Case study · Procurement intelligence
Pierpont Holdings gave its procurement leaders one view of 2.7M records.
Native helped Pierpont build EP2, three AI agents on its existing Microsoft SQL infrastructure, so procurement leaders can see revenue at risk, ask questions in plain English, and pull supplier reports on demand.
EP2 queries 2.7M records in 11 seconds and surfaces revenue risk in seconds instead of days, with a clear owner and escalation path for each issue. It runs on Pierpont's existing Microsoft SQL infrastructure with no disruption to it.
About Pierpont Holdings
Pierpont Holdings is a procurement intelligence firm. Its co-founder, Walt Charles III, is a seven-time Chief Procurement Officer.
The challenge
Procurement leaders control trillions in global spend with limited visibility. Escalations surfaced too late, revenue risk hid inside operational noise, and data was spread across dozens of systems.
Analysis depended on scarce data scientists, and executive reviews ran on anecdotes. Ticketing systems showed that someone needed to buy something, but not where the risk was, who was accountable or how much money was at stake.
What we built
How it works
- 01AskAn executive asks a question in plain English.
- 02GenerateThe agent routes it to the right data and writes candidate SQL, drawing on past validated pairs.
- 03Execute and validateThe query runs against live data, and the results are checked for completeness and correctness.
- 04RefineIf validation fails, the agent refines the SQL and runs it again until it passes.
- 05LearnThe successful pair is stored in a vector database, so the next similar question is faster and more accurate.
Results
Supplier research that took weeks of manual work and slide building now happens on demand, as executive-ready reports with embedded charts.
“I wanted a single pane of glass for my world, one place where I can see my people, my risks, my spend, and where value is leaking, and then act on it immediately.”
