How to make your business visible in ChatGPT and AI search

Three things: ChatGPT has to be able to read your website, understand what you do and see that others mention you.

According to Ofcom’s Online Nation report (December 2025), ChatGPT is already the second-largest search service in the UK, with 252 million UK web visits in August 2025, up 156 per cent on a year earlier. If your customers ask their buying question there and you are missing from the answer, it is your competitor who gets the call.

ChatGPT search is now an established part of the search landscape. It follows the whole conversation with the user and has grown into a search category of its own alongside Google.

If you want customers to find you where they ask their buying questions, read more about how we approach AI search, or book a free search audit.

ChatGPT search is the search function in ChatGPT. It retrieves current sources from the web and summarises them in an answer with citations.

The technique behind it is called retrieval-augmented generation (RAG). When a question needs up-to-date information, ChatGPT sends its own queries to its search providers, Bing among them, fetches the pages that rank and splits the text into smaller chunks. The chunks that answer the question best make it into the answer; the rest are discarded.

What I like about the interface is that it works more like a sounding board.

Search 1: I ask for what I want, in this case five suggestions for black walking shoes.

Search 2: I add the requirements I felt the results were missing: a high ankle, as light as possible and the price shown in the answer.

Search 3: I set a budget.

With the traditional interface on Google and Bing, you have to run a new search or filter the results yourself. When I am trying to find something, I prefer the conversation.

Once the user does decide to click through, the traffic is highly relevant. The likelihood of a purchase is high.

Optimising for ChatGPT search comes down to two things: being findable in the search indexes it draws on, Bing’s included, and having passages that answer the question better than your competitors’ do.

As you would expect, there is a strong correlation between rankings on Bing and the sources that turn up most often in ChatGPT search. The relationship is not one-to-one, and ChatGPT search uses more sources.

Bing works a bit like optimising for Google used to: older SEO techniques, with more emphasis on matching the exact keyword. It does not, for example, lean as heavily on semantic SEO.

With a fraction of the searches, Bing also cannot draw the same conclusions about what users like.

Google has always changed, and Bing will too. Most likely the two will always differ, but Bing will move closer to Google if it gains more users.

Bear in mind that a top position on Google is no guarantee. Traditional domain authority is only loosely linked to which sources ChatGPT cites, and a large share of the cited pages are not even in Google’s top ten results. ChatGPT does not pick the domain with the most links. It picks the passage that answers best.

It is worth watching how behaviour changes. Most of the search volume you find in keyword research today is long tail on Google, but there are still a handful of big keywords that a business has to win.

OAI-SearchBot is OpenAI’s search crawler, the one that fetches pages for ChatGPT search. If it cannot get into your website, you will not appear in the answers, however good your content is.

Many businesses block everything from OpenAI to protect their content. There is no need. Why? OpenAI splits the work between separate crawlers:

• OAI-SearchBot fetches pages for the search function in ChatGPT.
• GPTBot collects text that may be used to train OpenAI’s future models.
• ChatGPT-User visits a page when a user’s action in ChatGPT calls for it, for example asking ChatGPT to read a specific URL.

So you can say no to training and yes to search. They are two separate decisions.

How to do it: Add separate rules for each crawler in robots.txt.

The next trap is the firewall. If you have a web application firewall (WAF) in front of your website, for example through Cloudflare, Akamai or AWS CloudFront, it can stop OAI-SearchBot even when robots.txt says yes. The crawler then gets an HTTP 403 Forbidden response, and the whole website drops out of ChatGPT search. OpenAI raises the problem itself in its guidance. The fix sits with IT: let OpenAI’s crawlers through the firewall, ideally based on their user agents, as OpenAI recommends. OpenAI also publishes its IP ranges in its developer documentation.

The last piece is Bing. Because Bing is one of the search providers ChatGPT draws on, Google Search Console is not enough. Like Google, Bing only spends resources on visiting your site when it thinks something has changed. With the IndexNow protocol, you notify Bing directly when a page is created, changed or removed.

This is usually the least effort of anything in this article. But if a single line in robots.txt or the firewall is wrong, none of the rest matters.

Entity authority is how clearly and consistently your company is described, both on your own website and elsewhere. If sources say different things about you, ChatGPT hesitates to mention you with any confidence.

To become a primary source in AI-generated answers, you need to build entity authority. That means more than naming entities (brands, people, places): it means showing how they relate to one another.

1. Define your core entities – Which brands, products, places and people are central to your business?
2. Create an entity map – Document how they relate to one another (e.g. “Sunbird is an SEO agency in Malmö that specialises in…”)
3. Use structured data – Add “Organization”, “LocalBusiness” and “Person” schema to make the relationships explicit, and use sameAs to connect your website to your LinkedIn page and other official profiles
4. Build internal links between entities – Link deliberately between pages that cover related entities
5. Earn external mentions – Work towards being cited by authoritative sources that LLMs are trained on

Why it matters: AI models value consensus and clear relationships. The more structured and coherent your information is, the better your chance of being cited.

ChatGPT trusts what others write about your company more than what you write about yourselves. For buying questions, ChatGPT draws a large share of its sources from comparison posts along the lines of “Best X for Y”. Reddit and Wikipedia are among the domains it cites most often.

Buying links is no shortcut here.

Picture two companies in the same sector. One calls itself the market leader on its own website. The other appears in two independent comparisons, has reviews and is mentioned in a Reddit thread. Which do you think ChatGPT recommends? My money would be on the second.

If you are not mentioned anywhere outside your own website, a competitor will overtake you. Put your most important buying questions to ChatGPT, note which pages it cites and work on getting featured on exactly those. On forums such as Reddit, automated posts will not get you far, but you can answer factually, correct errors and show that you know your subject.

Content that ChatGPT cites is content where every paragraph answers a question, stands on its own and keeps a factual tone.

ChatGPT does not read your page from top to bottom. It splits the HTML into smaller blocks (this is called chunking), compares each block with the question and includes the ones that answer best. A paragraph that starts “This also means that…” means nothing once it is lifted out of context.

Tone really does matter. Sales copy with words like “market-leading” and “end-to-end solution” is linked to a lower chance of being cited, and the models prefer text that reads like an encyclopaedia (awkward for an agency that likes to blow its own trumpet).

Use names. “Our solution” and “the platform” tell a model that only sees an isolated paragraph nothing, whereas “Sunbird”, “WordPress” and “Malmö” do.

Sourced figures make a difference. In a 2024 study from Princeton University and IIT Delhi, adding statistics, quotations or source citations to a text increased its visibility in the answers of generative search engines by 30–40 per cent. Keyword stuffing performed 10 per cent worse than the original text.

A common objection is “but we have the keyword in every heading”. Unfortunately, that is not how it works. The model is looking for the answer, not the keyword.

How to do it: Go through your most important pages with this checklist.

• The heading is the question the customer asks.
• The first sentence under the heading answers the question.
• The paragraph makes sense without the rest of the page.
• Names appear where it would otherwise say “we”, “the system” or “the solution”.
• Every figure has a source with a name and a year.
• Comparisons sit in a table, not in running text.

In a table, it is hard for a model to mix up which price belongs to which product. The same goes for your customers.

Optimising for follow-up questions means a page answers the whole chain of questions a customer asks in ChatGPT, not just the first one. The same applies to deep research, where ChatGPT runs its own searches in several steps before the answer is ready.

When someone searches for “best walking boots for the Lake District in autumn”, it is rarely the only question. The next one might be “which of these suit narrow feet?” or “how does Gore-Tex differ from eVent?”.

1. Identify question sequences – Which follow-up questions do your customers typically ask after the first one?
2. Build comparison tables – Make it easy to compare products or services across several dimensions
3. Create FAQ sections – Answer the most common follow-up questions directly on the page
4. Define technical terms – Explain industry-specific terms so the AI can use them in its answer
5. Use structured data for FAQs – Add “FAQPage” schema to complement the questions and answers in the text itself

• First question: “Best walking boots for the Lake District in autumn”
• Follow-up 1: “Which of these suit narrow feet?”
• Follow-up 2: “How does Gore-Tex differ from eVent?”
• Follow-up 3: “Which is best under £150?”

That is four questions, and all of them can be answered on the same page.

Why it matters: ChatGPT search is good at understanding context from earlier questions in the conversation. If your content is structured to answer the whole sequence, and not just the first question, you have answers ready when the follow-ups come.

If you would like help mapping your customers’ question sequences, get in touch and we will go through it together.

Since spring 2025, ChatGPT search has offered shopping features that show product recommendations directly in the answer. These include:

• Product images
• Reviews and user ratings
• Links to retailers

In practice this resembles Google Shopping, but the results are organic.

Just as with an ordinary answer, you can carry on the conversation if something catches your eye.

So a larger part of the buying journey can happen right in the answer. In practice this will mean less traffic, but a higher conversion rate.

AI-generated answers are often better at picking up nuance. You could, for example, say more about your problem and what you are looking for to get an even better result.

To appear in ChatGPT’s shopping results, you need to go further than basic product schema. These tend to make a difference:

1. “Product” – Basic product information
2. “aggregateRating” – The combined rating from all reviews (AI places a high value on consensus)
3. “review” – Individual reviews with detailed feedback
4. “offers” – Price, availability and delivery options
5. “brand” – A clear brand identity (an entity signal)

• Answer comparison questions – “This model differs from X in that…”
• Cover use cases – “Best suited to you if…”
• Include dimensions and specifications – AI uses these to filter
• Explain what the product is NOT suited to – This helps the AI rule it out for the wrong use case

ChatGPT search values consensus. If you sell on several platforms (your own website, Amazon, eBay), reviews on all of them count. In your own structured data, though, mark up only reviews collected on your own website: Google’s guidelines do not allow ratings aggregated from other websites.

• Long-tail keywords
• Technical SEO
• Structured data
• Appearing on trusted sources that LLMs are trained on
• Going deeper in your content and stating exactly what you do
• Entity optimisation

• Set up an account in Bing Webmaster Tools
• If you track keywords on Google, start tracking them on Bing too

My view is that SEO is increasingly a common thread running through everything you do on your website, rather than a channel of its own. If you are a Sunbird client, we cover the technical side of search as well as the content.

ChatGPT search has become a traffic source in its own right, particularly for complex informational searches and product comparisons. For many businesses, the traffic from it is smaller than from traditional search.

This mainly applies to informational searches where AI (an LLM) can give an answer. Overall traffic is likely to fall if users like the AI results enough. In the end, it is us as users who decide (especially if competition increases).

But if you need a tradesperson or want a pair of walking boots, it is a joiner or an online shop you are looking for. So even if there is less traffic, we do not think the total number of leads will fall. There will be winners and losers.

If LLMs do not bundle as many searches together as Google does today, the traffic may also end up shared between more sites than it is now.

Today we mainly measure rankings and traffic. But for informational searches, the main measure may no longer be traffic. It may be how often your domain appears as a visible source in ChatGPT’s answers (“visible mentions”).

You measure whether a business shows up in ChatGPT by asking the same customer questions at regular intervals and counting how often the business is named, linked as a source or recommended.

Start simple. Write down 20 questions your customers ask before they buy, such as “which accountancy firm in Leeds suits a growing limited company?”. Put them to ChatGPT once a month and note three things:

1. Are you named in the answer?
2. Is your website linked as a source?
3. Which competitors are named instead?

If you are named in 4 of 20 answers and a competitor in 11 of 20, you know where you stand. You also know which questions to tackle first. Bear in mind that the answers vary from one run to the next, so ask the same question more than once before drawing conclusions.

Once the list of questions grows, there are platforms that do the same job automatically. Ahrefs Brand Radar and Authoritas are two examples: they collect AI answers and show how often a brand is named compared with its competitors. Before you pay, check that the tool can track your market in the UK and follow ChatGPT specifically, not just Google’s AI Overviews.

Then there is the traffic. Visits from chatgpt.com usually show up as referrals in your analytics tool, and they are worth following all the way to a lead.

What counts is not the number of mentions but what they bring in. Ten visits from ChatGPT that turn into two leads are worth more than 200 visits that turn into none.

AI search gives businesses four clear challenges: measurement, fewer clicks, generic keywords and the need for up-to-date information. Each challenge has a matching opportunity.

The challenge: Traditional analytics tools do not capture “mentions”, that is, when your content is cited in an AI answer without the user clicking through.

The opportunity: Build a system for tracking visible mentions. You can do that by:

• Monitoring when your domain is cited in ChatGPT answers (manually, or via an API where one is available)
• Measuring branded searches as an indirect signal
• Tracking direct traffic and “unknown source”, where AI traffic without a clear referrer can end up

The challenge: Fewer clicks overall when the AI answers directly.

The opportunity: Build your landing page to convert, not just to inform. The user has already had the basic information and now wants to act.

The challenge: For broad questions, the AI pulls together information from several sources.

The opportunity: Focus on niche, specific use cases where you can be the only authoritative source. “Best shoes” is hard, but “best walking boots for narrow feet with bunions” is your chance.

The challenge: When a question needs current details, ChatGPT looks for fresh information, and a page with old statistics risks being passed over or producing a wrong answer.

The opportunity: Set up a content calendar for updates. Even small updates (new examples, fresh statistics, current references) can give you an edge.

If you would like help with this, get in touch. As an SEO agency, we help businesses build strategies that work for both traditional search and AI search.

Google has long been moving towards “zero-click” searches, where it gives the answer directly. The difference with LLMs is that the AI generates an answer tailored to what you searched for, whereas featured snippets took text from a source Google trusted and showed it on Google.

You measure visibility in ChatGPT by tracking a fixed list of customer questions over time and seeing what share of the answers name or link to your business. Add referral traffic from chatgpt.com, and a tool that tracks AI answers once there are too many questions to run by hand.

ChatGPT chooses its sources by sending searches to its search providers, Bing among them, fetching the pages that rank and then picking out the passages that best answer the question. Passages with a clear answer, factual language and sourced facts make it into the answer. When ChatGPT is recommending a supplier, what independent sources write about the company also carries a lot of weight.

SEO agencies that work on the technical side, content and mentions beyond your own website help businesses become visible in ChatGPT. Sunbird does exactly that, so we have an interest here. But whoever you choose, ask to see how they measure mentions in ChatGPT before you sign. An agency that only reports Google rankings is measuring the wrong thing.

Ahrefs Brand Radar and Authoritas are among the tools that track how often a brand appears in AI answers. For most businesses, a spreadsheet with a fixed set of questions goes a long way at first, and a paid tool only becomes worthwhile when you want to compare yourselves with many competitors.

You become visible in AI search by being reachable for the AI services’ crawlers, having content that can be lifted out as an answer and being mentioned by independent sources. The foundations are the same for ChatGPT, Perplexity and Gemini, but the indexes differ, and Google’s AI Overviews are built on Google’s own index. Taken together, this is usually called generative engine optimisation, or GEO.

Focus your SEO work on the areas where ChatGPT search is strongest: in-depth guides that answer the whole chain of follow-up questions, and detailed product data for shopping. Ranking on Bing is only part of it. Structure your content so that it can become a primary source in an AI-generated answer.

1. Let OAI-SearchBot in – Check robots.txt and the firewall
2. Review your Bing rankings – Identify big differences from Google
3. Add product schema – If you sell products, prioritise “aggregateRating” and “review”
4. Map question sequences – Which follow-up questions do your customers ask?
5. Build entity authority – Structure your content around clear entities and relationships
6. Set up measurement – Start tracking mentions, with branded searches as a proxy

We are hoping for real competition, with Google and Bing/OpenAI both competing for users and advertisers.

ChatGPT cites whoever answers most clearly and whoever others talk about. Mentions take time to build, but that is also their strength: no competitor can buy their way past them with ads.

Sunbird is a search marketing agency, and we are happy to help with more than your SEO.

If you want ChatGPT to name you ahead of your competitors, see how we approach AI search or book a free search audit with me (Nils). It takes 30 minutes and there is nothing to prepare.