The Repair Order That Negotiated Itself.

Eighteen hundred locations. Forty thousand repair orders a month. Every order starts with an estimate — parts, labor, paint — and every estimate passes through a procurement gauntlet: carrier policy validation, multi-supplier sourcing, margin optimization. The process was manual, sequential, and slow.

A single repair order could touch five different systems before a part was ordered. Estimators toggled between screens, cross-referencing carrier guidelines against supplier catalogs against margin targets. Policy exceptions were caught late — after the part was ordered, sometimes after the repair was complete. Rework. Write-offs. Margin erosion that was invisible at the transaction level but staggering in aggregate.

Twelve percent of repair orders contained policy exceptions that were not caught until post-repair audit. Aftermarket parts substituted where OEM was required. Labor rates applied outside carrier-approved schedules. Paint material calculations that didn’t match the methodology specified in the policy. Each exception individually small. Collectively: millions in annual margin leakage.

Supplier sourcing was single-threaded. The estimator picked the supplier they knew, not the supplier the margin model would have chosen. Price variance across the same part, same region, same week — routinely fifteen to twenty percent. Not because better pricing didn’t exist. Because nobody had time to look.

The governance model encoded carrier policies as enforceable rules — not reference documents. OEM-required parts, approved labor rate schedules, paint methodology standards, supplement thresholds. Every recommendation the system produced would be validated against the carrier’s specific policy before it reached the estimator.

The sourcing layer connected to multiple supplier catalogs simultaneously, scoring options on price, availability, proximity, and margin impact. Where the policy allowed alternatives, it found the optimal one. Where the policy required OEM, it enforced the requirement and moved on. The distinction was automatic.

Optimization happened at the order level — not the line-item level. Total repair cost, total margin, total cycle time. Trade-offs made visible before the first part was ordered.

The repair order arrived pre-validated. Policy compliance confirmed. Supplier options ranked. Margin impact calculated. The estimator reviewed a recommendation, not a blank screen. Where the system found a clear path, it presented it. Where a judgment call was required — an ambiguous policy clause, a supplier substitution that needed approval — it escalated with the context needed to decide in seconds, not hours.

Every decision carried its reasoning. Why this supplier. Why this part. Why this price. Not hidden in a log file. Visible at the point of decision.

The procurement intelligence was location-specific. But the patterns were network-wide. Supplier performance, price trends, policy exception frequency — aggregated across eighteen hundred shops, the data revealed which carriers had the most ambiguous policies, which suppliers were consistently uncompetitive, and which locations were leaving the most margin on the table. The next question: predictive procurement. Not just optimizing the order in front of you, but anticipating the parts you’ll need next week based on the repairs in the pipeline today.