Frequently asked questions

Why does automation fail for freight rate sheet processing?

Automation works well on stable, standardised carrier formats. The problem is that formats are not stable. When a carrier changes their template, every automation rule breaks. Managing dozens of carrier relationships means managing dozens of formats evolving independently. The maintenance cost of keeping automation current often exceeds the cost it was built to replace.

Why can AI not fully solve freight rate sheet processing?

AI handles clean, structured rate sheets well. It fails on the hard ones — 700-page PDFs with misaligned data, amended tariffs where crossed-out rates disappear in text conversion, and complex tariffs where prices are governed by footnotes. These are not edge cases. They are regular occurrences at volume. Without domain expertise to guide it, AI accuracy falls on exactly the rate sheets that cause the most commercial damage.

Why does EDI not solve the rate sheet processing problem?

EDI works well where carriers have adopted it. The limitation is coverage. The long tail of smaller and regional carriers, often those with the most complex rate structures, never adopted EDI and are unlikely to. Rate network automation built on EDI also lacks a human validation layer, so non-standard formats still break the process.

Will carrier APIs eventually solve rate sheet processing?

Progress is real among the largest carriers. But smaller and regional carriers lack the incentive and investment capacity to build and maintain rate APIs. The timeline for broad enough adoption to materially reduce the processing burden for most freight forwarders is years, not months.

Why does human expertise alone not scale?

Specialist domain knowledge is genuinely valuable and irreplaceable for complex rate sheets. But headcount cannot scale at the same rate as carrier relationships. When key specialists are unavailable the pipeline stalls. And cost per sheet stays constant regardless of volume.

What is the right approach to freight rate sheet processing?

Each approach solves part of the problem. The right answer is a designed system that combines automation, AI and human expertise in the right measure, right where each is needed. That is what Part 3 of this series covers.