Automating tender pricing
Case study 03
From tender documents to a priced estimate.
The challenge
A construction contractor priced every tender by hand. The know-how lived in the heads of a few senior people, and one tender could take from days to weeks. The hard part was getting that know-how out of their heads and into an AI agent.
The work
We started on site, watching how the team actually prices a job.
From the first version, the agent was built to be taught. The team shows it how they do things, and it keeps what it learns as memories and rules it applies to the next tender.
Tender documents go in; the method, quantities, a priced spreadsheet, supplier quote requests and the estimate come out.
The outcome
The team reports that pricing is much faster, and that the tool flags risks and opportunities they used to miss.
What it demonstrates
Knowledge that lives in a few people's heads can become a workflow the whole team uses, without taking the experts out of the loop.