The quote queue that was quietly losing deals

What they thought the problem was

The owner was convinced the problem was capacity. Estimators were working late, the RFQ inbox never emptied, and the obvious fix was two more hires at roughly $180K/yr fully loaded. The ask when I walked in was not "should we build AI" — it was "help us onboard estimators faster."

What I actually found

I timestamped every RFQ from inbox to sent quote for the prior 90 days. The median quote took 4.2 days to leave the building, and 19% never left at all — buried under newer requests. On the deals that were quoted inside a day, win rate was more than double. This was not a headcount problem. It was a latency problem, and latency was costing them jobs they had already earned the shot at.

The opportunities, ranked by impact × effort

1. 01

   Quoting copilot

   Impact High · Effort Medium

   Drafts a first-pass quote the moment an RFQ lands, pulling from past jobs, current material costs, and the estimator’s own patterns. A human still owns the send.

2. 02

   Quote-age alerts

   Impact Medium · Effort Low

   Any RFQ older than 24 hours pings the owner and the estimator. Nothing goes stale silently.

3. 03

   Win/loss capture

   Impact Low · Effort Low

   One-click reason codes on every closed quote, so the model and the team learn what actually loses deals.

The numbers

What happened next

First build was live in 26 days. Within six weeks, 88% of RFQs got a quote out inside a day, up from 31%. They did not hire the two estimators. The retainer now covers material-cost sync and a second build in the scheduling queue.