Opportunities
A nightly scored queue of what to write or fix next, with a brief attached and somewhere to send it.
On this page
Turning it on
Opportunities are off until someone switches them on for the workspace — the tab opens with Enable opportunities. After that they're recomputed nightly at 04:00 UTC; the first run takes a couple of minutes.
The two feeds
Losing on collects tracked prompts where you're weak: a prompt qualifies when its mention rate is below half, or when any single engine is losing — a competitor doesn't have to be present for it to count.
Discover new proposes questions you aren't tracking yet, from two signals: Search Console queries with real impressions and a position inside the top 30 that have no prompt, and variations generated around prompts you already run.
How the score is built
Each row scores 0–100 from four named parts, and the modal shows them: visibility gap /40, reach /25, effort /20 (easier scores higher), competitive heat /15.
The brief
Every opportunity carries a brief: the angle, the H2s to answer, the entities to name, and the closest page you already have. You can swap that page for another, which regenerates the brief against the new page. Each row also carries a per-engine breakdown — winning, losing or untested on each LLM — with the evidence quotes behind it.
The eight recommendations
Create content · Optimize page · Rewrite for AI · Add schema · Internal links · Refresh content · Authority mentions · Track as prompt. The pill on each row says which one applies, and the filter narrows the queue to one kind of work.
Acting on one
The point of the queue is the hand-off. From an opportunity you can Send to Grid (mapping the brief onto that grid's columns), Add to Articles, or — in the Discover feed — Add as Prompt so it starts being tracked. Campaigns take this further, routing a kind of opportunity into a grid template automatically.
Statuses
Open · Watching · Acted · Resolved (auto) · Dismissed. The nightly run resolves rows whose mention rate has recovered, which is why "Resolved (auto)" exists. Run Scout re-scores prompts and can flag new rows; Refresh only recomputes the list.