The problem
A global food services company — about $11B in annual revenue and roughly $3B in food and supply spend — had negotiated strong rebate contracts with its suppliers. Those contracts reward volume: buy the right mix of products and the discounts unlock.
But purchasing ran location by location, and no single site had visibility into which items triggered the best network-wide rebates. The result was quiet leakage — savings the contracts explicitly allowed for never showed up, because day-to-day buying didn't line up with the rebate structure.
What we built
We started with the economics, not the technology: the money was already sitting in the contracts. So we built SupplyAI, a purchasing engine that reads what every location actually buys, checks it against the terms of each rebate contract, and recommends the product choices that maximize total rebates across the whole network — optimizing for the system, not one site at a time.
The recommendations land where buyers already work, so acting on them never means learning a new tool. Every suggestion traces back to the contract clause that justifies it.
The outcome
In its first year, rebate capture improved by 5% — measured against the same contracts and the same spend. Against roughly $3B of annual purchasing, a five-point improvement in capture compounds every year the contracts renew.