Search ·

How brands get found in AI answers: ChatGPT, Google AI Overviews and Perplexity

What a brand can do to be cited in AI answers: let the right crawlers in, be one clear entity, keep facts consistent, write pages that answer questions, use honest structured data and earn real mentions and reviews.

  • Author: Workly Up editorial team
  • 7 min read

To be found in AI answers, a brand needs the same foundations as good search visibility, applied with more discipline: pages that the right crawlers can reach, one clear identity, the same facts everywhere, content that answers real questions in plain sentences, structured data that matches the page, and genuine mentions and reviews elsewhere. ChatGPT search, Google’s AI Overviews and AI Mode, and Perplexity all draw on the web to build their answers. This guide sets out what a brand can do about each part.

How AI answers choose their sources

AI assistants with web search retrieve pages from a search index, summarise them and usually link to the sources they used. The brand that gets mentioned is the one whose pages are reachable, easy to understand and consistent with what other credible sites say about it.

Google’s own guidance on AI features and your website is direct: there are no additional requirements or special optimisations needed to appear in AI Overviews or AI Mode. A page must be indexed and eligible to show with a snippet, and the usual fundamentals apply. Traffic from these features is included in the Performance report in Search Console.

Other engines run their own crawlers, so access matters as much as content.

1. Let the right crawlers reach your pages

Check your robots.txt and your site’s settings before anything else. A single line can keep a brand out of an engine’s answers.

  • Google. Googlebot must be able to crawl the pages you want cited, and they must not carry noindex or nosnippet directives that remove them from snippets.
  • ChatGPT. OpenAI’s crawler documentation says OAI-SearchBot is used to surface websites in ChatGPT’s search features and recommends allowing it. GPTBot, which collects content that may be used for training, is a separate setting.
  • Perplexity. Perplexity’s crawler documentation describes PerplexityBot as the crawler that indexes sites for its search results, and says it is not used to train foundation models.

Keep the key facts in HTML text, not only in images, PDFs or scripts that load after the page. Whether to allow training crawlers is a separate business decision; it does not need to be bundled with search access.

2. Be one clear entity

AI systems work with entities: a named organisation with attributes such as a location, a founding date, products and official profiles. If your brand appears under three spellings, two descriptions and an old address, the engine has to guess which is right.

  • One name, written one way, on the website, social profiles, marketplaces and directories.
  • One short description of what you do, for whom and where, reused everywhere a profile asks for it.
  • One About page that states the facts plainly: legal name, where the company is registered, what it offers, which markets it serves and how to contact it.
  • Organization structured data on the home page or About page. Google’s Organization markup documentation lists properties such as the legal name, logo, address and sameAs links to your official profiles on other sites.
  • A Business Profile that matches reality. If you have a physical location, Google’s guidelines for representing your business ask that the name matches what customers see on your signage, website and stationery.

A clear positioning makes this easier, because the description says the same thing everywhere; our guide to brand positioning covers how to settle it.

3. Publish facts that agree everywhere

AI answers often combine facts from several sources. When those sources disagree, the answer may be wrong or the brand may be left out.

Make a single fact sheet and use it to update every place your brand appears:

  • legal name and trading name;
  • what you offer, in the words customers use;
  • locations and the markets you serve;
  • opening hours, delivery areas and returns policy, where relevant;
  • materials, origin and certifications you actually hold;
  • official profile links.

Review it whenever something changes, and update the website first, then your profiles and listings.

4. Write pages that answer questions

AI answers are built from passages that state something clearly. Pages that answer the questions customers actually ask, in plain sentences, are easier to quote.

  • Answer first. Open each section with a one- or two-sentence answer, then explain.
  • Use the customer’s words in headings, such as “How long does delivery to the UK take?” rather than “Logistics”.
  • Define your terms. A short definition of what a product, treatment or service is helps engines and readers alike.
  • Use lists and tables for steps, specifications and comparisons.
  • Date and sign content. Show when a page was published or updated, and who stands behind it.
  • Cite sources for any figure or factual claim, and link to them.

Good candidates are product and service pages, sizing and care guides, delivery and returns pages, and a frequently asked questions section written from real enquiries.

5. Use structured data that matches the page

Structured data makes true facts unambiguous to machines. Use the types that describe what your pages really are: Organization, Product with offers and availability, LocalBusiness or Hotel for a physical place, and Article for guides.

Google is clear that no special schema is needed for its AI features, and that markup should match what visitors can see. Never mark up reviews, prices or answers that are not on the page.

6. Earn genuine mentions and reviews

Engines weigh what others say about you, not only what you say about yourself. Consistent, credible mentions on other sites help them confirm who you are and what you are known for.

  • Press and editorial coverage in publications your customers read.
  • Directories and associations relevant to your sector and markets, with the same name and description.
  • Partner and stockist pages that link to you and describe you accurately.
  • Genuine reviews on the platforms customers use.

Never buy, write or selectively hide reviews. In the UK, the Competition and Markets Authority’s fake reviews guidance explains that fake reviews and concealed incentivised reviews are banned practices under the Digital Markets, Competition and Consumers Act 2024, and that misleading cherry-picking or suppression of reviews is covered too.

How to check where you stand

There is no single report for AI visibility, so build a simple routine:

  1. List ten to twenty questions your customers ask before they buy, in each language you serve.
  2. Ask them each month in ChatGPT with search, Google, and Perplexity, and note whether your brand appears, what is said and which sources are cited.
  3. Correct the sources that are wrong or out of date, starting with your own site and profiles.
  4. Watch Search Console for impressions and clicks on the pages that answer those questions.

Answers vary between sessions and users, so look for patterns over months rather than reacting to one result.

What not to do

  • Writing separate pages only for AI. Write for people and organise clearly; that serves both.
  • Hiding text or stuffing keywords into FAQs or markup.
  • Marking up content that is not on the page.
  • Blocking crawlers by accident during a redesign or a security change.
  • Relying on a new file format alone. Proposals such as llms.txt may be read by some tools, but Google says no special files are needed, and none replaces the fundamentals above.

AI visibility checklist

  • Googlebot, OAI-SearchBot and PerplexityBot can reach the pages you want cited.
  • Key pages are indexed, eligible for snippets and carry their facts in HTML text.
  • The brand name and description are identical across site, profiles and listings.
  • The About page states legal name, registration, offer, markets and contact.
  • Organization markup includes legal name, logo and sameAs profile links.
  • A fact sheet exists and every listing matches it.
  • Main pages answer real questions in their first sentences, with dates and sources.
  • Structured data matches visible content.
  • Reviews are genuine; none are bought, written or hidden.
  • A monthly check of key questions is recorded.

Where to start

Start with the two checks that most often fail: crawler access in robots.txt, and whether your name and description match across your site and profiles. Then rewrite the opening of your five most important pages so each answers its main question in the first sentence.

If your website needs the structure, markup and content to support this, Workly Up builds brand websites with it from the start, and our care and growth retainers keep technical SEO and content in order month by month.

Questions

Questions clients often ask

If your question is not here, write to us.

Contact us

How do AI assistants decide which brands to mention?

Assistants with web search, such as ChatGPT search, Google's AI Overviews and AI Mode, and Perplexity, retrieve pages from their search indexes and summarise them with links to sources. A brand is more likely to be mentioned when its pages can be crawled, state clear and consistent facts, answer the question directly, and are supported by other credible sites that describe the brand in the same way.

Do we need special markup or files to appear in Google's AI Overviews?

No. Google's Search Central guidance says there are no additional requirements or special optimisations for AI Overviews or AI Mode. A page must be indexed and eligible to appear with a snippet in Google Search. The usual fundamentals apply: crawlable pages, important content in text, internal links, a good page experience and structured data that matches what visitors can see.

Should we block AI crawlers in robots.txt?

It depends on what you want. OpenAI and Perplexity separate their search crawlers from other uses: OpenAI's OAI-SearchBot surfaces sites in ChatGPT search, while GPTBot collects content that may be used for model training, and each setting is independent. A brand that wants to be cited usually allows the search crawlers and decides on training crawlers as a separate business question.

What is generative engine optimisation (GEO)?

Generative engine optimisation is the practice of making a brand's content easy for AI answer engines to find, understand and cite. In practice it builds on good search engine optimisation: crawlable pages, clear entity information, consistent facts, content that answers questions directly, structured data and a reputation built on genuine mentions and reviews. It does not require separate content written only for AI.

Does structured data help with AI answers?

Structured data helps search engines understand what a page is about and which organisation stands behind it, for example through Organization markup with your legal name, logo and official profiles. Google says no special schema is required for its AI features, and markup must reflect what is visible on the page. Treat it as a way to make true facts unambiguous, not as a shortcut.

Can we pay for reviews or write our own to look better in AI answers?

No. In the UK, fake reviews and concealed incentivised reviews are banned practices under the Digital Markets, Competition and Consumers Act 2024, and the CMA's guidance also covers cherry-picking or suppressing reviews in a misleading way. Ask genuine customers for honest reviews, make it easy to leave one, and reply calmly to criticism.

Insights

Book a call Start a project