The things people ask first.
If yours isn't here, ask us directly — you'll get a real answer from a person.
How it works
For estimators and bid managers. It reads the tender pack itself.
What do you need from me to start?
One live tender. Drop the folder in and it’s read. That’s it. Later — once it has earned it — one finished job you wish you’d understood better.
What do I send you?
The tender as you got it: specs, drawings, BOQ, contract, addenda, scans. Drag the whole folder in — no renaming, no sorting. Anything we can’t take comes back named, with the reason.
What comes back?
A short list — three to seven things — that could change your price, scope, programme or submission. Each one quotes the page it came from, says why it matters, and what to do about it. The full list sits behind it, and anything it held back is listed with the reason.
How do I know it isn’t making things up?
Every quote it shows is checked word for word against your document. If the words aren’t on that page, the finding is thrown out before you see it. It can’t tell you about a clause that isn’t there — the quote wouldn’t match.
What if it finds nothing?
Then it says so. A clean tender is a clean tender, and padding the list would make it worth less. If it couldn’t read one of your documents, it names that document — “found nothing” and “couldn’t read it” never look the same.
Will it tell me whether to bid?
It gives you a straight answer in words, and says how much of the tender it could read. It won’t give you a win percentage — it refuses to make one up. A number like that looks official, would be fiction, and would make everything beside it harder to believe.
Does it write my RFIs?
It drafts them from each problem and its proof, along with qualifications and exceptions. It doesn’t send them. You do.
Is my tender data pooled with other contractors’?
No. Your tenders are never pooled with another builder’s or shown to one. Only two things can ever count across firms, and each needs your firm’s own opt-in: the totals on our website (rounded down, published only once five firms that opted in are counted, never naming a firm), and patterns across at least five firms that opted in, shown only as counts. Both are off unless your admin switches them on, and your admin can switch either off at any time.
What makes this different from the other AI tender tools?
Plenty of tools read tender documents now, and some do parts of it better than we do. The difference is what happens next: when an addendum changes something a decision rested on, that decision comes back to you with the reason it may no longer be safe; and what actually happened on the job is remembered for the next tender.
Your data, and what it costs
The answers we would want in your position.
Is my data safe?
Your tenders are virus-scanned before they’re read, stored encrypted, and walled off from every other firm — by the app, and again by the database underneath it. We never train AI on your data, and nothing from your tenders is pooled with another firm’s.
Have you been independently security tested?
Not yet. Every change is tested automatically, but that testing is ours — no independent penetration test has been done. One is scheduled. We’d rather tell you that than let you assume otherwise, and we’ll tell you when the report exists.
Does it do things on my behalf?
Only when you say so. It reads, finds and drafts. Nothing is sent, priced, submitted or changed until a named person approves that exact thing — and when it does act in your own systems, it keeps a record of what it did.
What does it cost?
Your first tender is free. After that you pay per tender, not per person: buy tenders when you need them, by card or invoice, and nothing renews by itself. We’ll give you the price on a call rather than a made-up one on a web page.
Have another question? henry@commonmind.co