Why agentic TMS isn't an LLM wrapper — and why that matters for your margin.
A teardown of the difference between a reasoning agent and a chatbot sitting on top of a workflow engine. Plus, the economics of each in real freight ops.
Deep notes on agent architectures, freight economics, voice in logistics, and the teardown of legacy TMS myths — from the people building Veltrum and the operators using it.
A teardown of the difference between a reasoning agent and a chatbot sitting on top of a workflow engine. Plus, the economics of each in real freight ops.
What the average broker's week looks like by the minute. Where voice automation actually works, and where it still gets in the way.
The training corpus, evaluation benchmarks, and failure modes of our freight-tuned foundation model. With charts.
A cost model for US-based and nearshore dispatcher teams, with a break-even analysis for autonomous alternatives.
A field report from a mid-market broker's transition to agentic operations — with the metrics, the setbacks, and the playbook.
Notes on the failure modes we hit when we first shipped an autonomous rate negotiator — and why the policy layer is now non-negotiable.
How we built an autonomous claim filer that recovered $1.4M for a single customer in twelve months. And why it keeps getting better.
An engineering walkthrough of how we got voice agents fast enough that drivers don't realize they're not talking to a person.
Why every Oracle TMS contract up for renewal between now and 2028 is a deployment window — and what the market is actually doing with it.
One essay every other week. No lead magnets, no sponsorships, no email flooding.