Skip to content
Competitor Watch

AI visibility and estimated demand

Measure public buyer-answer mentions separately from social views.

Coming soon

  • Approved Copilot/Bing answer API, automated demand capture and full contextual verification on new engines.

Formula and example

Share of answer = complete sampled runs naming a brand / complete sampled runs. Average position = mean named order. Share of citations = brand-matched citations / all citations. AI-cited videos = distinct tracked video URLs cited.

Example (fictional): Kite Beauty appears in 2 of 3 completed answers: 67%, with low confidence. This is not 67% of consumer prompts.

Steps

  1. Open AI visibility and save 15–30 public questions grouped by intent and locale.
  2. Choose configured engines; missing credentials show Needs key. Set a weekly cap before starting.
  3. Every engine samples each question three times. Read matched previous-panel trends and repeat agreement.
  4. Open cited sources and Brief this to create an answer-first video draft.
  5. Question demand is estimated from search proxies, never actual LLM prompt volume.

What good looks like

The API, fixed model, period, complete repeats and sources are visible.

Common mistakes

  • Treating API answers as consumer-app rankings.
  • Calling a citation a view or sale.
  • Treating moved together as causation.
  • Assuming no registered contradiction means all product claims are correct.

Related articles