Automating Duty Drawback for SMBs

Diving deeper into

Penny Chen, CEO of Pax, on building AI-powered tariff refunds

Interview
AI-native startups like Pax and tech-forward logistics leaders like Flexport have the potential to expand duty drawback across Fortune 500 through SMB
Analyzed 5 sources

The real unlock is not better tax advice, it is lower processing cost on messy trade data. Duty drawback already returns up to 99% of duties, taxes, and fees on eligible claims, but most providers still depend on manual cleanup of invoices, bills of lading, and HTS data, which makes smaller claims uneconomic. Pax attacks that bottleneck directly, while Flexport can bundle the same automation into the freight and customs workflows customers already use.

  • Legacy drawback firms were built for large accounts. Charter Brokerage was large enough to be acquired by Berkshire Hathaway in 2014, and the core model was still specialist service work, not software that cheaply handles thousands of small and mid sized claims.
  • Pax describes the old workflow as months of manual record collection and spreadsheet cleanup, with many providers unwilling to touch claims under about $100K. Its product uses LLMs for document extraction, then deterministic optimization to match imports to exports and maximize the refund.
  • Flexport comes at the market from the other side. Because it already sits inside booking, customs filing, ERP, and shipment records, it can turn drawback into one more software feature on top of an existing logistics system, and says established clients can receive payments in as little as six weeks.

As tariffs spread across more products and supply chains fragment across more countries, drawback is likely to shift from a niche enterprise service into a standard trade operations tool. The winners will be the companies that can ingest raw shipping records, classify them correctly, and file compliant claims at software speed, not analyst speed.