Definition
Automation Readiness is the honest answer to a simple question: is this decision flow clear enough, defined enough and owned enough to hand — safely — to a system?
Why it matters
Automation applied to an unclear process produces unclear outcomes, faster. Readiness is the prerequisite that most AI programmes skip and later pay for.
Symptoms
- No shared definition of the entity the automation acts on.
- Exceptions are handled by tacit knowledge in three people's heads.
- The decision rights are informal or contested.
- Success and failure are not distinguishable from the data alone.
Typical mistakes
- Automating a workflow because a vendor demo makes it look easy.
- Confusing pilot success with production readiness.
- Treating governance as a signoff at the end rather than a design input.
Examples
A claims process reads clean end-to-end on paper. In practice, twelve percent of cases branch on judgment that is never captured. Automating the happy path is trivial; automating responsibly means first making the branching explicit.
How Zero-Gravity approaches it
We score readiness across five dimensions — definition, ownership, data, exception handling, measurability — and produce a per-domain Automation Readiness Level (ARL). We then either raise the level deliberately, or narrow the automation scope until it matches the current level.
A reliable flow must do four things: complete correctly, stop or escalate safely, recover from partial execution, and keep information and action within permitted authority.
- Correct completion
- Safe abstention or escalation
- Recovery or compensation
- Authorised information and action flow
These are qualities of reliable execution within Automation Readiness — not a separate Framework concept.