Court-admissible evidence, from operational data.

Context

A logistics company suspected systematic violations of non-compete agreements by drivers and subcontractors. Conventional controls produced suspicion. They never produced the hard evidence required for legal action.

Problem

Internal monitoring caught the symptoms: scheduling anomalies, unusual idle time, missing capacity. The behaviour driving those symptoms stayed hidden. Cross-referencing patterns against external transport marketplaces and GPS at scale was beyond the operations team's tooling.

Solution

We built a system that correlates driver behaviour with external marketplace activity. Algorithms fused three data streams into one behavioural model:

• GPS tracks
• Work-time records
• Listings from external transport marketplaces

The output: a court-admissible evidence report per driver and per subcontractor.

Implementation

Integrating the three sources (GPS, work-time, external marketplaces) into a single evidence report was the first hard problem. Formats, units and identifiers had to be normalised before anything else worked. We then built analytical models that flagged anomalies indicating competitive activity, and assembled the findings into a defensible report.

Outcome

The system detected the violations. The report, the evidence trail and the methodology were accepted by court as credible material. The client successfully enforced its non-compete provisions on the strength of it.