Design Pattern · September 2026
From human work to AI workforce
Automation follows understanding. Work becomes explicit, then observable, then repeatable, and only then is it a candidate for AI execution with human exception handling.
Under the pillar Zero Neo
The progression
Human work becomes explicit workflow, then observable workflow, then repeatable workflow, then AI-assisted workflow, then AI-executed workflow with human exception handling. Skipping a stage does not accelerate the sequence; it just moves the failure later, where it costs more.
Start with the work, not the agent
The common approach is to ask where an AI agent could be deployed. That question presumes the work is already understood, which is usually the thing that is missing.
Only automate work that is sufficiently explicit, observable, testable, and governed. Each of those is a precondition, and each is checkable before anything is built.
Consolidate before automating
This is where Zero Neo departs from the current fashion of layering AI agents over already fragmented application estates. An agent stitched across nine systems inherits every inconsistency between them.
The open research question is whether application consolidation should precede AI automation as a rule. The concern is straightforward: automate a fragmented environment and you have automated the fragmentation, at higher speed and lower visibility.