Enterprise motor carriers operating 250+ power units face increasing pressure to extract maximum utilization from assets while maintaining service levels. Part 9 of the Applied AI for Enterprise Motor Carriers Systems Playbook focuses on three production-grade systems that deliver measurable ROI: real-time telematics ingestion pipelines, mixed-integer linear programming (MILP) dispatch optimization engines, and straight-through processing (STP) billing architectures.
Strong Tower Consulting has helped carriers running Truckmate, McLeod, and TMW systems implement these patterns since 2008. The architectures described below are vendor-neutral and designed to augment - not replace - your existing TMS.
Real-Time Telematics Ingestion Architecture
Modern fleets generate millions of GPS, engine, and sensor events daily. A robust ingestion layer must normalize data from multiple ELD providers (Samsara, Geotab, Motive, PeopleNet, etc.) and make it immediately available for downstream optimization and compliance systems.
The reference architecture uses Apache Kafka as the event backbone with schema registry enforcement. Each telematics provider streams via MQTT or HTTPS webhook into isolated topics. A fleet of containerized normalization microservices then transform vendor-specific payloads into a canonical event model stored in both PostgreSQL (for recent 72-hour operational data) and a columnar store optimized for analytics.
| Fleet Size (Power Units) | GPS Pings | Engine Events | Exception Events | Total Events |
|---|---|---|---|---|
| 250-500 | 1.8M | 420K | 18K | 2.24M |
| 500-1,000 | 3.9M | 910K | 37K | 4.85M |
| 1,000-2,500 | 8.2M | 1.95M | 82K | 10.23M |
MILP Dispatch Optimization Engine
Rule-based dispatch systems reach their limit at scale. Mixed-integer linear programming provides mathematically optimal load-to-driver and load-to-asset assignments while respecting hard constraints such as hours-of-service, equipment compatibility, driver preferences, and customer requirements.
The formulation we deploy for clients typically includes:
- Binary decision variables for each feasible driver-load pair
- Continuous variables for arrival times and fuel purchases
- Objective function balancing revenue, deadhead miles, and driver dwell time
- Constraints for 14/11/70 hour rules, equipment types, and detention windows
Production implementations solve nightly using Gurobi or CPLEX with warm-start from the previous solution. Real-time re-optimization is triggered by exception events (driver illness, load cancellation, or significant traffic events ingested from telematics). Typical results show 4-7% reduction in empty miles and 6-9% improvement in on-time performance versus heuristic dispatch methods.
Straight-Through Billing Pipeline
EDI 210 invoicing remains a manual bottleneck for many carriers. A mature STP pipeline ingests EDI 214 status updates, cross-references them with telematics proof-of-delivery events, applies fuel surcharge and detention logic, then generates and transmits the 210 without human intervention in over 85% of loads.
Key integration points include:
- SQL Server or PostgreSQL stored procedures that calculate line-haul, fuel surcharge (using EIA weekly averages or carrier-specific matrix), accessorials, and detention based on actual in-yard and at-shipper times from telematics.
- Driver settlement integration that pulls from the same source data to ensure consistency between what is billed to the shipper and paid to the driver.
- Exception queue for the remaining 15% of loads requiring manual review (missing signatures, disputed detention, etc.).
Exception Management and Human-in-the-Loop Design
No optimization system is perfect. The most successful implementations include a purpose-built exception management console that surfaces only the loads or drivers that fall outside confidence thresholds. Dispatchers and billing clerks work from a prioritized queue rather than scanning the entire operation.
This design pattern reduces decision fatigue and allows carriers to reallocate experienced personnel from routine work to high-value exception handling and customer relationship management.
Implementation Roadmap for Enterprise Carriers
- Assess current TMS (Truckmate, McLeod, TMW) data model and identify gaps in real-time event capture.
- Implement canonical telematics ingestion layer with initial focus on GPS and HOS events.
- Build the MILP model using historical load and settlement data. Validate against 90 days of manual dispatch decisions.
- Develop straight-through billing rules engine starting with line-haul and fuel surcharge logic.
- Integrate exception management UI and begin phased rollout by region or customer segment.
Frequently Asked Questions
How long does it typically take to achieve positive ROI on a MILP dispatch optimizer?
Most carriers see measurable improvements in asset utilization within 90-120 days of production deployment. Full payback on the systems engineering investment usually occurs between 7 and 11 months depending on average length of haul and current empty mile percentage.
Can these architectures work with both Truckmate and McLeod TMS platforms?
Yes. The patterns are deliberately TMS-agnostic. We have successfully implemented them against Truckmate SQL Server databases, McLeod SQL Server and Oracle instances, and TMW systems. The key is creating clean, version-controlled data access layers rather than direct table queries.
What telematics providers do you most commonly integrate?
We maintain production connectors for Samsara, Geotab, Motive (formerly KeepTruckin), PeopleNet, and Omnitracs. The ingestion architecture is extensible - adding a new provider typically requires 3-4 weeks of development and testing.
How do you handle fuel surcharge and detention calculations in the straight-through billing pipeline?
The billing engine stores both contract-specific surcharge tables and EIA-published weekly averages. Detention logic uses actual GPS-derived arrival and departure times cross-referenced with EDI 214 events. All calculations are fully auditable with complete lineage back to source telematics and load data.
Strong Tower Consulting, LLC and Exspeedite LLC are separately organized sister companies under common ownership. Founded in Minnesota in 2008 by Scott Elliott, we have been providing independent systems architecture and technology advisory services to motor carriers for more than 18 years.
Our only product is better decisions.