Framework
The Gottfredson-Cavalier Matrix
How hard is the task, and how much does it cost when it goes wrong? The answer sets how much human accountability a task keeps.
Impact of failure: high
Lower difficulty, high impactSimple, but a miss is costly. Keep a human sign-off.
Higher difficulty, high impactHard and high-stakes. Human review on every output, or human only.
Low
Lower difficulty, low impactStrongest candidates to automate.
Higher difficulty, low impactAutomate with exceptions, errors caught downstream.
Task difficulty: lower
Higher
Risk before automation.
The matrix places every task by its difficulty and by the critical impact of failure, then reads the Human-AI Task Scale against that position. A task that is easy to automate can still be the wrong one to hand over, if a single failure is expensive.
It builds on Dr. Conrad Gottfredson's task difficulty and critical impact of failure matrix, owned by APPLY Synergies, LLC, and combines it with the Human-AI Task Scale. It is co-authored by Conrad Gottfredson and Josh Cavalier.