Back to Case Studies
A Fortune 500 Secure Logistics Leader Turns Decades of Contract Data into Auditable Intelligence
Structuring 7,000+ customer contracts, correlating against operational data, and surfacing multi-million-dollar revenue recovery previously invisible to the business
Case study at a glance
- Customer: A Fortune 500 provider of secure logistics and cash management services
- Industry: Secure logistics, cash management, security services
- Challenge: Decades of unstructured customer contracts disconnected from operational data
- Solution: Seekr’s AI-ready data engine: structured extraction, correlation with a modern cloud data warehouse, full provenance
- Results: Multi-million-dollar revenue recovery opportunity identified across approximately 5% of North American locations; deployment expanded across five business areas under a multi-year enterprise license
The Customer
A Fortune 500 provider of secure logistics and cash management services, operating across more than 100 countries, managing billions of dollars in customer assets every day, and running an operating model built on customer contracts dating back to the 1990s that govern service frequency, pricing, and billing terms.
The Challenge
The customer’s operational infrastructure was modern, but its contract environment was not. Service records – every truck stop, every cash pickup, every change order – flowed into a a modern cloud data warehouse data warehouse cleanly. The commercial and finance teams had good visibility into what was happening operationally.
But more than 7,000 customer contracts, many dating back to the 1990s, lived as PDFs in a document archive. Each contract held the terms that were supposed to govern service frequency, pricing escalation, change order handling, and invoicing. None of it was structured, queryable, or connected to the operational systems.
This gap created a structural risk: If a contract specified five service stops per month at a location and the field team was making seven, whether or not the invoicing reflected the additional service was ambiguous. If pricing terms had escalated under a contractual clause, but the billing system had not picked up the change, the customer would under-collect. None of this was the result of bad intent or bad operations. It was the result of running a multi-billion-dollar operation against a contract portfolio that no one could query at scale.
Industry research from Sirion places typical revenue leakage in large enterprises at 2% to 9% of annual revenue. The customer’s leadership suspected they were somewhere in that range. What they needed was a system that could tell them where, by how much, and with enough auditable evidence that finance could act on it.
The Approach
Schema definition. Before any extraction ran, Seekr worked with the customer’s data and commercial teams to define the structured schema—which contract fields mattered, how they should map to the operational data model, and what edge cases the extraction pipeline needed to handle. Every extraction decision downstream was governed by the schema agreed at this stage.
Structured extraction. Seekr’s AI-ready data engine ingested all 7,000+ contracts. For each one, the system extracted the fields defined in the schema (pricing terms, service frequency, location identifiers, change order history, escalation clauses, renewal dates) and produced structured records that could be analyzed against the customer’s operational data.
Provenance preservation. Every structured value carried a pointer back to its source: the specific contract, the specific page, the specific clause. When a discrepancy surfaced in the joined view, the finance team could click through to the underlying language. The output was not a recommendation. It was auditable evidence.
Correlation with the data warehouse. The structured contract data landed in a modern cloud data warehouse alongside the customer’s existing service records to surface gaps between what the contracts said should be happening and what the operational data showed was happening.
The Results
The initial deployment identified a multi-million-dollar revenue recovery opportunity across approximately 5% of the customer’s North American locations.
The finding was consistent with what industry research would have predicted. It was also, for the first time, actionable. Every flagged discrepancy came with a source contract clause and a corresponding operational record. The finance team could review, prioritize, and act.
Beyond the initial financial finding, the deployment produced three durable outputs:
- A queryable structured dataset covering the customer’s full contract portfolio, joined against operational data in a modern cloud data warehouse.
- An extraction pipeline that continues to process new contracts and amendments as they are executed, keeping the dataset current.
- An auditable evidence trail supporting every flagged discrepancy, ready for finance, compliance, and commercial review.
Why It Worked
Several factors separated this deployment from enterprise AI pilots that stall short of production. The customer’s team came in with a clear definition of success and the operational infrastructure to act on findings the moment they surfaced. The scope of one use case, one data domain, and one integration path was narrow enough to prove out in a defined window. And the schema work happened before any extraction ran, which avoided the semi-structured mess that most extraction pilots end up in.
None of these factors are specific to secure logistics. Any large enterprise with a legacy document corpus and a modern data warehouse can apply the same solution.
What’s Next
The customer has expanded the deployment across five business areas under a multi-year enterprise license. The AI-ready data infrastructure will support additional use cases beyond revenue leakage, including contract renewal management, commercial analytics, and compliance monitoring among them.
About Seekr
Seekr is the leader in explainable, defensible AI built for critical decisions in environments that demand accuracy and accountability. Seekr’s technology and products help enterprises and government agencies deploy domain-specific large language models (LLMs), vision language models (VLMs) that understand the physical world, and AI agents trained on their own data – across any infrastructure, including sovereign deployments. Backed by robust verification and validation tools that surface the provenance and intent behind every model decision, Seekr delivers AI that organizations can audit and defend across all modalities. Learn more at seekr.com.
Deployment details anonymized pending customer disclosure approval. Industry research citations: Sirion (contract revenue leakage benchmarks); IDC (enterprise data composition); Gartner (AI-ready data).