Can Semrush or SpyFu show ChatGPT ads, or do you need a specialized tool?
Our view from running AI visibility programs: classic SEO suites are not built to capture sponsored answers in ChatGPT, and the ad question needs prompt-level data.
What clients ask us
When sponsored placements started appearing in ChatGPT answers, the first question we heard from clients was whether the SEO suite they already pay for would show them. Usually Semrush, sometimes SpyFu.
Our honest answer is that we do not rely on either for this, and we have no basis for telling you they cannot do it. We have not tested SpyFu's AI features, so we will not describe them. What we can tell you is how we approach the problem and why we ended up on a dedicated tool.
Start with the question, not the tool
Before picking software, decide which of two jobs you have.
Job one: see who is advertising against you. For a list of buyer prompts, which advertisers appear in the answer, what do they say, and where do they send people. This is competitive monitoring.
Job two: report on your own paid campaign. That reporting sits in OpenAI's ads tooling, not in any third-party tracker. We tell clients up front that nothing we use replaces it.
Most of the confusion comes from treating these as one. A suite that tracks keyword rankings and ad history on Google search is answering a different question from "who paid to appear in this generated answer."
Why prompt-level data matters
Ads in AI answers are triggered by the conversation, not by a keyword bid in the old sense. The same advertiser may show on one phrasing of a buyer question and not on another. So the useful record is the ad plus the exact prompt that triggered it.
That is the shape of the data in Promptwatch's Ads Radar, which is the feature we use. For tracked prompts it captures sponsored ads in ChatGPT Search, Google AI Overviews and Google AI Mode: the title, description, image or price, advertiser name and domain, target URL and the triggering prompt. The views we find useful are ads per day, top advertisers, and a table by prompt that shows ad count, your average visibility and the latest ad.
It does not cover Perplexity, and it will not tell you how your own OpenAI campaign performs. We say that to clients before they buy so nobody expects it.
How we use it in a client program
The workflow is less exciting than the feature name suggests.
- Build the prompt set first. Ads Radar only watches prompts you track, so the quality of the prompt list decides the quality of the ad data. We organize prompts by topic and intent, and we separate organic, brand and competitor-comparison prompts, since ads cluster differently on each.
- Let it run for a few weeks. One snapshot says little. Ads per day gives you a trend, and top advertisers shows who is consistently present.
- Read the By Prompt table next to visibility. The interesting rows are prompts where a competitor is advertising and your organic visibility is low. Those are the places where the answer is already being bought.
- Decide the response. Sometimes the right move is a paid placement. More often the cheaper fix is a better page, more citations from the sources the model already trusts, or fixing crawler access so the content is readable at all. The same platform holds the citation analytics and AI crawler logs, so the diagnosis does not need a second tool.
- Review the landing pages competitors use. Because the target URL is captured, we can compare their offer and claims against ours.
What it costs and who can get it
Plan gating is the part people miss. Ads Radar starts on Business, at $579 a month, and it is also on every agency plan, starting at Kick-off at $199 a month. It is not on Essential or Professional. If a client wants it and is on a smaller plan, we tell them that early, because it changes the budget conversation.
For agencies the math often works out differently. Because all agency plans include it along with unlimited projects and prompts, running ad monitoring across several clients can be cheaper than moving each one to Business.
What an SEO suite might still be good for
We still use general SEO tools for what they do well: keyword research, backlink data and classic rank tracking. If Semrush or SpyFu answers a question you have, keep using them. Our point is only that a sponsored answer in a chat interface is a new surface, and a tool should document that it captures it before you assume it does.
When you evaluate any vendor, ask three direct questions and get written answers:
- Does it capture sponsored ads shown inside ChatGPT answers, or only organic mentions?
- Can you see the ad text, advertiser domain and landing page, not just a count?
- Is the ad linked to the prompt that triggered it?
If the answer to the second or third is no, you are looking at a mention tracker with an ad checkbox.
Limits to keep in mind
No tracker can promise it saw every ad. Placement is decided by the platform and varies by user and time. What you get is a sample taken from the prompts you track. We treat it as directional evidence for decisions, not as a full census.
Our recommendation, then, is plain. Test the suite you have by asking it a direct question. If it cannot document ad capture, use Promptwatch for the ad monitoring and keep the suite for classic search work.