Most of your inbox doesn’t need your most expensive AI
In a 100-email freight broker test, a lightweight classifier sorting first kept 99% accuracy and cut cost 57%. Two-thirds of emails never needed Claude.
Insights
Where AI actually pays off
Practical writing for business leaders deciding what to automate, what to leave alone, and how to tell the difference. No hype, no vendor pitches.
A Practical Reading of the NIST AI Framework
The NIST AI Risk Management Framework, reduced to what an operations-heavy business can actually implement: four functions, four questions, and a one-session start.
Automate the copying before you automate the deciding
Most logistics teams aim AI at decisions. The bigger win is the re-keying between systems, and it fails visibly, which makes it the safer place to start.
The AI compliance deadline moved. The build decisions didn't.
Europe delayed its high-risk AI rules to December 2027. The parts of compliance that shape how a system is built cannot wait, because they cannot be retrofitted.
GTM Engineering: How Fast Can Your Revenue Team Safely Move?
GTM engineering helps revenue teams build and automate faster. The challenge is knowing when an experiment has become production software.
When Software Gets Faster, Something Has to Change Besides the Timeline
AI did not just make writing code faster. It changed where the time actually went in software projects, and how they should be specified, built, and priced.
Want this applied to your own operations?
A free AI assessment is a working session on your actual processes — which are strong automation candidates, and which are not worth the effort.
Book a free AI assessment