Procurement was the biggest cost line and the messiest. Inflation and supplier price rises were eating margin, spend was split across four separate systems, and buyers had no way to compare contracts or even see that two departments were paying different prices for the exact same service.
Because spend lived in four separate systems, buying power was fragmented. Different departments bought identical equipment from different suppliers at different prices, and the company never used its own scale to negotiate as one.
Contracts made it worse. No one could track thousands of agreements and their annexes, so the business missed volume discounts it had already earned and took penalties because contract terms were never enforced with suppliers in practice.
And negotiation ran on instinct. Buyers haggled based on their personal relationship with a supplier, with no benchmark of competitor pricing or current market rates to anchor a hard target.
Turn four systems into one clear picture, then let AI catch the leaks and set the negotiation targets.
We pulled all purchasing data from the ERP systems into one repository and a machine-learning model auto-classifies every spend line, folding hundreds of scattered entries into clean categories, giving complete visibility for the first time.
A language engine reads thousands of scanned procurement contracts, extracts the payment terms, renewal dates and supplier commitments, and cross-checks them against the actual invoices to surface overcharges and duplicate payments.
Buyers get a tool that analyses purchase history, market trends and competitor pricing and produces a negotiation brief with firm price targets, ready in hand the moment before they sit down with a supplier.
The same cross-check that reads contracts against invoices flags invoices paid twice and discounts that were promised but never applied, turning silent leakage into money the company can claim back.
With every category visible across departments, the business consolidates suppliers and negotiates on its full scale, so it stops paying three prices for the same thing.
Purchasing is steered to approved suppliers automatically, which ends the off-contract spend that used to leak value and speeds the approval of legitimate new suppliers.
The biggest cost line, finally under control.
In under a year, indirect spend fell 12% through supplier consolidation and data-backed negotiation, and the company recovered significant sums that had been leaking through duplicate invoices and unclaimed volume discounts.
Approving new-supplier contracts sped up 40%, with off-contract buying shut down and purchasing automatically steered to approved suppliers, so control and speed improved at the same time.
Procurement leaks because the spend is invisible and the contracts go unread. The value here is one categorised view across every system, an engine that reads every contract against every invoice, and a buyer who walks in with a market-anchored target, so the saving is repeatable, not a one-off audit. It draws on the same document-and-anomaly discipline as our audit-analytics work, aimed at spend and suppliers rather than the ledger.
An audit firm relied on small random samples that missed sophisticated fraud. AI analysing 100% of the general ledger flagged fraud patterns and control gaps, moving the firm to full coverage and cutting fieldwork sharply.
Read the case → Portfolio & cost · Consumer goodsAn over-extended catalogue of 1,500+ SKUs hid its true cost to serve. AI cost analysis cut the range 25% with no revenue lost, dropped line downtime 35%, and freed working capital tied up in surplus stock.
Read the case →
Reef TRH