How Do We Appear
in ChatGPT? Our
Proven Process

On this page 11 sections

1. The Short Answer
2. Why Ranking on Google Is Not Enough
3. Where ChatGPT Actually Looks
4. What Decides Whether You Get Quoted
5. Case Study 1: From 117 AI Visitors to $29K
6. Case Study 2: The Brand ChatGPT Now Names First
7. If You Sell Products, There Is a Second Door
8. Our ChatGPT Visibility Process
9. What We Will Not Promise You
10. Where to Start if Your Budget Is Small
11. FAQs

Clients ask us this on nearly every call, usually with some frustration. They rank well on Google. So why does ChatGPT never say their name?

Fair question. Ranking on Google and getting named by ChatGPT are two separate jobs, and most brands only do the first one.

The second job goes by a few names. Answer engine optimization, or AEO. Generative engine optimization, or GEO. The goal is the same either way. You want a model to say your brand name inside its answer.

Here is what worked on two of our campaigns, with the numbers and the slow parts.

Why Ranking on Google Is Not Enough

Google gives back ten blue links and lets the searcher pick.

ChatGPT gives back one answer with two or three brand names in it. If you are not one of them, you are invisible. Position two does not exist here.

The audience is not small either. OpenAI said in February that ChatGPT had passed 900 million weekly active users, with more than 50 million consumer subscribers and 9 million paying business seats.

Source: Search Engine Land

Here is the part that surprises people. Being on page one of Google does not carry you across. When Josh Blyskal ran 600 queries through both, ChatGPT’s overlap with Google’s top 50 was 37 percent, and separately, in a comparison covering more than 600 queries, Claude and ChatGPT shared only 8 percent of citation domains.

That changes the job of your page. A page that ranks can be vague and salesy and still win the click. A page that gets quoted has to say something exact. The model wants a fact it can repeat without adding the word “maybe.”

One of our clients, a high-ticket jewelry store, ranked well on Google before we started. In January, they got 117 visitors from AI tools and made nothing from them.

Where ChatGPT Actually Looks

There are two sources, and only one is worth your budget.

The first is training data, a snapshot of the web from long before your campaign started. You cannot edit it, so leave it alone.

The second is live search. When ChatGPT searches, it pulls pages, reads them, and cites them. This is the part you can change this quarter.

More than 90% of ChatGPT users are on the free tier. Free means instant, and instant means labrador. (Search Engine Land)

The overlap with Bing is almost nothing: 1.5% of labrador URLs showed up in Bing’s top 20 for the same searches. Free Instant mode takes 74.7% of its results from labrador, just 3.1% from scraped Google, and 22.2% from oxylabs. Paid Thinking mode flips it and leans mostly on scraped Google rankings.

So “get indexed in Bing” is stale advice. Stay crawlable everywhere and let OpenAI’s own crawler build its picture of you.

The same teardown shows how brutal the funnel is. Across the sample, 61,332 URLs were surfaced into the sidebar, 5,032 became the lead source behind a citation, and only 759 pages were actually opened and read.

That last number is the one to care about. Pages ChatGPT opened were cited 74% of the time. Pages it merely surfaced and skipped were cited 7% of the time. Getting retrieved is easy. Being worth opening is the job.

What Decides Whether You Get Quoted

Four things, in the order we fix them.

Crawler access

OpenAI runs two bots, and its crawler documentation treats them as separate switches. GPTBot handles training. OAI-SearchBot handles search. You can allow OAI-SearchBot so you show up in answers while still blocking GPTBot from training use.

Block both and ChatGPT cannot read you at all. Same story if your main content only loads after JavaScript runs. We check this on day one of every technical SEO job, and it is broken far more often than people expect.

Position on the page

Kevin Indig analysed 1.2 million AI answers and 18,012 verified citations and found that 44.2% of citations came from the first 30% of a page. Only a quarter came from the final third. Bury your answer under 800 words of preamble, and it may as well not be there.

The same study found Q&A structure cited roughly twice as often as narrative prose, and that 78.4% of question-related citations came from headings. Write the heading as the question. Answer it in the next two sentences. We fold this into every on-page SEO pass now.

Mentions, with or without a link

If a review site names you next to three rivals, the model learns you belong in that group. No link needed. One caution, and Google says the same thing: manufactured mentions do not work. A paid-looking listing on a site nobody reads is noise. An editorial mention on a real publication is signal.

Language a model can lift

Indig found cited passages were nearly twice as likely to contain a clear definition, and that heavily cited text ran about 20.6% proper nouns against a normal 5 to 8%. Name the products, the platforms, the standards, the competitors. Vague copy gives a model nothing to hold. That is now the standard brief for our SEO content writing team.

Google backs the general shape of this in its guide to optimising for generative AI, where it says AEO and GEO are still just SEO, and that a page has to be indexed and eligible for a normal snippet before it can appear in an AI answer.

Case Study 1: From 117 AI Visitors to $29K

This is the jewelry store from above. We ran the work across one full year, in three parts.

First we found the sources. We asked ChatGPT, Perplexity, Gemini, and Copilot the exact questions their buyers ask, then traced each answer back to where it came from. Most led to Reddit threads, buying guides on big publications and review sites. Almost none led to the brand’s own product pages.

Then we rewrote pages a model could quote. Direct answers instead of hedged ones, comparisons with real numbers, buying guides good enough to be the reference. That is a different brief for a writer than the one most agencies hand out.

Last, we put the brand on the sites those tools read. Editorial posts on retail and lifestyle publications, listings on comparison and review sites. The goal was not link juice. It was that when a model builds its picture of the category, this brand turns up in several places, described the same way each time.

Growth was slow for two quarters, then it broke open.

Swipe to see the full table

Full year: 8,381 new users from AI tools, 31 purchases and $29,174 in revenue from a channel that made nothing in January. ChatGPT sent 6,309 of those users directly, and another 1,513 through referral links, so more than nine in ten came from one tool.

Engagement rate on that traffic was 73%, well above what the same store gets from Google. That makes sense. Someone who arrives from ChatGPT has already been told the brand is worth a look. They land half sold.

The full breakdown sits in our AI SEO case study.

Case Study 2: The Brand ChatGPT Now Names First

Our second client sells agency ad accounts to media buyers. Their buyers check everything before they spend, so many of them now open ChatGPT and ask for a shortlist.

We have worked on that site since launch, close to two years. Three things carried the result.

Data pages instead of keyword pages

We built pages full of numbers: Facebook ads statistics, Google ads statistics, and ad cost benchmarks. Each one holds figures a media buyer actually needs, laid out so a person or a model can lift a clean line out of it.

Over six months, the Facebook ads statistics page grew 675% in Search Console and the Google ads statistics page grew 3,850%. Those two now bring in most of the site’s AI traffic, and we reuse the format on every statistics page build.

One description everywhere

Directories, profiles, and other people’s write-ups all describe the brand identically. What it is, what it does, which platforms it covers. SEOs call this entity consistency. When those descriptions clash, a model hedges instead of naming you, and it is the most underrated part of AI search.

Mentions on the sites these models read

Much of this work happened nowhere near the client’s own domain. That is the job of digital PR, and most SEO plans skip it because it does not move a rank tracker.

Where it landed: ChatGPT now names this client first for its main category question, and so do Gemini, Perplexity and Claude.

AI sessions reached 1,639 for the period, up 111% on the six months before. ChatGPT sent 936 of them, Gemini 342, Perplexity 189 and Claude 173. Those sessions produced 91 contact conversions, up 98%. The numbers are all in our AI search optimization case study.

ChatGPT936

Gemini342

Perplexity189

Claude173

1,639 AI sessions (up 111%) and 91 contact conversions (up 98%).

If You Sell Products, There Is a Second Door

Everything above is about getting quoted in a written answer. If you run an ecommerce store, there is a separate route that most SEO advice ignores.

OpenAI now pulls product data straight from merchant feeds through its Agentic Commerce Protocol, so catalogues appear inside ChatGPT with images, prices and side-by-side comparisons. Target, Sephora, Nordstrom, Lowe’s, Best Buy, The Home Depot and Wayfair are all wired in directly.

Shopify merchants get a shortcut here. Product data already flows into ChatGPT through Shopify Catalog, with nothing to set up. What decides whether you show well is the quality of that feed: titles, attributes, stock status, pricing.

We treat feed hygiene as part of Shopify SEO now rather than a separate merchandising job. A brand can be quoted in the text answer and still lose the sale because its feed is thin.

Our ChatGPT Visibility Process

Every campaign we run follows these eight phases, in this order. The same work feeds Google AI Overview and AI Mode, so we never quote it as a separate project.

Swipe to see the full table

What We Will Not Promise You

Three things, said plainly.

• We cannot guarantee a spot

  Nobody can. Ask the same question two ways and the brand order shifts, because the answer gets written fresh each time. What we can do is build the signals that raise your odds, then show you what moved each month.

• Your reports will understate the traffic

  Many AI sessions arrive with no referrer, so they land in GA4 as direct. Treat the number you can measure as a floor. There is a second wrinkle worth knowing: a ChatGPT interface change in May 2026 sent a much larger share of referrals to homepages rather than deep pages, which makes year-on-year comparisons messy.

• Timelines depend on your category

  The jewelry store made its first AI sale in month two. The B2B client needed close to two years to lead every major tool, because its buyers are pickier and its rivals are sharper. If someone promises you 90 days in a hard category, ask what they are counting.

Where to Start if Your Budget Is Small

Order matters more than volume here.

1. 01

   Fix your own pages first. A model that finds you and then reads a vague product page will not quote you. Put the answer in the opening third, use question-shaped headings, name real entities, and keep the structure plain. This part is cheap, and everything after it depends on it.

2. 02

   Then get mentioned inside your own niche. For most brands that means real publications in your industry, not general blogs. We place those through niche guest posting. A DR 45 site that genuinely covers your industry teaches a model which category you belong to. A DR 78 general blog does not.

3. 03

   When you need speed, use pages that already rank. New guest posts take time to get indexed. Link insertions put your brand inside articles that already rank, which helps if you want to shift an answer before a launch or a busy season.

4. 04

   Then hold the pace. AI visibility rewards a slow drip far more than one big burst, which is why most of our clients keep a monthly rhythm with link-building packages instead of buying 40 posts in one quarter and stopping.

5. 05

   After that, say the same thing about your brand for a year. That is the dull part, and it is the part that decides it.

Want the full numbers behind these campaigns? They are in our SEO case studies. If you would rather talk it through, book a call and we will run a prompt audit on your category first.

FAQs