Article based on what I see at a Telco and one of my Tech clients.
We are spending a lot of time discussing how to make AI agents more capable. Are we overlooking another problem in the process?
What happens when the agent works, but nobody trusts the output enough to let it run?
A recent G2 report calls this the “Distrust Tax.” The productivity gains we lose to human oversight, rework, and verification. I think there is more to this.
The Distrust Tax is a tax on ambiguity. AI just collects it.
Based on my experience, in many organizations, human judgment has quietly compensated for poorly defined processes, conflicting objectives, and decisions that rely on years of individual experience. AI agents expose that problem. And suddenly, what was previously absorbed by people becomes an operating cost.
I explore this through a simple framework: Ambiguity × Stakes, and why defining what good looks like may be more important than adding another layer of human supervision.
