Success is the day the central team can step back and nothing falls over.
Most programmes measure success by their own activity. This one should measure it by how little the line needs the centre. From day one, the central team is designed as a diminishing function — building local capability, transferring ownership, and stepping back into a coaching role as directorates gain confidence. The external partner's role becomes explicitly finite, with capability transfer as a contracted outcome, handled with respect for the depth they've brought.
The aim isn't a permanent programme office. It's the line's capability to run this without one — with the central function designed to shrink as that capability grows.
Diagnose, then shift the unit of change, then transfer ownership, then augment.
Each phase opens to its detail. The sequence matters: we earn each phase by completing the one before. Select a phase to expand.
Diagnose traction, not activity
Before adding a single tool, establish the truth: where is the operating model genuinely live, and where is it theatre? The toolkit being in place tells us about adoption. It tells us nothing about embeddedness. This phase closes that gap and gives the executive sponsor an honest baseline.
- Run a maturity diagnostic across directorates — rating whether huddles resolve issues, whether frontline ideas get implemented, whether leaders coach or direct
- Distinguish "the board exists" from "the board works" in every area
- Listen at the source — gemba, not reports — and read the engagement and safety-culture signals as live data
- Map the current external-support footprint and where dependency sits
- An honest traction baseline — green-on-the-surface vs. genuinely embedded
- A clear view of where leaders have changed and where they haven't
- The real dependency map: what would stop if external support paused tomorrow
- A focused plan, not a generic scale-up
Shift the unit of change to the leader
The largest failure mode is leaders not changing. So the core work of this phase is behavioural, not mechanical: install leader standard work, coaching-based problem-solving, and going-to-the-source as the actual content of how leaders lead — and make those behaviours the ones the organisation develops and reinforces.
- Define and embed leader standard work — the daily disciplines beneath the rituals
- Coach senior and middle leaders through the shift from directing to developing
- Fix or retire any huddle or scorecard that generates motion without resolved problems
- Protect the framing relentlessly — anchor everything to safety, quality and respect for people; manage the cost-cutting misread before it starts
- Leaders who lead differently — visible, coaching, at the work
- Rituals that resolve problems rather than report on them
- Early, defensible wins tied to patient flow and safety
- Union and workforce trust actively maintained, not assumed
Transfer ownership into the line
This is the phase the whole role exists for. Capability moves out of the central team and external partner and into local leaders and embedded facilitators. The centre deliberately shrinks. The method gets wired into the performance lifecycle so it is reinforced by how the organisation hires, expects, reviews and promotes — not by central push.
- Build local capability through train-the-trainer and embedded "navigators" in each directorate
- Wire the programme behaviours into leader expectations, capability profiles, appraisal and promotion
- Step the central team back from doing to coaching; convert external support to a finite, capability-transfer engagement
- Connect — not duplicate — what already exists across the People function (see section C)
- Directorates running improvement without central prompting
- A defined, dignified off-ramp for external spend
- The method embedded in the performance lifecycle, so it can't be switched off
- A central function that is smaller than the year before — by design
Augment with AI — once the human system is real
Only now does technology earn its place. With a healthy improvement culture established, AI compounds it — and because the culture is healthy, the technology accelerates rather than automating broken processes. Governance, transparency and human judgement lead; the tool follows.
- Introduce live dashboards that feed huddles automatically, ending manual data prep
- Use pattern-finding across patient and staff feedback to surface root causes faster
- Pilot predictive flow so teams act before bottlenecks form
- Hold firm governance — bias, transparency, human oversight, respect for people preserved
- A continuously-learning operating model, not a periodic exercise
- A genuine forward story for the CEO's flagship
- Augmentation that strengthens the human core rather than eroding it
- A defensible, governed approach the board can stand behind
Connect what an organisation already has. Build only what's genuinely missing.
A common error is to stand up a parallel programme that competes with existing People & Culture activity. The better move is the opposite — find the capability already present across the function and wire the programme through it, so the method amplifies existing investment rather than duplicating it.
Existing leadership development
Make leader standard work and coaching the live content of current leadership programmes — not a separate track competing for the same leaders' time.
Existing engagement & culture data
Repurpose engagement and safety-culture instruments already in place as the live measure of whether the operating model is real in each area.
Existing performance & capability frameworks
Thread the programme behaviours into the appraisal, capability and promotion architecture that already exists, rather than adding a new layer.
Existing change & quality teams
Partner with quality, safety and improvement functions already operating — align language and method so the organisation speaks one improvement dialect.
The external partner
Reposition deep external expertise toward capability transfer and the hardest technical problems — a finite, valued role with a clear endpoint.
Clinical leadership
Equip clinical leaders to own the model on their own turf, supplying the floor credibility and clinical legitimacy the central function should never claim for itself.
Measure traction and ownership — not the volume of programme activity.
Care is measurably safer and flows better
Shorter delays and smoother handovers, faster access to emergency and mental health care, fewer safety incidents, problems resolved closer to the source.
Staff feel safe to surface problems
Engagement and safety-culture signals rising in the areas where the model is embedding; frontline improvement ideas implemented, not just submitted.
The line needs the centre less each quarter
Directorates running their own rhythm and problem-solving without central prompting; external dependency falling on a planned curve.
Leaders lead the new way
Leader standard work observed in practice; coaching and developmental behaviours rising in 360° and on-the-ground evidence, not just self-report.