Choose Custom AI for Complex Logistics Carrier Management

Many small logistics companies rely on the rating module within their Transportation Management System (TMS). These systems work well for major national carriers with standardized API connections. The problem starts with the regional or specialized carriers that provide a competitive edge but only offer PDF rate sheets or an online portal.

The TMS data model is fixed; it cannot adapt to a carrier's unique pricing structure, like a 10% discount on a specific lane only on Tuesdays.

Consider a 15-person freight brokerage managing LTL shipments. Their TMS integrates with three national carriers, but they have pricing agreements with seven other regional carriers. For every new load, a dispatcher must manually check rates by opening seven separate web portals or PDF documents. They might miss a crucial accessorial charge for liftgate service on one carrier's 12-page PDF, leading to an unprofitable load. This manual process takes 20 minutes per quote and introduces significant risk of human error.

The structural issue is that off-the-shelf software is built for standardization, not specialization. These platforms cannot ingest and interpret unstructured documents like a carrier's uniquely formatted rate agreement. They force you to fit your business into their predefined logic. But a small logistics firm's advantage often comes from these exact non-standard agreements that generic software cannot comprehend. The result is a permanent ceiling on operational efficiency.

The first step is a carrier data audit. Syntora would collect every rate sheet, portal login, and API documentation for all your carriers. We map every pricing rule, fuel surcharge calculation, and accessorial charge. You receive a structured data map showing exactly how each carrier's pricing logic will be modeled before any code is written.

The core system would be a FastAPI service deployed on AWS Lambda for low-cost, on-demand processing that costs under $50 per month to run. For PDF and Excel rate sheets, the Claude API parses the documents into a structured JSON format that the rating engine can use. Pydantic models enforce data consistency for each carrier. This service exposes a single API endpoint that accepts a shipment query and returns a ranked list of carrier rates in under 500ms.

The delivered system integrates directly with your existing TMS, writing the best rate back to a custom field, or can be used via a simple web interface. When a new load is entered, the system queries all carriers in parallel and displays the results in one place. You receive the full Python source code, a runbook for adding new carriers, and a Supabase dashboard to monitor processing logs and any parsing errors.