How do I get ChatGPT to recommend my consumer app?
People ask ChatGPT about a situation ("an app that turns my lecture notes into flashcards"), and it names the products whose pages describe that situation and that its crawler can read. Write one page per situation in your users' words, answer free or paid in plain text, and make the pages readable without JavaScript. On our own consumer web app, launched from zero on 21 September 2026, ChatGPT sent 333 visitors in the first 7 days, and 72 of the first 102 accounts came from it.
How consumers ask
A consumer rarely asks an assistant for a product category. They describe their moment: “I have an exam in two weeks and 40 pages of notes, how do I study faster?” or “is there an app that reminds my kid to take their medicine?” The assistant answers with advice and, more and more often, with a few named apps and links.
Two kinds of questions behave differently:
- “Top X” questions (“top note-taking apps”) tend to get the big brands. A new app rarely breaks into that list early.
- Situation questions (“an app that turns my lecture notes into flashcards”) get whatever page matches the situation and can be read. Here a small app can be named in days.
What happened on our own app
We launched an AI-powered consumer web app, our own product, on 21 September 2026 and built its public pages with this method. From the site’s own analytics, human sessions only:
| Measure | First 7 days |
|---|---|
| Visitors from ChatGPT | 333, 63% of human traffic |
| First ChatGPT visitor | about 1.5 days after launch |
| Accounts whose first visit came from ChatGPT | 72 of 102 |
| Paying customers whose first visit came from ChatGPT | 11 of 17 |
| ChatGPT visitors who landed on two situation pages | 59% |
It lasted: by day 26 (17 October 2026) ChatGPT had sent about 1,800 people and was the app’s biggest revenue channel, with 44% of sign-ups and 46% of monthly subscription revenue, and people from ChatGPT signed up at 31% against 9% from Google.
Source: the site’s own analytics, 21 to 27 September 2026 for the table (528 of 1,546 sessions counted as human), 21 September to 17 October 2026 for the 26-day figures (by the channel of each account’s first visit). One app, our own, not a client: an example of what the method did, not a forecast. The case study has the details, including the mistakes.
The pages a consumer app needs
- One page per situation. The H1 is the user’s sentence; the first two or three sentences say how your app solves it and what it costs. Find the sentences in support emails, app reviews, Reddit threads and the searches people already use.
- A pricing page in plain text. Free, trial, paid: say which, with the numbers in the HTML.
- An honest comparison or two. Users ask how you differ from the app they know. Name it fairly, with dated facts and the cases where it fits better.
- An about page with one sentence. What the app does, for whom, and who makes it. Keep that sentence identical on the homepage, the meta description, llms.txt and the store listings.
The technical traps
Consumer apps are often built as one JavaScript app, and that hides the public pages from crawlers. Test before you write anything:
- Run the free AI crawler checker on your homepage. It requests the page as the AI crawlers do and shows what text they get.
- If the text is missing, prerender or server-render the public pages. Do AI crawlers render JavaScript? explains what is measured and what is assumed.
- Keep the logged-in screens out of the index, return real 404s for missing pages, and list only public pages in the sitemap.
- Count ChatGPT visits properly: many arrive with no referrer. Tracking ChatGPT traffic in GA4 shows the channel rule.
Doing it yourself or having it done
If you built the app with an AI coding tool, Katman Pro ($99 a month) audits the live site, hands your tool each fix with an acceptance check, and briefs every situation page. If you would rather have the research and writing done for you, Growth writes 30 pages a month for one app site, with the fixes and a dated report every 30 days, for $1,490 a month.
Written by the Katman team, Rnify Limited.
Who this is not for
- Apps that live only in the App Store with no public website: start with the store listing.
- B2B tools sold to teams: the buyer questions differ; see the B2B SaaS shortlist page.
- Anyone who wants a promise of being recommended: nobody can promise an assistant will name you.
Frequently asked questions
Do people really find consumer apps through ChatGPT?
On our own consumer web app, yes: 333 real visitors came from ChatGPT in the first 7 days, 63% of human traffic, and 11 of the first 17 paying customers first arrived from ChatGPT. By day 26 ChatGPT brought 46% of monthly subscription revenue, more than any other channel. That is one app, our own product; read it as an example, not a forecast.
My app is a single-page React app. Can ChatGPT read it?
Often not. AI search crawlers generally read the HTML without running JavaScript, so a page whose text appears only after scripts load can look empty. Prerender or server-render the public pages and keep the logged-in app out of the index.
Should I write about my app or about the user's problem?
About the problem, in the user's words, with your app as the answer in the first sentences. On our sample app, 59% of ChatGPT visitors landed on two such situation pages.
Does the App Store listing matter?
It helps when your app is also on iOS or Android, because assistants read store listings too. Keep the name, one sentence and price the same on the listing and the website.
Is it free?
Users ask that constantly. Say it in plain text on the pricing page and in the FAQ: what is free, what is paid, and what it costs. An assistant that can't find the answer may say something wrong.
Can you do this for my app?
Yes. Growth writes 30 situation, pricing and comparison pages a month for one app site, with the technical fixes and a dated report every 30 days, for $1,490 a month. The one-time Launch Pack is $990 for 30 pages in 14 days.