommendations because its registered name and its trading name have never been connected in a way a machine can read.
How to pass the check
The work is unglamorous and entirely within your control:
- State your registered firm name and ICAEW firm details in plain text on your site, and connect your trading name to them.
- Make your name, address and phone number identical across your website, Google Business Profile, the ICAEW record and every directory you appear in.
- Reinforce your status and specialisms — chartered, registered auditor, the services and sectors you want recommendations for — clearly and consistently across your public footprint, not asserted once.
- Add structured data (schema) that encodes your firm as a regulated
AccountingServiceentity with its ICAEW details, so assistants can resolve you to the register without guessing.
Done properly, this turns four sources of doubt into four sources of confirmation — and moves you from the "cannot safely recommend" pile into the "named firm" pile.
The scale of the gap
For accountants the gap is different from most sectors — and that is the opportunity. TendorAI maintains a dataset of more than 63,000 UK regulated firms drawn from the official registers — ICAEW, SRA, FCA and Propertymark. Unlike solicitors, where many firms have no website at all, nearly all ICAEW-registered accountancy firms have a website. The problem is not absence; it is that most of those sites carry the consistency gaps above and almost none expose their regulated status as structured data an AI assistant can read. In other words: the firms that get recommended are usually not the best firms — they are the firms whose chartered status an assistant can actually confirm. Almost everyone is leaving that confirmation to chance.
That is a fixable position, and being early to fix it is the advantage.
Frequently asked questions
Does ChatGPT really check ICAEW registration?
Is accountant a protected title in the UK?
We are ICAEW-registered. Why are we still not recommended by AI?
What is the single most important fix to get recommended by AI?
How do I know if AI assistants currently recommend my firm?
TendorAI is the AI visibility platform for UK regulated professional services firms. We maintain verified profiles for more than 63,000 ICAEW, SRA, FCA and Propertymark-registered firms, install structured data on your own domain, and track how AI assistants describe and recommend you. Check your firm's AI visibility score.