Every contractor who's touched an AI tool has seen the same trick. You ask it for a price and it gives you one: confident, specific, and completely made up. In a chatbot, that's an embarrassing quirk. In an estimate, it's a lost job or a margin that never existed. A wrong number on a proposal isn't a typo. It's your name on a promise you can't keep.
Most software tries to fix this with a warning label, something like "AI can make mistakes, double-check important info," and calls it handled. We didn't think a disclaimer was good enough for a bid with your name on it. So instead of asking PQ nicely not to invent numbers, we built it so an invented number has nowhere to live.
That's the whole philosophy, in four words.
What it means on your screen
If PQ can't point at exactly where a price came from (a receipt you scanned, a price you told it yourself, a supplier's written quote, your own crew's logged hours on a real prior job) it doesn't show you a number. You get an honest blank instead.
Every price that does show up wears a badge. Tap it and the source opens right there: the store, the date, the record. Green means fresh and real. Yellow means it's aging. Blue means it's your own stated figure: trusted, but never dressed up as more than that.
Labor works the same way. PQ prices labor only from your own logged hours on a comparable job, never from "framing typically runs" guesses. When you do have a comparable job, it shows you the implied hourly rate so a bad comparison is obvious at a glance.
Quantities and measurements are a different matter. Those are yours to verify in the field, and PQ flags every one of them for exactly that. The rule is about the numbers PQ proposes, not the tape measure in your hand.
And when there isn't enough history to back a number up, the blank shows up honestly instead of guessing. That's not a missing feature. That's the rule doing its job.
The part we can't make up, even if we wanted to
The system grades the source, not us. That grading happens in a place we can't reach into and dress up, because a grade we could talk up would be worthless to you. Same with the report card: when PQ scores a proposal, the grade comes from your own records, and if there isn't enough history yet, it says so. No bullshit.
Why it gets better instead of stale
PQ only speaks from your records. Every receipt you scan, every walkthrough you confirm, every quote you import makes it sharper. That's because it's learning your business, not scraping the internet for regional averages that don't fit your crew or your market. That's why the honest blank you see on day one becomes a sourced, badged number by month three. It's not a workaround for a limitation. It's the whole point.
We're not going to tell you this makes your estimates error-proof. Nobody can promise that, and we're not going to insult you by pretending otherwise. What we can promise is narrower and truer: PQ can't hand you a number it can't back up. That's a much smaller claim, and it's the one that actually matters when you're the one signing the bid.
No source. No number. That's the K.O. Rule.