Track your prompts in named groups, so you can see which buyer questions you own and which competitors take.
It’s usually one engine that dropped you, or one source it trusts more than your site.
Roman sorts every group by what’s actually wrong, so you know what to write next.
See whose page took the answer, and by how much weight.
Questions the engines answered without ever naming you.
AI mentions you, then cites a review site instead of your page.
Roman writes the article or refresh from your product, you approve it, and the group reports back.
Roman groups your prompts by topic during setup, and you can rename or regroup them anytime.
ChatGPT, Gemini, Perplexity, and Google AI Overviews are covered on every plan. Claude is available on the paid add-on.
Every monitor breaks its score down by engine, so you can see which one stopped recommending you rather than only that the average moved.
No. Monitors group the prompts you already track, and one prompt can sit in several monitors at once — a prompt can count towards both a comparison group and a competitor group.
Roman also clusters your generated prompts by topic during setup, so your first monitors arrive already built and named.
Add or remove prompts whenever you like, and the score follows the group. Archived prompts drop out of the numbers and are reported separately, so a score never shifts silently.
Deleting a monitor removes the grouping only. The prompts and their collected history stay.
Each monitor carries the number of answers behind it. A group running on a handful of answers reads as directional rather than exact, and Roman shows that count rather than hiding it.
Answer engines vary run to run, which is why every metric is reported against the previous window instead of as a single absolute figure.
No. Roman drafts the article or refresh from your product, positioning, and existing content, and you review and approve it before anything reaches your CMS.
Bring your prompts. We’ll group and score them, then pick the first articles to write.