Operating model · 9 minute read
Designing the human-agent operating model
How to define roles, decisions, handoffs and performance when people and agents share work.
- Operations
- Reinvent
- Global
- Design
Decision lens
How should work, authority and accountability change when agents join an operating team?
Deploying an agent without redesigning the operating model leaves ambiguity exactly where risk and adoption concentrate: decisions, exceptions and ownership.
Redesign work, not job titles
Break the workflow into information, recommendation, preparation, execution and exception activities. Assign each activity according to capability, impact and accountability.
- Preserve accountable decisions
- Expose invisible coordination work
- Design for exceptions first
Make authority operational
A responsibility chart is insufficient unless identity, permissions, approval interfaces, escalation and stop authority enforce the agreed design.
- Map decision rights
- Test prohibitions
- Give people recovery authority
Measure the combined system
Track workflow quality, cycle time, human intervention, exceptions and adoption together. Optimising agent utilisation alone can damage service performance.
- Measure outcome, not automation rate
- Review role load and exception volume
- Recertify as work changes
Practical checklist
Evidence to bring into the decision.
- 01Workflow decomposition
- 02Decision-rights matrix
- 03Exception ownership
- 04Adoption measures
- 05Periodic recertification
Continue with evidence
Connect the perspective to an operating method.
Human-agent design must reflect local employment, regulatory, cultural and operating context. This perspective is not workforce or legal advice.