Where they started
A commercial builder in south east Queensland doing fitouts and light industrial work between two and fifteen million dollars. Two estimators, one senior and one junior, and a pipeline of tender invitations that had grown faster than the team.
A tender took about nine working days from invitation to submission. Roughly half of that was measuring quantities from drawings and chasing supplier and subcontractor pricing. Another chunk was formatting the same scope into whichever template the client wanted. They were declining about a third of invitations because they had no time to get to them.
What we found in discovery
Three things stood out after a fortnight sitting with the estimators.
First, the business had eleven years of priced jobs sitting in their estimating software and an accounting system that nobody queried. Every tender started from scratch as if that history did not exist.
Second, the quantities work was almost entirely repetitive. The same building elements, measured the same way, from drawings that were more consistent than the estimators gave them credit for.
Third, the senior estimator’s real value was in pricing risk. Which subcontractor will turn up, and where the client will push back. Almost none of the nine days was spent on that.
What we built
A first pass workflow that reads the tender drawings and scope and drafts quantities against the elements the estimators already use. It prices them from the business’s own historical rates and current supplier lists, and produces a draft in the client’s template.
Three decisions mattered more than the model.
- The rates came from their own history rather than a database. We loaded eleven years of priced jobs and let the estimators see exactly which past job each rate was pulled from.
- Each line showed its confidence. Every quantity and rate carries a confidence flag. Low confidence items are listed first so the estimator’s attention goes where it is needed.
- The estimator signs off everything. No tender leaves the building without the senior estimator reviewing the draft. The workflow changed what they review.
Where it landed
Turnaround went from nine days to three. The team went from pricing roughly five tenders a month to twelve with the same two people, and stopped declining invitations. Win rate held, which was the thing the managing director was most worried about.
What we would do differently
We started with the full range of project types. The right call was to start with fitouts only, which were the most consistent, and expand once the estimators trusted the drafts. The first four weeks had more low confidence flags than they needed to, and it slowed adoption.