Skip to content

Operating model redesign

Design an operation
that can adapt.

Connect work, decisions, people and knowledge around the outcome that matters. See how AI can contribute, what must change, and how to know whether it is working.

Find your starting point

Start with the challenge.

Recognize the concern. Explore the operating choices behind it. Find a practical way forward.

Six connected dimensions.
More than one way in.

12 challenges across six dimensions

Choose a challenge to explore a first step.

Value & outcomes

Connect the investment to an accepted result.

See the connections in the model

Work & decisions

Follow the work across the whole operation.

See the connections in the model

Knowledge & context

Establish what a decision can rely on.

See the connections in the model

Authority & controls

Make information rights and action limits real.

See the connections in the model

People & capability

Give people the means to own the change.

See the connections in the model

Evidence & adaptation

Keep learning from the operation in use.

See the connections in the model

Each challenge has a starting point. The response often connects several dimensions.

Show all challenges
Recently added

How do we know when to expand, change course or stop?

The operating question

Which evidence should change the scope, authority or design of the operation?

A practical first step

Agree a baseline, acceptance rules and decision thresholds with the Outcome owner. Review service quality, Cost per accepted outcome, control performance and effects on other teams. Record the case for Expand, Continue, Narrow, Redesign or Stop.

Keep in view

Count review, correction, exceptions, support and recovery in recurring cost. Match the cost and accepted-outcome periods; with zero accepted outcomes, the unit metric is not calculable.

when-to-change-course

Evidence & adaptation

120

Recently added

The pilot worked. Will it remain dependable in everyday use?

The operating question

Which assumptions could fail when volume, knowledge, reviewers or dependencies change?

A practical first step

Turn the service claim into acceptance criteria for representative operating conditions. Test peak demand, exceptions, degraded sources and recovery. Assign monitoring, correction and re-evaluation responsibilities to named owners before wider use.

Keep in view

A successful demonstration is limited evidence about sustained operation at a different scope or volume.

pilot-to-dependable-service

Evidence & adaptation

110

Recently added

Someone owns the outcome, but cannot secure commitments from other teams.

The operating question

Who can commit the shared budget, capacity and service obligations this outcome depends on?

A practical first step

Record what the Outcome owner needs from each shared team and what each team can commit. Identify an authorized decision-maker for competing priorities, the limits of that authority and who implements the agreement. Set a follow-up against the promised service.

Keep in view

Accountability for an outcome does not imply unlimited authority over other teams.

ownership-without-commitments

People & capability

100

Recently added

People cannot see how their roles will change.

The operating question

What will people stop doing, start doing and remain accountable for as routine work changes?

A practical first step

Walk one team through the proposed work, including the exceptions that remain. Explain decisions already made and what is still uncertain. Give managers and employees funded time to test the work, build capability and raise consequences that the design has missed.

Keep in view

Generic AI training cannot resolve unclear responsibilities, an unmanageable review queue or uncertainty about staffing.

roles-in-transition

People & capability

90

Recently added

We do not know what AI should be allowed to do.

The operating question

Which specific actions may be delegated, under which conditions, and who can intervene?

A practical first step

For one workflow, agree Recommend, Act, Escalate and Never Delegate arrangements. Have the Outcome owner approve the boundary, the Control lead define checks and the Technology owner enforce them before action. Test denied requests and recovery.

Keep in view

A prompt is not an authorization control. The system cannot widen its own authority.

permission-to-act

Authority & controls

80

Recently added

Privacy concerns are holding up progress.

The operating question

What information would this use expose, to whom, for which purpose, and for how long?

A practical first step

Trace one proposed use through retrieval, model context, outputs and retained logs. Have the Knowledge owner and privacy specialists define permitted use and minimum data. Require the Technology owner to demonstrate the relevant access and retention controls.

Keep in view

Permission to retrieve information does not automatically permit disclosure, retention or reuse. Resolve the specific information path before widening the use.

privacy-and-progress

Authority & controls

70

November 18, 2026

Recently added

Different sources give different answers. Which can we trust?

The operating question

Which source has authority for this fact, policy, case and point in time?

A practical first step

Make the source, version, effective date and applicability visible at the decision. Define which owner resolves a conflict and how an approved correction reaches the next case. Keep the system of record explicit.

Keep in view

A confident explanation or plausible citation does not establish that the underlying fact is correct.

conflicting-sources

Knowledge & context

60

Recently added

We keep hearing that our data is not ready.

The operating question

Ready for which decision, with what consequences if the information is wrong or incomplete?

A practical first step

Choose one decision and identify its authoritative sources. Agree thresholds for completeness, freshness and accuracy; confirm permitted use; test representative exceptions. Name the Knowledge owner who can correct gaps and maintain the sources.

Keep in view

Information suitable for drafting an answer for expert review may be unsuitable for authorizing an action.

data-readiness

Knowledge & context

50

November 18, 2026

Recently added

We need to change the operation while keeping daily work running.

The operating question

What can change within a bounded trial while the service remains adequately staffed and supported?

A practical first step

Define the trial cohort and minimum service commitments. Have the Operating owner allocate real time for learning, exceptions and recovery. Agree pause conditions and test whether the fallback route can handle the expected volume.

Keep in view

A manual fallback is only credible when people have the capacity, access and authority to use it.

change-without-disruption

Work & decisions

40

Recently added

Individual tasks are faster, but the complete process is not.

The operating question

Where does the case still wait, repeat work or lose the context needed for a decision?

A practical first step

Trace one ordinary case and one exception from request to accepted outcome. Separate working time from waiting time, identify the remaining constraint, and agree receiving conditions with the next team before adding more automation.

Keep in view

A faster upstream task can increase the queue or correction workload downstream.

faster-tasks-same-delay

Work & decisions

30

Recently added

We need results now while preparing for what comes next.

The operating question

What improves with this investment, and which capabilities should the next change be able to reuse?

A practical first step

State the outcome that matters now and the future operating direction. Identify the shared knowledge, interfaces, controls and people each change needs. Fund a bounded first step with named dependencies and an explicit review point.

Keep in view

Track present benefits separately from future capability. A possible later benefit is not realized ROI.

results-and-direction

Value & outcomes

20

Recently added

We are saving time, but cannot demonstrate business value.

The operating question

Which operating decision turns released time into an outcome the business will accept?

A practical first step

Choose one cohort and reporting period. Ask the Outcome owner to commit the released capacity to a named service or workload change. Track accepted outcomes and the full recurring cost, with Finance challenging any financial claim.

Keep in view

Hours multiplied by salary are not automatically cash savings. Better service can be valuable without reducing spending.

capacity-to-value

Value & outcomes

10

Connect the dots

One outcome. A connected operating model.

Select a dimension to explore the decisions it introduces and the other parts of the operation it depends on.

Start with any dimension. Follow the decisions it connects.

Six dimensions. One shared outcome.Choose a card to explore
Selected dimensionKey connectionsAlso part of the model

Highlighted connections are a starting point. Every dimension contributes to the shared outcome; this is not a sequence of steps.

A RePerspective in practice

Make a choice. Follow its consequences.

A RePerspective examines one end-to-end outcome. This small example shows how an operating choice connects value, evidence, authority and people.

Illustrative operating scenario

Pay suppliers accurately and on time.

Routine invoices wait in an approval queue. Late payment can interrupt supply, while accounts payable spends time assembling context and chasing decisions.

Assumed starting point

Approved orders, receipt confirmations and supplier records are reliable for a defined routine cohort. Human review capacity is the main constraint.

Receipt confirmations for part of the routine cohort are now unreliable. The team can no longer treat a recorded match as sufficient evidence of receipt.

The wider outcome starts with the purchase need and ends with a fulfilled, correctly paid order. This decision concerns one part of that journey.

Change a condition:

Receipt evidence becomes unreliable.

Explore the initial assumptions, then change the condition to reconsider your choice.

The reasoning behind the choice

Under the starting conditions

Support the approver.

A useful starting point when leaders need to learn how reliable the evidence is while people retain approval authority.

Value and the trade-off

Reduce time spent finding and reconciling information. Approval queues may remain if review capacity is the main bottleneck.

Work, people and ownership

Accounts payable owns the case and approval queue. Buyers and receiving teams resolve missing evidence; reviewers need time to assess it.

Knowledge, authority and controls

Show the source behind each material fact. Separate verified evidence from inference and keep approval with the authorized person.

Evidence before expanding

Compare handling time, rework, queue age and payment accuracy against a baseline. Check whether the actual constraint has moved.

What would change this decision?

If evidence proves dependable but approval queues persist, test a bounded route for a defined cohort. If sources deteriorate, narrow what the assistant can assert.

When the condition changes

Support the approver.

With unreliable receipt evidence, assistive support can still help locate gaps. Reviewers must see the uncertainty and verify the receipt before approving the affected payment.

Value and the trade-off

Reduce time spent finding and reconciling information. Approval queues may remain if review capacity is the main bottleneck.

Work, people and ownership

Accounts payable owns the case and approval queue. Buyers and receiving teams resolve missing evidence; reviewers need time to assess it.

Knowledge, authority and controls

Treat unsupported receipt claims as unresolved. Route them to the receiving team and keep the evidence gap visible to the authorized approver.

Evidence before expanding

Compare handling time, rework, queue age and payment accuracy against a baseline. Check whether the actual constraint has moved.

What must happen next?

Reassess once receipt accuracy and completeness recover. A cleaner summary cannot substitute for verified evidence.

Under the starting conditions

Delegate within tested limits.

With reliable evidence and an approval bottleneck, this is a candidate to test for a tightly defined routine cohort. Readiness still has to be demonstrated.

Value and the trade-off

Remove routine waiting where the approved controls allow it. Exception handling, monitoring and support remain part of the full operating cost.

Work, people and ownership

The Outcome owner in Finance approves the operating boundaries. People own exceptions, intervention, reconciliation and any change in authority.

Knowledge, authority and controls

Require validated eligibility, trusted records, evidence of authorized action, independent checks and a way to halt the affected route. The system cannot grant itself authority.

Evidence before expanding

Test source completeness, incorrect releases, exception demand and on-time payment. Define thresholds and a stop rule before expanding.

What would change this decision?

If receipt evidence becomes unreliable, restrict the affected cohort and restore verified receipt handling before widening authority again.

When the condition changes

Delegate within tested limits.

The changed condition invalidates an assumption behind this route. Restrict the affected cases; human review also needs verified receipt evidence.

Value and the trade-off

Remove routine waiting where the approved controls allow it. Exception handling, monitoring and support remain part of the full operating cost.

Work, people and ownership

The Outcome owner in Finance narrows the operating boundary. Receiving validates the evidence; accounts payable manages the affected queue and supplier communication.

Knowledge, authority and controls

Prevent affected cases from proceeding on unreliable receipt records. Preserve case evidence, use the approved exception route and define the conditions for reopening.

Evidence before expanding

Test source completeness, incorrect releases, exception demand and on-time payment. Define thresholds and a stop rule before expanding.

What must happen next?

Reopen only after evidence quality and control performance meet the agreed criteria. Verify that exception-handling capacity can absorb the change.

Under the starting conditions

Repair the operating handoffs.

A strong choice when recurring evidence gaps or unclear ownership cause the delays. Under the initial assumptions, first test whether handoffs explain the queue.

Value and the trade-off

Address repeat exceptions and avoid chasing the same missing context. Coordination takes effort, and approval capacity may still need attention.

Work, people and ownership

Receiving owns receipt confirmation; procurement owns order quality; accounts payable owns payment coordination. Define who resolves each exception and when.

Knowledge, authority and controls

Make evidence requirements and escalation paths explicit. AI may help classify and assemble cases; receipt or approval facts must remain verifiable.

Evidence before expanding

Track repeat exceptions, time waiting on each owner, evidence completeness and end-to-end payment performance. Check for work merely shifted to another team.

What would change this decision?

If the information is dependable and the queue remains, revisit review capacity and appropriate delegation. If source evidence weakens, fixing ownership becomes more urgent.

When the condition changes

Repair the operating handoffs.

The changed condition strengthens the case for repairing the receipt handoff. Keep affected payments controlled while the evidence and ownership gap is resolved.

Value and the trade-off

Address repeat exceptions and avoid chasing the same missing context. Coordination takes effort, and approval capacity may still need attention.

Work, people and ownership

Receiving validates the facts and owns source correction. Procurement and accounts payable agree the exception route and keep suppliers informed.

Knowledge, authority and controls

Make evidence requirements and escalation paths explicit. AI may help classify and assemble cases; receipt or approval facts must remain verifiable.

Evidence before expanding

Track repeat exceptions, time waiting on each owner, evidence completeness and end-to-end payment performance. Check for work merely shifted to another team.

What must happen next?

Once dependable receipt handling returns, retest the whole journey. Then reconsider whether routine cases need less manual review.

Where this thinking goes next

Recognizable challenges. New operating possibilities.

Planned examples will examine a stakeholder outcome across teams, partners, decisions and controls. The procure-to-pay scenario above illustrates one decision within that wider journey.

Banking

Planned example

Commercial lending

What must connect to reach a timely, well-supported credit decision with clear accountability?

Insurance

Planned example

P&C claims

How could claims move from first notification to fair resolution while keeping customers, adjusters and partners aligned?

Pharma

Planned example

Clinical trial site activation

What would make readiness, dependencies and ownership visible as a study site moves toward activation?

Healthcare

Planned example

Referral-to-care coordination

How could a referral progress to appropriate care with fewer administrative delays and clearer handoffs?

Cross-industry operations

Planned example

Procure-to-pay

Follow the complete journey from a purchase need to fulfillment and correct payment. Connect supplier experience, operational continuity, evidence and decision authority.

Progress you can explain

The design improves when the evidence does.

Define the baseline, owner and decision threshold for each measure. Use the evidence to choose Expand, Continue, Narrow, Redesign or Stop.

Stakeholder outcomes

For supplier payment: timeliness, accuracy and the effort needed to resolve a problem.

Operating economics

Cost per accepted outcome includes recurring technology, review, correction, exceptions, support and recovery. Count failed and repeated work in the cost. Track how released capacity is used.

Control and reliability

Unsupported actions, recurring exceptions and the time needed to intervene or recover.

Ability to adapt

How quickly a changed condition is detected, assessed and addressed without losing control of the service.

The same evidence may reveal a reusable capability, a better service or a new opportunity worth testing.

Build a scorecard around your outcome →
From perspective to practice

What could change in your operation?

My approach connects the operating problem, design choices and evidence needed to shape a practical next step.

Explore approach