Skip to content
AI DISCOVERY12 min

AI Discovery: Winning ChatGPT, Perplexity & Gemini Recommendations

Assistants now recommend apps by name. If yours isn't in the answer, no amount of App Store ranking recovers the loss. This is the operating discipline that puts you back in the sentence.

// contents
  1. 01.Why assistants matter now
  2. 02.The tracked-prompt panel
  3. 03.The audit surface
  4. 04.Evidence engineering
  5. 05.Third-party citation strategy
  6. 06.Metadata rewritten for machines
  7. 07.The monthly ship cadence
  8. 08.How we defend the number
01

Why assistants matter now

People increasingly ask ChatGPT, Perplexity, Gemini or Copilot which app fits a specific need. If your app is absent from those answers, a competitor can occupy that consideration set.

02

The tracked-prompt panel

Build a fixed panel of prompts representing your buyer's language across capability, competitive, compliance, price and geography intents. Run the same panel monthly across named assistants. The result is Share of Recommendation: the proportion of prompts in which your app is named.

03

The audit surface

The first audit surfaces the gap: which assistants recommend you, on which prompts and beside which competitors. Competitors that appear repeatedly but are absent from existing dashboards become part of the roadmap.

04

Evidence engineering

Assistants cite evidence: reviews, listings, articles, category rankings, and public claims. We engineer evidence into the sources they read, the app description, review-theme distribution, third-party listicles, comparison pages, press mentions. This is not link-building; it's citation engineering.

05

Third-party citation strategy

Assistants weight sources: Wirecutter, The Verge, category-specialist blogs, aggregator pages, comparison sites. We map which sources each assistant favours in your category, then run editorial outreach targeting those sources with legitimate stories, not paid placements.

06

Metadata rewritten for machines

Descriptions become training data. Capability-first phrasing, entity-clean naming, feature language machines can lift verbatim. See our Apple AI Tags guide for the field-level detail; the same principles apply to how large language models parse your listing.

07

The monthly ship cadence

Panel run. Gap analysis. Two-week ship of description + review-theme adjustments + one editorial placement. Re-run panel next month. Compound.

08

How we defend the number

Share of Recommendation is a defensible metric because the panel is fixed, the sampling is public, and the assistants are named. Boards accept it when the methodology is transparent. We include the raw panel outputs in every monthly report, no black box.

// next step

Want this applied to your app?

Request a free app audit
// ready to ship this?

Get an audit that lands actions, not slides.

Request a free app audit