What an AI assistant actually looks at
When someone asks ChatGPT, Gemini or Perplexity a question like "who does web design in Ipswich", the assistant is not consulting a private list of approved businesses. It is pulling from a mix of indexed web pages, structured data on those pages, and third-party sources such as directories, review sites and forums. A business shows up when enough of that evidence points to it clearly and consistently.
Why Bing indexing matters more than most people expect
ChatGPT's search feature runs on Bing's index, not Google's. Many UK businesses set up Google Search Console and stop there, assuming that covers everything — it does not. Bing Webmaster Tools is a separate, free step, and skipping it is one of the most common reasons a business that ranks fine on Google is completely absent from ChatGPT answers.
The five things that actually help
1. Answer-first content
State the direct answer in the first sentence or two of a page, before any build-up. AI systems weigh opening content heavily when deciding what to quote.
2. Structured data (schema markup)
Markup such as FAQPage, LocalBusiness and Article schema gives machines an unambiguous version of the same information a human reader gets from the page.
3. A consistent business identity everywhere
The same name, address, phone number and one-line description across your website, Google Business Profile, and any directory listing. Inconsistency makes it harder for any system, human or AI, to confirm which business is being discussed.
4. Genuine third-party mentions
Directories, local press, review platforms and community discussion all feed AI answers. A brand-owned website is only one signal among several; independent mentions carry real weight.
5. Technical access for AI crawlers
Robots.txt files sometimes block crawlers by accident, including the ones AI systems use to read the web (GPTBot, PerplexityBot, ClaudeBot and similar). Checking that these are explicitly allowed is a five-minute technical fix with an outsized effect.
What about llms.txt?
llms.txt is a proposed file format — a plain summary of a site for AI systems to read. It is inexpensive to add and cannot do harm, but no major AI provider has confirmed using it in production, and adoption across the web remains low. Treat it as tidy housekeeping, not a guaranteed lever.
A realistic timeline
Technical fixes and indexing can happen within days. Consistent AI mentions tend to build over months, as more of the underlying signals — content, structured data, third-party mentions — accumulate. Anyone promising an overnight change is describing luck, not method.