For app founders
How to get your app recommended by ChatGPT, Perplexity and Gemini
Every answer below comes from 50 dated audit runs we ran ourselves across three engines, not from a vendor blog. Where the honest answer is “you can't measure that yet”, it says so.
Vandro Labs is a Generative Engine Optimization agency for astrology and spiritual wellness apps — we work on getting apps named inside ChatGPT, Perplexity and Gemini answers. What follows is our own field notes, not a brochure. What we do · Free audit of your app
How do I get my app recommended by ChatGPT?
Get named on the pages the engine already cites for your category. Across 50 dated runs and 219 citations, the two largest source types were app store listings (26.9%) and comparison pages published by apps ranking themselves (20.1%). Genuine editorial publications accounted for 5.0%.
That distribution is the whole strategy, and it is not the one most founders expect. The instinct is to chase press. Press is one twentieth of the citations. The two surfaces that actually feed the answer are ones you control outright: your store listing and a comparison page on your own domain.
In practical order:
- Rewrite the store listing in the vocabulary of the question. Engines lift text straight out of listings. If users ask for “transit alerts” and your listing says “celestial notifications”, you are invisible to that question no matter how good the app is.
- Publish a comparison page you own. We repeatedly found pages that rank their own product first, with no disclosure, cited by all three engines — including one that was the single most-cited source for the highest-intent question we tested.
- Get named in the specific threads and directories that recur in your citations. Not “do Reddit”. The specific threads that already show up when you audit.
- Answer your reviews. We logged an engine citing a developer's reply to a review. Unanswered one-star reviews about your paywall are raw material for the paragraph the assistant writes about you.
Do ChatGPT recommendations actually drive installs?
They drive consideration, and the attribution is not clean. The user asks for a shortlist, decides inside the conversation, then opens the store and searches your name — so the install lands in your dashboard as organic store search.
This is a zero-click discovery layer sitting upstream of the app stores. It is genuinely hard to measure, and we would rather say that than sell you a number. What can be measured, dated and tracked over time is whether you are in the shortlist at all — and that is the metric we report.
The comparison that makes it worth doing anyway. Paid iOS installs cost $5.84 on average and that figure rose 19% year over year. A citation in an AI answer is high-intent, has no marginal cost per user, and does not get more expensive when your competitors raise their bids. You already see the first number on your Apple Search Ads invoice every month.
GEO vs ASO vs SEO — what's actually different?
| SEO | ASO | GEO | |
|---|---|---|---|
| Surface | A ranked list of links | Store search & browse | One synthesised answer |
| Slots available | 10 per page | Dozens | 2 to 5 apps. That's it. |
| Unit of success | Position | Rank & conversion | Being named in the answer |
| Where the user is | Deciding what to read | Already inside the store | Deciding before the store exists to them |
| Variance | Stable day to day | Stable day to day | Same question, different answer next run |
The two disciplines overlap in one place that matters: your ASO text is GEO surface. Store listings were the single most-cited source type in our runs. The engines are reading your listing whether or not you wrote it for them.
Why does my competitor show up and I don't?
Almost never because their product is better, and — this surprised us — not because they have more reviews. In our dataset, review count predicted neither citation nor revenue.
The recurring causes, in the order we see them:
- They own the comparison page for the category and you don't.
- They sit in a forum sidebar or a resource list the engine keeps citing. Sidebars and community wikis are cited far more than their traffic would suggest — that is a PR target most people never think of.
- Their listing matches the vocabulary of the question and yours doesn't.
- Sometimes it isn't about you at all. Visibility is violently engine-dependent. We recorded an app appearing in 6 of 17 Gemini runs and in zero ChatGPT and Perplexity runs, and another appearing in 4 of 17 ChatGPT runs and nowhere else. Before you conclude anything, check which engine you were asking.
A trap worth knowing about. When we asked an engine about a small named app, one of the three answered helpfully and then closed with a table titled “better alternatives”. Asking an assistant about your own app can actively route the user to your competitors. That is worth checking before you assume silence is your worst case.
How do I check whether the assistants mention my app?
Run your customers' questions in a clean session on each engine, on the same day, and repeat across days. A single screenshot is a screening signal, never a measurement.
Clean session means: temporary chat in ChatGPT; incognito in Perplexity with search-history personalisation switched off; incognito or a separate Google profile for Gemini. A logged-in assistant with memory on measures your own history, not the market — this is the single most common way founders convince themselves they are more visible than they are. The full method is published here, in enough detail to run it yourself this afternoon.
How long does it take?
| What you move | Realistic latency |
|---|---|
| An identified, useful comment in a thread that is already cited | Hours to days |
| Store listing rewrite | Days to weeks |
| A comparison page on a domain that already has signals | Weeks |
| Anything at all on a brand-new domain | Months |
Which is why, if you are choosing where to start: an app that already has its own website will see movement far sooner than one that only has a store listing. If you have no site, building one is step zero and it is a slow step — start it now and do the fast surfaces in parallel.
What does this cost?
GEO retainers in 2026 run roughly $2,000–$8,000 per month for mid-market work and $5,000–$12,000 for serious programmes. Below about $3,000 per month you are usually buying relabelled SEO. Our entry retainer is $3,000 per month.
We publish the number because a founder who has to email an agency to find out the price has already lost an afternoon. One founder-client place is open at a reduced rate until , in exchange for a published case study and a testimonial — one, not several.
When GEO is the wrong thing to buy
- Your retention is broken. Getting more people to a leaky bucket faster is not a win. Fix day-30 first.
- You have no website and no plans for one. We can still move your store listing, but the ceiling is low and you should pay accordingly.
- You want a guarantee. Answers vary between identical runs. In our own re-runs, an app cited on one day was gone from the same engine two days later. Anyone guaranteeing a position is selling you something else.
- You want it done quietly with fake independent review sites. Not our business. It also stops working the moment the models get better at spotting it.
Find out where you actually stand — free
Send the name of your app. We run it through the same question set on ChatGPT, Perplexity and Gemini, in clean sessions, and reply with the dated results — including the questions where you are absent and which apps are recommended instead. No call, no deck, no obligation.