Case studies / helix-freight

Helix Freight — autonomous dispatch

Client — Helix FreightSector — LogisticsDuration — 14 weeksTeam — 4 engineers
71%
faster assignment cycle
4 depots40 agents14 weeks
Problem

Load matching ran on four spreadsheets and a phone tree. Dispatchers held 40 minutes of latency per assignment, and empty-mile cost had grown 19% year over year.

Intervention

A 4-agent fleet on Mesh: one planner per depot, tool-scoped executors against the TMS API, an eval gate on every commit, and a policy engine holding hours-of-service constraints hard.

Outcome

71% cycle-time reduction, 12.4% empty-mile savings, and dispatchers moved to exception handling only — verified against six months of post-deploy ledger data.

dispatch queue depth — 30s window
71%faster assignment cycle
12.4%empty-mile reduction
0.4%escalated to a human
planner topology
throughput / depot
0.4%
escalated to a human

"Our dispatchers stopped firefighting in week three. By week nine we had removed a whole shift pattern."

HFHelix FreightVP Engineering
Share this case study
← Previous case
Aster — intake summariser
Next case →
Meridian — ZK settlement rail
Q4 2026 — two slots

Let's Build the Thing You've Been Putting Off.

15 minutes with the engineer who'd actually build it. If we're not the right fit, we'll tell you straight and point you somewhere better if we can.

No sales pitchNDA on requestReply in 1 business day