Building the estimating algorithm a decade before AI could do it

Every trade prices its work off a repeatable metric, room count, linear feet, square feet, but the standard industry process was to send drawings out to trades and wait weeks for bids to come back. I encoded those pricing metrics myself so the firm could produce a complete estimate in house instead of sending drawings out and waiting three weeks for someone else's number. The single largest source of delay in preconstruction was waiting on other people's bids. If you already know, from experience, roughly what a linear foot of a given trade costs, there is no reason to wait weeks to confirm it. The firm stopped waiting on trade bids for its baseline numbers, and that in house algorithm is the direct ancestor of Kanopi, the AI driven estimating engine I built more than a decade later. The engine is twelve years old. The AI part is what's new. The tool people think I invented in 2026 is really a twelve year old idea that finally got the technology to catch up to it. I didn't wait for software to solve a problem I already understood well enough to solve by hand first.