product guide

Insights and Recommendations

How to turn automated observations into decisions without treating every alert as equally important.

Updated 2026-05-29

The difference

Insights explain what the data is noticing. Recommendations suggest what to do with that signal. Use insights to understand the situation and recommendations to choose an action.

How to triage recommendations

Read recommendations in this order:

  1. Does it point to a repeatable format, topic, CTA, or packaging pattern?
  2. Is the supporting metric meaningful for the video type?
  3. Can you act on it in the next planning cycle?
  4. Does it conflict with a stronger business goal?

Recommendation states

Use saved or active recommendations as a planning queue. Dismiss recommendations that are already resolved, based on messy data, or not aligned with the channel's current direction.

Use recommendations as AI context

When a recommendation is specific enough, it can become the context for a growth tool. For example, a recommendation about a strong format can become a batch of new ideas, an outline, or a strategy bet.