How close did Go/No-Go come to an expert’s call?
He ran it on four tenders where he already knew what he would decide, and reported back on 25 August 2026.
4 of 4
Tested on four tenders. Every time, within a gnat's whisker of an expert's own call.
Our advisor's own test, on tenders where he already knew the answer.
“All four of those occasions, the Go-No-Go information that the app came back with was not exactly what I would have done, but within a gnat's whisker.”
“If I take it sensibly as a, this is a list of things for you to think about, it's brilliant.”
His caveat, which the site repeats: Go/No-Go only knows the documents you upload, and like any AI it can make mistakes, so check it.
How long did a first draft take?
About an hour, on 5 October 2026, for the technical questions of a tender.
£3,500 · 3–5 days
What it would cost, and how long it would take, to have a bid writer do the work TenderSeal now does.
Our advisor's own estimate, on a call in November 2026.
“To get to a first draft of the technical questions has only taken me maybe an hour … an hour, as opposed to days, which is why I'm grinning like a loon.”
“I can't begin to describe how pleasantly surprised I was with the result. It's spot on … it's done everything it was supposed to do.”
He still edits a draft so it sounds like him. It’s a first draft to work from, not a finished bid.
How does it compare with his own AI?
He had spent a long time training his own Claude.
“Yours is as good as the one I've been training for a long time.”
What did he change about TenderSeal?
He suggested calling the Go/No-Go score a compatibility score, not a win score, because it measures how well you fit the tender, not your odds of winning. That’s what this site calls it.