Find Which of Your Prompts Trigger ChatGPT Ads

The first question in ChatGPT ad tracking is not who is advertising. It is which of your prompts even produce a sponsored unit. Most prompts do not. ChatGPT does not show a paid slot on every query, and the prompts that do return one are a small subset of the ones you track. If you start from the advertiser view you read a list of rivals with no idea where they showed up. If you start from the prompt view you find the queries where the auction is real, and that is the set worth working. This is how to find those prompts with Promptwatch Ads Radar, and the point of finding them is to turn them into tickets. Review: Promptwatch. Full directory.

This site ranks tools that act. A list of active prompts is a report. A list of active prompts that drives which rows get escalated and which get dropped is a workspace. The steps below lean toward the second. The data behind this is observation, not a feed. There is no official ChatGPT ads API and no vendor has an OpenAI partnership for this. Ads Radar monitors the real ChatGPT interface and stores the sponsored units it sees, tagged to the prompt that produced them. Any tool claiming an official integration is misrepresenting how it works, and getmint is one example of that false claim pattern.

Start with the prompt view

Open Ads Radar and list the prompts whose answers contained sponsored ads. Each row has an ad count and a latest capture time per prompt. This is the list you want. A prompt on this list returned at least one sponsored unit in the window you are looking at. A prompt not on this list did not, at least not in a captured check.

Sort by ad count descending. The prompts at the top are the ones where competitors are buying repeatedly, which is a different signal from a one off. A prompt with one captured ad may be a test by a rival. A prompt with twelve captured ads is an active auction, and that is where the surface is real for your category.

Use the latest capture time

The latest capture time field tells you when a prompt last returned a sponsored unit, and it is more useful than the ad count for one job. It tells you whether the surface is still active. A prompt with a high ad count but a latest capture time three months ago is a dead auction. A prompt with a low ad count but a latest capture time yesterday is a live one. Sort by latest capture time descending when you want to know what is happening now, and sort by ad count when you want to know what happened overall.

The two sorts answer different questions and you need both. Ad count tells you where rivals have spent. Latest capture tells you where the surface still loads. A prompt that ranks high on both is the one to watch first.

Filter by intent and prompt type

The prompt view is more useful when you filter it. Ads Radar lets you filter by intent, which is BRANDED, INFORMATIONAL, NAVIGATIONAL, COMMERCIAL, or TRANSACTIONAL, and by prompt type, which is ORGANIC, BRAND_SPECIFIC, or COMPETITOR_COMPARISON. Filter to COMMERCIAL and TRANSACTIONAL first. Those are the prompts where a sponsored unit is most likely, because they are the prompts where a buyer is close to a purchase, and that is where the ad inventory concentrates.

Then filter to BRAND_SPECIFIC. A prompt where someone types your name and gets a rival's sponsored unit is the row that matters most, because that is a rival buying your own demand. A prompt where someone types a generic category term and gets a rival's unit is a different, lower priority signal. The intent and type filters are what separate those two, and a single unfiltered list blurs them together.

Confirm by hand before you act

Before you build a plan around a prompt the dashboard says is active, run it in ChatGPT yourself. This calibrates the dashboard against ground truth and catches two things. It confirms the surface still loads on that prompt, because captures can lag. And it shows you where the sponsored unit sits in the answer, which is the position field on the row but is easier to trust when you have seen it.

If the prompt returns no sponsored unit by hand but the dashboard says it did, the capture may be stale or the surface may have rotated off. Either way, you act on the live check, not the stale row. If the prompt returns a sponsored unit by hand and the dashboard has not caught it yet, give the capture a cycle to land before you assume the dashboard missed it.

Turn the answer into work

Once you know which prompts trigger ads, the work splits in two. For the prompts that trigger ads on your branded terms, escalate. A rival buying your name is a ticket. For the prompts that trigger ads on category terms you also target, compare. A rival buying a category term you also want is context for your own bid thinking, even though the auctions are not the same. For the prompts that do not trigger ads at all, stop spending time on them. Ad tracking on a prompt that never returns a paid unit is a metric about a surface that does not exist for that query.

This is not bid management. ChatGPT sponsored units are not Google Ads, and importing a Google mental model breaks here. Ads Radar is a monitoring tool for the ChatGPT surface. Treat the output as observation with capture gaps, because no feed exists.

Ads Radar is a platform feature on Professional at $245/mo and above, and on the self serve agency plans (Kick-off $199, Growth $399, Scale $799). Essential at $95/mo does not include it. Explore is free with 10 ChatGPT prompts. AthenaHQ at $295/mo and Otterly.AI at $29 do not store the sponsored unit against the prompt.

What to do this week

1. Open the Ads Radar prompt view and sort by ad count descending.
2. Re-sort by latest capture time to find the prompts that are still live.
3. Filter to COMMERCIAL and TRANSACTIONAL, then to BRAND_SPECIFIC.
4. Run the top branded prompt by hand in ChatGPT to confirm the surface loads.
5. Escalate the branded prompt rows as tickets and stop spending time on prompts with no ads.