Get the AI-insights report envelope for a single store: a live per-platform performance breakdown computed from the latest 30 non-suppressed, on-platform reviews within a 6-month window, plus stubbed narrative fields (summary, key_insights) reserved for a later AI-generation work-stream.

Auth: Both — authentication. No prefix/partner-only access; scoping is in-controller. The store_id query value must be in the caller's accessible store set or the endpoint returns 403. As a partner-tree user, the reachable set is every store under the partner's company hierarchy (getStoreListByCompany). The same endpoint serves both roles — only which store_ids are reachable differs.

Rate limit: No rate limit.

Request

• Query: store_id (integer, required) — target store PK; must resolve inside the caller's access scope.
• Query: platform (string, optional) — filters reviews to a single review site's name (e.g. "hipages", "Google").

Response — { data. } envelope.

• summary (string|null) — narrative summary; currently the stub string "AI insights pending generation." when reviews exist, otherwise null.
• key_insights (array) — AI narrative bullet objects; currently always [].
• category_performance (object) — keyed by platform name (reviewsite.name, or "Unknown"). Each value has count (int), avg_rating (float|null, rounded to 2dp over rates > 0), positive_count (int, rate ≥ 4), negative_count (int, rate 1–2).
• review_count_used (int) — number of reviews aggregated (0–30).
• lookback_days (int) — days from oldest used review to now; defaults to 180 when no reviews found.
• generated_at (string) — ISO-8601 timestamp of generation.

Errors: 422 on validation failure.