Prompt research reports
A prompt research report compares the questions people ask AI platforms with the prompts you track, and lists the ones you're missing.
A prompt research report puts two lists side by side: the questions people ask the AI platforms about your category, and the prompts you already track. What's on the first list and missing from the second is a gap in your tracking.
It's worth running because your numbers only cover what you track. If your prompts miss the questions people ask most, your visibility can look fine while customers are hearing about your competitors somewhere you never look.
What a report shows you
Most reports point to three kinds of problem:
- Common questions you don't track at all
- Prompts you track that hardly anyone asks
- Prompts that are near copies of each other
Each one skews your numbers in its own way. Missing questions leave you blind. Rare prompts take up room on your plan and can make you look better or worse than you are. Duplicates count the same question twice.
An example
Say Fernhill Coffee tracks 40 prompts, mostly about coffee subscriptions. A report shows that "office coffee" questions are asked often in its market, and Fernhill tracks none. It also shows five of Fernhill's prompts get almost no demand, and two ask the same thing.
Fernhill archives the rare five and one of the duplicates, and adds six office coffee prompts. Its prompt count barely changes, and its numbers now describe the questions customers actually ask.
In AnswerOrbit
The Coverage tab in Topics and prompts checks your set of prompts as a whole and lists the top fixes, like questions people ask that you don't track yet, prompts with the lowest demand, or duplicates. Click Review next to a fix to deal with it. The demand figures are estimates from a panel of conversations and searches on the six platforms, and they may lag by up to two weeks.