Infrastructure for Mode Selection & Intermodal Optimization

Mode selection requires documented, findable rules defining when intermodal is appropriate: minimum transit time buffers for rail, commodity restrictions (hazmat, perishables), origin-destination pairs with rail ramp access, and customer service level agreements that prohibit mode switching without approval. The freight baseline confirms that customer-specific handling knowledge — including mode preferences — is tribal. For the AI to automatically route freight to intermodal, these eligibility criteria and customer constraints must be current and queryable, not held by individual freight brokers.

Mode optimization requires systematic capture of rates and transit times across all modes — truck, rail, intermodal, air, ocean — for each origin-destination pair, along with capacity availability signals and carbon emission factors by mode. TMS captures truck-mode transactions, but intermodal and rail rate data often arrive through separate carrier EDI feeds. Template-driven capture ensures that multi-modal rate comparisons include all required attributes: transit days, reliability score, carbon factor, and ramp-to-ramp versus door-to-door cost.

Mode optimization requires consistent schema across all rate and service records: origin, destination, mode, carrier, rate, transit days, reliability percentage, and carbon emission factor. TMS provides structured fields for truck lanes, and the same schema must extend to intermodal and rail data for comparative analysis. L3 consistent schema enables the AI to rank mode options on a common cost-service-carbon framework without manual normalization of differently structured rate sources.

Mode selection AI must query TMS for shipment characteristics, rate management systems for multi-modal pricing, capacity availability feeds for rail and intermodal, and customer SLA repositories for mode eligibility. API access to these systems enables automated mode recommendations without requiring a planner to manually gather rates from multiple systems. The freight baseline confirms legacy TMS API limitations, but achieving API access to the primary rate and SLA systems is the minimum for automated intermodal routing.

Mode selection models must update when rail rates change, when intermodal ramp schedules shift, or when carbon emission factors are revised per GLEC or EPA updates. Event-triggered maintenance ensures that when a rail carrier adjusts transit times for a corridor, the mode comparison engine reflects the change before routing decisions are made on stale data. The baseline confirms freight rates lag by days — for mode optimization this causes systematic misrouting when cost parity between truck and intermodal shifts mid-contract period.

Mode selection requires integrating TMS shipment records, multi-modal rate systems (truck, rail, air, ocean brokers), customer SLA repositories, carbon emission factor databases, and tendering systems for each mode. API-based connections allow the AI to pull current rates and transit options across modes and push mode recommendations to the appropriate tendering workflow. The freight baseline confirms siloed systems require manual reconciliation. L3 API integration across the core mode data sources is necessary for the multi-modal comparison to function without manual assembly.