Question mining
ICP query patterns from live search and Search Console.
Know the questions buyers actually ask.
One run returns a cited report in five parts: product map, competitor profiles, search topics, market summary and the questions buyers ask. The other agents draft from it.
Free analysis · 5 credits · no card · first score in about 30 seconds
5 parts · a source on every claim · current for 7 days
01 / The problem it removes
Positioning guesses age badly.
The questions buyers ask change every month. Last quarter's research spreadsheet is already wrong, and nobody can trace where its claims came from.
Without it
- Content gets written from hunches, not from questions buyers ask.
- Competitor gaps go unnoticed until they rank.
- Research lives in someone's head and leaves with them.
With the Research agent
- Buyer questions arrive with the source behind each one.
- Every report names the gap and the competitor currently winning it.
- Articles and posts draft from the report, not from a blank page.
02 / How it runs
A loop, not a task list.
- 01
Gather
Exa search results, your Search Console queries and your named competitors are read.
- 02
Rank
Topics and questions are scored by demand and fit to your product.
- 03
Report
A five-part report lands with a source envelope on every reference.
- 04
Feed
Articles, posts and ads draft from the report. Nothing starts from a hunch.
03 / What it can do
What the agent does
Competitor profiles
The companies you name, read on every run.
Topic map
SERP topics your buyers search, ranked.
Source envelopes
Every reference kept with a SHA-256 digest.
Agent handoff
Articles, posts and ads draft from the report.
Freshness window
Current for 7 days, then marked stale.
04 / In the product
A report lands. You read it in minutes.
Every report arrives with a source link on each claim. Read it in the queue, then hand it to the Articles, X or LinkedIn agent as the brief they draft from.
05 / Questions
Before you install it.
Five artifacts in one cited report: a product and service map, competitor profiles, a SERP topic map, a market summary and ICP query patterns. Each carries its schema version, UTC observation time, a source envelope per reference and a SHA-256 digest.
Live retrieval through Exa search, your Search Console queries and the competitors you configured. Every claim the report keeps carries its source link. A query with no traceable evidence is dropped rather than kept.
A completed run stays current for seven days, then is marked stale: readable, but prompt regeneration refuses it. On-demand runs are limited to one start per workspace per hour.
Research is on Lite and Pro. A run reserves at most 6.90 credits, authorizes each provider call against that hold and settles at the recorded cost. Lite is $129 a month, or $1,290 a year, with 2,000 credits each month.
Ships on the Lite plan
Start with the free analysis.
Free analysis, 5 credits that never expire, no credit card. The free analysis previews the Research agent; it briefs itself on your site once you are on Lite, $129 a month.