Related Searches
See the searches AI runs before it answers you
Engines do not just answer, they search first. Recited AI captures every one of those related searches alongside the answer, so you can see the questions the model asked itself and target those instead of guessing.
Captured on every scan. No extra configuration.
Searches captured
4,912
best ecommerce analytics tools 2026
northlane pricing
everline vs northlane
shopify analytics app reviews
ecommerce attribution software
Every run
Fan-outs captured with the answer
Branded
vs shopping vs generic
By topic
Or by individual prompt
Named
Brands appearing in the queries
The hidden layer
The question you asked is not the question it searched
Ask an engine something broad and it quietly decomposes it into several narrower searches, retrieves for each, and merges the results. Those sub-queries decide which pages get read, which means they decide the answer. Almost nobody measures them.
- Captured during stage one of every scan, alongside the answer itself
- Grouped by topic or by the prompt that produced them
- Distinct queries, total searches and branded share at a glance
Searches captured
4,912
best ecommerce analytics tools 2026
northlane pricing
everline vs northlane
shopify analytics app reviews
ecommerce attribution software
Intent split
Branded, shopping and generic behave differently
A related search that includes your brand name means the engine already knows to consider you. One that is purely generic means the shortlist is still open. A shopping query means the user is close to buying. These need different responses.
- Branded share as a running measure of category awareness
- Shopping queries flagged separately, because intent is highest there
- Generic queries showing where the category is still up for grabs
GA4 vs dedicated ecommerce analytics
Cheapest analytics tool for small stores
Get listed on G2 ecommerce category
Integrations coverage
Brands and phrases
See which competitors get searched for by name
The queries themselves frequently name brands. When an engine searches your competitor by name while answering a neutral question, that is a measurable head start, and it is one you can see coming.
- Brands named inside those searches, counted and ranked
- Common phrases across your whole prompt set
- The vocabulary the model uses for your category, in its words
Retrievals
4,234
Also included
Why this ends up being the most quoted screen
It is the one that changes what people write, because it shows the target rather than implying it.
Real query language
The phrasing engines generate is not the phrasing in your keyword tool. This is the closest thing to seeing the retrieval layer directly.
Content targets
Each one is a page brief that has already been validated: the engine looked for exactly this and read whatever it found.
Outline skeletons
The content engine uses them as the section structure for a draft, so the page answers the sub-questions the engine actually asks.
Filter with the suite
Region, engine, brand, topic and date range apply here exactly as they do everywhere else in the product.
Per-prompt drill-in
Open any monitored prompt and see the related search terms that specific question produced.
API access
Pull the full query set through the project API and treat it as a keyword list that reflects how engines actually search.
Find out what the engines are really searching for
Run a scan and read the searches behind your own prompts. It is usually the report people screenshot first.