Returns one page of the prompts an industry runs — the ids you pass to POST /org/industries/{id}/prompts/import to adopt them, and to POST /org/content-generation/industry-prompt-draft to draft from one. You can read any industry in the public catalogue (GET /org/industries) and your organization's own industry; any other id is a 404.

Only the industry's active prompts are listed. A prompt switched off at the industry collects no new results and is rejected by the import, so it is left out.

funnel_stage is the stage the prompt would have in your organization once imported, so the stage you pick by is the stage you get.

Each prompt also carries how it went over the date window. models has one entry per AI model that answered it, with your own brand overlaid: how many answers that model gave, and how many of them named you. top_three_mentioned is the three brands named most often on that prompt by anyone — so a prompt whose models show no mentions of you, next to a full top_three_mentioned, is one worth looking at: the AI answers it, and the brands it names are not yours.

The window bounds the FIGURES, never the list. Every active prompt is returned whether or not it has results in the window; one without them carries [] for both.

Figures are grouped into whole weeks, Monday to Sunday, UTC. A week is counted when its Monday falls inside from–to, so a from that lands mid-week starts the figures at the following Monday, and the last week is counted in full even where it runs past to. Pass Mondays for exact weekly slices. window echoes the dates you asked for.

Every prompt also says whether your organization already has it. org_prompt_id is your own prompt with this text — the one an import created, or one you wrote by hand — or null. Pass imported=false to list only the prompts you do not have yet, the natural call after importing one; imported=true lists the ones you do. A prompt counts as yours when a live prompt in your organization has the same text, ignoring case and surrounding whitespace. total follows the filter.

Prompts come back oldest first, in a stable order. limit and offset are validated rather than adjusted: a limit above 100 is a 400, not a silently smaller page. Requires the GEO product.

{
  "prompts": [
    {
      "text": "Which bank has the best mortgage rates?",
      "funnel_stage": "consideration",
      "models": [
        {
          "model": "chatgpt",
          "answered_count": 8,
          "brand_mention_count": 2,
          "brand_mention_rate": 0.25
        }
      ],
      "top_three_mentioned": [
        {
          "brand_name": "Acme Bank",
          "mentioned_count": 6
        }
      ]
    }
  ],
  "total": 212,
  "limit": 50
}