An AI that tells you
what it will not do.
Most tools like this fail one of two ways: they bury you, or they get the data somewhere you would rather not explain. We built against both.
Where every answer comes from
- Company newsrooms, filings, investor material, earnings calls
- Employer careers pages and public job boards
- Funding databases, regulatory registers, trade press
- Public pricing and product pages, compared over time
Every single card names its sources. If you cannot trace it, we do not send it.
What we will never do
- Scrape LinkedIn or anything a platform forbids
- Buy personal contact data - enrichment stays your own step
- Guess things about a person they never made public
- Score anything in secret, or hide what we held back
You can open the withheld list any time and see exactly what you were not shown.
Ten kinds of document, every one of them checkable
Nine credible classes plus your own material, indexed together so one question reads all of it at once. Every one is a document you can open yourself.
Every single card names its sources. If you cannot trace it, we do not send it.
The withheld list is part of the product
Anything that did not clear the bar is still readable, with the reason it was dropped attached. Silence is never unexplained.
Collected
Your whole market, researched in full.
Deduplicated
The same event told twelve ways becomes one story.
Shortlisted
Scored on trigger strength, source weight and recency.
Sent
Cross-checked against a second source, then written up.
How the product itself is locked down
Who can sign in?
How is sign-in handled?
What is kept, and where?
Can we bring our own sources?
Read the sources yourself
Every claim in your first brief opens the document it came from.