Best tool to track sponsored placements in ChatGPT answers: what we use and why

When ads started showing up inside AI answers, the first question we got from clients was not "how do we buy them?" It was "is a competitor buying ours?" That is a monitoring question, and the answer has to come from something that captures what a searcher actually sees.

This is how we approach it in client programs, and why we ended up on one platform for it.

Why sponsored placements need separate tracking

Organic AI visibility and paid placements get mixed up in screenshots and in some dashboards. A client sees a competitor next to their answer and assumes the model recommends them. Sometimes it does. Sometimes the competitor paid for the slot. If you cannot tell which, you cannot decide whether the fix is content, PR, or budget.

So the first rule in our reporting is to keep the two apart. Organic mentions and citations go in one view. Sponsored placements go in another, with the advertiser, the copy, the landing page, and the prompt that triggered it.

What to capture for every ad

We ask for the same six fields on every ad we log:

• advertiser name and domain
• ad title and description
• image or price shown
• target URL
• the tracked prompt that triggered it
• date seen

Without the triggering prompt, an ad is trivia. With it, you can see that a competitor only bids on comparison prompts, or only on prompts from one buyer persona, and plan around that.

The tool we run it on

We run client programs on Promptwatch, and the feature we use for this is Ads Radar. It captures sponsored ads shown inside ChatGPT Search, Google AI Overviews, and Google AI Mode answers to the prompts we track, and stores exactly the fields above.

The views we lean on are Ads per day, Top advertisers, the By Ad table, and the By Prompt table. The By Prompt table is the one we show clients, because it puts ad count next to your own average visibility on that prompt. A prompt with heavy advertiser activity and low client visibility goes to the top of the list. We filter by model, advertiser domain, prompt type, and date range when a client asks a narrow question.

Limits we tell clients up front:

• Ads Radar does not cover Perplexity.
• It does not report on a client's own ChatGPT campaigns. That reporting lives in OpenAI's own ad tools.
• It starts on the Business plan and is also on every agency plan.

That last point shapes how we scope work. A brand-side client on a lower tier would not have it, so for those engagements we either recommend the step up or we handle ad checks by hand and say so.

Why one platform instead of a separate ads tool

We have tried the split approach: one tool for ads, another for organic visibility, another for traffic. Reconciling them costs more time than the data is worth. With everything in one platform, the same prompt set drives all of it.

For a single prompt we can see:

1. Whether the brand is mentioned, and in what position.
2. Which pages and domains are cited.
3. Which advertisers appear in the sponsored slot.
4. Whether AI crawlers are reaching the pages we expect to be cited.
5. What AI-referred visitors do after landing.

That chain is what turns "a competitor is advertising" into an action. If the competitor is buying a prompt where we have no organic presence, the fix is content and citations, and the Action Board and content tools in the same platform handle that work. If we are already well cited and the ad is just noise, we note it and move on.

How we set it up in the first month

1. We build the prompt set from real buyer questions and label each by type: organic, brand, or competitor comparison.
2. We turn on ad capture for the commercial prompts first, since that is where bids are likeliest.
3. We let it run for two weeks before drawing conclusions. A single day tells you almost nothing.
4. We pull the top advertisers per prompt and read their actual copy for repeated claims.
5. We compare against organic visibility on the same prompts and write the plan from the gap.

What this will not tell you

Be honest with clients about the edges. Ads data is a sample of what appears for tracked prompts, not every ad everywhere. It will not tell you what an advertiser paid. It cannot guarantee that an ad will or will not show for a given prompt tomorrow. And watching competitors' ads is not the same as running your own campaign well.

What it does give you is evidence. A client who walks into a budget conversation with a By Prompt table showing which competitors sit on their best prompts has a much stronger case than one with a hunch.

The short version

If you manage AI visibility for clients and sponsored placements are on the agenda, pick the tool that logs the ad, the advertiser, and the triggering prompt, and keeps it next to organic data. That is the combination we have found useful, and it is why Ads Radar inside Promptwatch is what we use.