Comparing bids by hand after the tender

The problem

When a tender result comes out, the firm wants to know where its price was too high. To find out, it builds a sheet comparing its own bid with the winning one and the closest one. This was done by hand, in a new spreadsheet every time, copying prices out of PDFs item by item.

What we did

Build

We set up an assistant that reads the price lists in the bids, whether they arrive as spreadsheets or PDFs, and produces the comparison in the format the team already used: green or red on each item and a summary by section.

The figures do not come from the AI. Code works them out and checks on every line that quantity times price gives the total. The AI only writes the conclusions from those figures.

To test it, we compared what the assistant produced with two old comparison sheets the team had made by hand.

The result

• In both test tenders, more than half the gap to the winning bid sat in a single section.

With the comparison ready as soon as the result is out, the team knows in which sections it is strong and where it is more expensive. Each analysis shows where to bid and where to rethink the price, so the team doesn’t waste time on tenders it can’t win.