The App Store is no longerthe only store.
A growing share of "which app should I use" is now settled inside an AI assistant: someone asks for the best app in your category, gets three names, and installs one. That answer is built from the open web, not from App Store rankings, so it has its own leaderboard and its own levers. MetaRun measures where you stand on it, who is beating you, and which pages the assistant actually read to decide.
Recommended in 1 of 3 answers. Ranked #1 on "best AI calorie tracker", absent from both category questions. MyFitnessPal named in all three. 10 sources read; you appear in none of them.
- 01
Track the questions your buyers actually ask
Not keywords: questions. "Best AI calorie tracker app for iPhone", "alternatives to MyFitnessPal", "is Nutrio any good". MetaRun proposes them from your store category, the ASO keywords you already track, and your tracked rivals, then keeps the set fixed so the trend line means something.
- 02
Ask a real answer engine, cold
Each question goes to a live, web-grounded assistant exactly as a person would type it, with no mention of your app. Anything else measures the prompt instead of the market. Every result is labelled with the engine that produced it, so nothing implies coverage of an assistant we did not call.
- 03
Read the answer the way a person would
Were you named at all, were you on the actual recommendation list or just mentioned in passing, and at what position. Those are different outcomes and they are reported separately, because a closing "other options include" is not slot two.
- 04
Follow the sources, because that is the lever
An assistant's recommendation is downstream of what it retrieves. Every check records the pages it read and how many of those answers left you out, which turns "be more visible to AI" into a specific list of pages to go and look at.
A second ranking system, with different winners
App Store search rewards downloads, ratings, and the keywords in your listing. An answer engine rewards none of those. It searches the open web, reads a handful of pages, and synthesizes a recommendation from what it finds: Reddit threads, best-of roundups, comparison posts, review sites. The two systems disagree constantly, and the disagreement is invisible unless you measure it.
In practice an app routinely owns its keyword in the store while being absent from the assistant's answer for the same intent, and the rivals named back are often not the ones that outrank it in search. Both facts are useful, and neither shows up in any store analytics dashboard.
It asks about your niche, not Apple's genre
Apple's categories are far too broad to be a fair question. "Health & Fitness" puts a habit tracker in a fight with every workout app ever shipped, and nobody types that phrase into a chat window anyway. MetaRun asks the narrow question first, taken from how your app actually describes itself, and keeps the broad category question as a second data point.
The difference is not cosmetic. The same app can score in the thirties on its store genre and in the sixties on the words its users would use, and only one of those numbers tells you anything you can act on. The question being asked is always shown, and always editable.
What MetaRun will not do
MetaRun never posts on your behalf. Not a Reddit comment, not a review, not a forum reply, not a direct message, not a listicle entry. That is a permanent product boundary rather than a feature waiting to be built.
The reasoning is straightforward. Every platform worth appearing on is actively demoting machine-written content, and a tool that manufactures it becomes a spam tool and takes its users down with it. The parts of distribution where automation genuinely compounds are the unglamorous ones: capturing demand that already exists, converting it when it arrives, and measuring which of it worked. That is the part MetaRun automates, all the way into the store.
- Posts or reviews we write for you
- 0
- Click from any app's command bridge
- 1
- Point visibility score, published formula
- 100
- Ways to run it: chat, voice, or your IDE
- 3
Frequently asked questions
What is answer engine optimization for apps?
It is optimizing to be recommended when someone asks an AI assistant which app to use, rather than optimizing to rank in App Store search. The inputs are completely different: assistants compose answers from web pages about your category, so presence and accuracy in those pages is the lever, not your keyword field.
How do I know if ChatGPT recommends my app?
Measure it. MetaRun asks a live answer engine the questions your buyers actually type and reports whether your app was named, whether it made the recommendation list, and what position it held. There is also a free checker that does this for any App Store app without an account.
Why do AI assistants recommend different apps than the App Store ranks first?
Because they rank on different evidence. Store search rewards downloads, ratings, and listing keywords. An assistant reads Reddit threads, roundups, and review sites, then synthesizes. An app can own its store keyword and be completely invisible in the assistant's answer for the same intent.
How do I get my app recommended by AI assistants?
Work on the retrieval rather than the wording. The answer is downstream of the pages the assistant reads, so the levers are being genuinely present and accurately described in the community threads and roundups that cover your category, and being clear on your own site about what the app does and who it is for. MetaRun shows you exactly which pages were used.
Which AI engines does MetaRun check?
Today it checks a live, Google Search grounded assistant surface, and every result is labelled with the engine that produced it. We report only on engines we genuinely call rather than implying blanket coverage of every assistant, and more surfaces are being added.
Does MetaRun write posts or reviews to improve my visibility?
No, and it never will. MetaRun does not post, publish, send messages, or write reviews and forum comments anywhere on your behalf. It measures, diagnoses, and points you at the pages that decide your category. What you do about them is yours.
Can MILO run this from my IDE?
Yes. Answer-engine tracking is part of the MetaRun tool registry, so MILO can run it in chat or by voice, and any MCP client such as Claude Code or Cursor can call it from your editor alongside the rest of your store operations.
Related capabilities
Capability file 05
ASO rank tracking
Track keyword ranks by country with an hourly sentinel and real Apple search data.
Open fileCapability file 06
Metadata editor
Bulk-edit every locale of your listing with preview, apply, and one-tap revert.
Open fileCapability file 07
MCP server
Drive your entire App Store from Claude Code or any MCP-capable IDE.
Open fileMILO standing by
14-day free trial. No credit card. Your keys stay encrypted.