When someone asks an AI search tool to recommend a local accountant, hotel, consultant or retailer, the answer depends on more than familiar SEO signals. The system must also be able to work out who the business is, what it does and whether the information it has found can be trusted.
A recent audit reported by Search Engine Land suggests that many websites make this surprisingly difficult. The study examined 71 businesses across several sectors and found substantial gaps in how clearly their identities could be retrieved and verified. The author described this as an “identity leak”: useful facts exist somewhere, but they are incomplete, hidden, contradictory or difficult for an automated system to read.
This does not prove that a particular website will be excluded from ChatGPT, Google’s AI features or another answer engine. Different systems gather and rank information differently. It does, however, highlight a sensible website-maintenance issue for businesses in Bath and across the South West: can an unfamiliar visitor — human or machine — quickly establish the basic facts about you?
What the audit found
The audit used a points-based framework covering company identity, technical foundations, external signals, content and whether key information could be retrieved. According to the published findings, the average business satisfied only 15.6% of the checks used to judge whether it was trustworthy and retrievable for AI.
That number should be treated as the result of one audit and its own methodology, rather than a universal industry benchmark. The practical problems it uncovered are more useful than the headline figure.
Three recurring issues stood out:
- important content appeared only after JavaScript ran, leaving little useful text in the original HTML;
- names, history, leadership and policies were tucked away on secondary pages with weak links from the homepage;
- a business identity was split across old domains, directory entries and social profiles, with no obvious canonical source.
These are not exotic AI problems. They are familiar website, content and local SEO problems seen through a new lens. A prospective customer can be confused by the same missing address, vague service description or abandoned domain.
Why this matters for local businesses
Local recommendations depend heavily on clarity. A Bath customer looking for a supplier may include a service, location, opening requirement and trust question in one conversational search. If a business website never states its service area plainly, names no people and disagrees with its Google Business Profile, it gives a search system little firm evidence to work with.
The risk is not simply “ranking lower in AI”. A third-party directory or booking platform may provide a cleaner, more complete version of the facts and become the source an answer engine cites. For a hotel, attraction or restaurant, that could push a customer towards an intermediary rather than the business’s own booking page. For a professional service, it could mean the company is omitted from a shortlist because its expertise and location are unclear.
Work on AI search optimisation therefore starts with accurate, accessible business information, not with attempts to game prompts or manufacture mentions.
A practical five-minute identity check
Start by opening your website in a private browser window and imagine you know nothing about the company. Can you answer these questions from the homepage and its main navigation?
- What is the exact business name?
- What does it sell or provide?
- Who is behind it?
- Where does it operate, and which areas does it serve?
- How can somebody contact it?
- Is there a current privacy policy and other appropriate legal information?
Next, compare those answers with your Google Business Profile, LinkedIn page and the most visible local directories. Names, addresses, telephone numbers, domain names and service descriptions should not contradict one another. Old websites should redirect properly rather than remaining as broken or competing versions of the business.
Finally, test whether the important wording is present without relying entirely on visual effects. A developer or SEO specialist can check the HTML response, but a simple warning sign is a page that becomes empty or meaningless when scripts fail. Core facts such as the company name, address, services and main explanatory copy should be available as readable page content.
Make the website the clearest source
An About page should name real people where appropriate and explain the company’s background without vague agency language. Service pages should say precisely what is offered and for whom. Contact pages should include current details. Those pages also need obvious internal links so neither visitors nor crawlers have to hunt for them.
Structured data can reinforce this information, but it cannot rescue an unclear website by itself. The visible content and the markup should agree. Our earlier guide to schema markup and AI search explains why entity clarity comes before adding more code.
It is also worth reviewing pages after a rebrand, office move, acquisition or change of domain. These are the moments when inconsistent identities multiply. Update the website, profiles and key citations together, redirect old URLs and decide which domain is the authoritative home.
What to check next
For most Bath and South West businesses, the useful response is a short factual audit rather than a wholesale website rebuild. Make the important content readable, put real people and contact details where visitors can find them, remove contradictions and maintain one clear source of truth.
These changes support conventional SEO in Bath, conversion and customer confidence as well as emerging AI discovery. Even if AI referrals remain a small part of your traffic, a website that explains the business clearly is rarely wasted work.
Source: Search Engine Land: AI search can’t verify your business — here’s how to fix it.
