Dependable outcome
A result from an AI-assisted workflow that is actually usable without further repair — the right denominator for measuring the human attention and cost AI automation really consumes.
A dependable outcome is a finished result you can rely on: complete, correct enough for its purpose and ready to be used, without someone quietly fixing it first. It is the unit that matters when you ask what an AI-assisted process really delivers.
Counting the agent's runtime, the number of tasks started or even the number of outputs produced overstates progress. Many of those outputs need review, correction or rework before they count. The useful denominator is therefore the outcome that survived that process.
Why it anchors the measurement
Once you count dependable outcomes, you can ask how much human attention each one required: context supply, prompting, review, approval, correction and exception handling. That is supervision load. You can also ask what each cost in money, which leads to the cost per accepted outcome.
It also keeps automation rate honest, because a high rate only matters if the results are dependable.
Defining what counts as dependable, for a given workflow and risk level, is a design decision worth making explicit before launch.
Read more in The hidden metric in AI automation is supervision.
Related terms
Supervision load
The human attention an AI system still consumes — context supply, review, approval, correction and exception handling — measured per dependable, completed unit of work.
Cost per accepted outcome
The total cost of getting one AI-assisted result that is actually accepted and used, counting model spend, retries, human review, exceptions and recovery — a truer measure than cost per call.
Automation rate
The share of work an AI system completes without a person doing it. A useful headline number that says nothing on its own about the human attention left behind.