Most software pilots end with a demo and a vague sense that it went well. A good proof of concept ends with a number. Here is how we structure ours.
Week 1: scope
Choose the right contracts. Pick three to five contracts that are high in value, high in volume and complex in their terms. Outsourcing, logistics, catering, facilities and technology contracts are usually ideal.
Find the data. For each contract, list where invoices, purchase orders, delivery records and service data arrive. Email inboxes, ERP exports and file drops are all fine.
Agree success criteria. Write down what would make this worthwhile, in money. For example: “Identify at least £250,000 of recoverable value across the selected contracts.”
Weeks 2 and 3: configure
Contracts are ingested and the measurable obligations extracted: pricing, KPIs, penalties and remedies. Your team verifies anything with low confidence. System references are mapped.
In parallel, we run data-source workshops. Rather than assume where data lives, we ask the people who know. This is often the longest task, and the most valuable.
Weeks 4 to 7: monitor
Data sources are connected and live checking begins. Every incoming invoice, delivery record and SLA report is checked against the verified terms. Findings appear on a dashboard with the clause, the calculation and the source document.
Week 8: report
We quantify every finding, separate what is recoverable from what is a process fix and set out the business case for a wider rollout.
Before you start
Use the leakage calculator to estimate the prize, then read what our proof of concept includes.