Articles
Operator notes.
Problem-first writing on AI, automation, and the process underneath. Each piece names a real bottleneck, exposes the broken layer nobody owns, and shows what the working system actually looks like.
When a carrier exposes its automation stack as an API, the failure mode moves
Amazon-grade logistics automation is becoming reachable by API. The catch: an automation-first network won't tolerate messy data. Automating shipment status updates starts with the prep layer nobody wants to own.
4 min read → Enterprise AIMulti-agent orchestration won't fix a process that's already broken
Multi-agent orchestration beat a newer model on a 2026 benchmark. The ablation shows the gain came from when agents shared findings, not from adding agents.
9 min read → AI governanceShadow AI isn't a policy gap. It's an inventory gap.
Shadow AI governance fails at inventory, not policy. 47% of enterprises can't list their agents, and the new AI control planes only discover what they're wired to.
10 min read → Optimization / ORCustom constraint optimization: the utility scheduler no off-the-shelf product could build
When scheduling constraints interact, custom constraint optimization beats any product. A 96% workday-efficiency engine, and why an AI agent can't replace the solver.
9 min read →