Redesign the work, not just the technology.
Turn an AI-enabled workflow into a clear operating agreement between people and agents—who owns the outcome, who may decide, how handovers work and what proves the model is taking hold.

Decision
Output
Method Boundary & Connections
Start here when
Does not decide
Reassess when
Connection to IDAO
Connection to Agent Authority
A rollout installs a tool. An operating model changes how work runs.
- Configures access and integrations
- Explains features and prompts
- Tracks attendance, licences and usage
- Supports initial technical adoption
- Changes roles, accountabilities and spans
- Reallocates decision and intervention rights
- Redesigns handovers, controls and incentives
- Proves outcomes persist in live operation
Start with one real workflow. Finish with an operable design.
- 01
Trace the live work
Map the work as it happens: demand, decisions, exceptions, queues, evidence and the people who carry hidden coordination.
OutputWork and handover baseline - 02
Redesign roles and handovers
Decide what people lead, what agents support or execute, and how ownership moves without losing context or accountability.
OutputRole and handover design - 03
Set decision rights
Name who proposes, approves, acts, intervenes and remains accountable for each consequential decision and handover.
OutputDecision-rights map - 04
Build operating capability
Define the judgement, supervision, exception handling, evidence literacy and improvement routines each role needs.
OutputCapability and enablement plan - 05
Align incentives and measures
Remove targets that reward unsafe automation or hidden rework. Measure confident use, intervention, quality and sustained outcomes.
OutputAdoption measures and incentive changes
Handover Choreography across IDAO
Identify roles and constraints.
Test handover logic safely.
Install explicit rights pointing to Agent Authority.
Measure and loop back.
Make authority visible at every move.
A role title is not a control. The map records what the person and agent may do at each point, the evidence they need and the condition that moves authority back to a person.
| Right | Person | Agent |
|---|---|---|
| Frame | Sets the outcome, boundaries and evidence required | Contributes options and operating evidence |
| Recommend | Challenges assumptions and interprets context | Produces traceable analysis or a proposed next action |
| Approve | Retains authority where exposure requires it | Waits at the defined gate and preserves the approval record |
| Act | Handles exceptions and actions reserved for people | Executes only within its permitted authority and constraints |
| Intervene | Pauses, overrides, escalates or demotes authority | Surfaces thresholds, uncertainty and control breaches |
Consequential handovers need an authority ceiling.
For every explicit propose, approve, act, intervene, and demotion right that can materially affect a person, record, system or service, point to the Agent Authority Model calculation to set permitted autonomy.
Prepare people to operate, challenge and improve the system.
The capability plan is role-specific. It combines practice in live scenarios, observed proficiency and support at the moment of work—not a generic course-completion target.
Role practice
Rehearse the normal route, exceptions and fallback with the people who will own them.
Control fluency
Recognise uncertainty, challenge evidence, intervene and document why authority changed.
Handover discipline
Pass context, state and accountability without creating shadow coordination.
Improvement ownership
Use operating evidence to adjust roles, controls and workflow—not only the model.
Measure behaviour, control and outcomes—not logins alone.
Activation begins when the redesigned route can run safely in live work. Sustained operation begins when the client team owns the routines, evidence and improvement cycle.
Use
- Eligible work using the new route
- Active use by role and team
- Fallback to the old process
Control
- Interventions made in time
- Exceptions resolved by the named owner
- Authority demotions and control breaches
Capability
- Observed proficiency in live work
- Confidence to challenge an agent output
- Time to independent operation
Outcome
- Quality and cycle-time movement
- Rework displaced rather than hidden
- Value sustained after handover
Design before launch. Learn after it.
Activate uses the role, rights, capability and measure designs to prepare the live workflow, rehearse exceptions and confirm ownership.
Operate uses real performance, interventions and workforce evidence to improve the model and return weak assumptions to the right IDAO stage.
Explore the IDAO lifecycleBring one workflow where people and agents must work together.
We will identify the first operating-model decision, the evidence it needs and the right IDAO entry point.
Book a Value Scan