Update microsoft visio model

Will upload once assigned

To Be Option A

To be Option B

SECTION 7

To-Be Process Models Options A&B

Option A Operational Quick Wins & Waste Elimination

Option A focuses on fixing the obvious first, removing friction, eliminating duplicate effort and giving

staff the information they already need but currently can’t access in a timely manner. Most interventions

require minimal new infrastructure and can be delivered within 618 months using existing vendor

ecosystems.

Guiding principle: Stop doing things badly before trying to do new things intelligently.

This guidance is designed to support – not prescribe – your process modelling work. Each set of points gives

you a way of thinking about your To-Be design and a set of critical questions your model must be capable of

answering. The answers themselves are for you to work out.

Option A

Quick wins & waste elimination

Your To-Be model should show how fixing the obvious friction points – without rebuilding the

whole system – produces measurable improvements for patients and staff. Think: same

actors, better tools and better information.

PROCESS DESIGN GUIDANCE

1 Start with the handover chain

Every time information moves from one actor to another in the As-Is model, ask: is that

handover currently manual, verbal or paper-based? For each one that is, your To-Be model

should show what replaces it and what the receiving actor now has available that they did

not have before.

What does the triage nurse know the moment the patient arrives, versus what they know

now?

2 Target duplicate effort first

Look for every place in the As-Is model where the same information is entered, recorded or

communicated more than once. Your To-Be model should eliminate each instance of

duplication – and show clearly which actor is freed from that task and what they do with the

recovered time.

MAN6910 – Perth Health Emergency BPM Case Study 2025

Page 16

ECU Internal Information

MAN6910 BPM Case Study | Perth Health System Crisis

CONFIDENTIAL For Educational Use Only

Where does data exist in one place but get manually re-created somewhere else?

3 Make waiting visible and time-bounded

Every waiting step in the As-Is model is currently invisible – no one is automatically alerted

when a threshold is breached. In your To-Be model, identify which waits can be eliminated

entirely and which must remain but should trigger an automatic notification when they

exceed an acceptable limit.

Who needs to know what, and by when, to stop a wait from becoming a crisis?

4 Keep existing swim lane actors but change what they see

Option A does not restructure roles or introduce new organisational actors. The same

people do the same jobs – but with better information at the right moment. Your model

should reflect this: same swim lanes, but with improved data flows and eliminated manual

steps within each lane.

What information would each actor need at each step to make a faster, better decision?

5 Show the cascade effect of each change

Many improvements in Option A have knock-on effects upstream and downstream. Fixing

one bottleneck relieves pressure elsewhere. Your model and written analysis should

explicitly trace these cascades showing how an improvement in one swim lane reduces

waste in another.

If the bed status update delay is eliminated, what else improves as a result?

MODELLING REMINDERS

Annotate each changed process step with what specifically has changed and why.

Where a manual step is replaced by a digital one, show the new data flow explicitly.

Where a step is eliminated entirely, document it – removed steps are as important as new ones.

Use consistent notation to distinguish new steps from retained steps from eliminated steps.

Every claim about time or quality improvement must be traceable to a specific process change

in your model.

QUESTIONS YOUR MODEL MUST BE ABLE TO ANSWER

Which of the identifies issues in the Issue Register does this option address – and which

does it leave unresolved?

Be systematic. Work through the register item by item.

MAN6910 – Perth Health Emergency BPM Case Study 2025

Page 17

ECU Internal Information

MAN6910 BPM Case Study | Perth Health System Crisis CONFIDENTIAL For Educational Use Only

MAN6910 – Perth Health Emergency BPM Case Study 2025 Page 18

ECU Internal Information

For each improvement, what is the minimum technology or process change required to

deliver it?

Simpler solutions are more credible. Avoid over-engineering.

What does a shift handover look like in your To-Be model compared to the As-Is?

Trace it step by step. What information is transferred, how, and by whom?

How does the ramp wait time reduce – trace the causal chain step by step.

The answer must be traceable through your process model, not asserted.

Option B

Strategic AI adoption targeting root causes

Your To-Be model should show how AI changes the fundamental information architecture of

the system not just how individual tasks are done better, but how the system as a whole

becomes capable of seeing, anticipating and coordinating in ways that are structurally

impossible today.

PROCESS DESIGN GUIDANCE

1 Identify the root causes first

Before designing any process change, clearly articulate what the root causes of the current

crisis are, not its symptoms. Your To-Be model should be designed around resolving those

root causes. Every AI intervention you include should map directly to at least one of them.

What structural conditions make the current crisis unavoidable, regardless of how hard

individuals work?

2 Distinguish AI-assisted from AI-automated

Not every step in your To-Be model should be automated. For each AI intervention, be

explicit about whether a human remains in the decision loop and why. Some decisions carry

clinical or ethical risk that makes full automation inappropriate. Your model should show

where the boundary is and how it is enforced.

For each AI-supported decision, who retains authority and what happens if they override the

AI recommendation?

3 Redesign the information architecture, not just the tasks

Option B is not Option A with more technology. The fundamental change is that information

flows across the system in real time and reaches decision-makers before they need to ask

MAN6910 BPM Case Study | Perth Health System Crisis CONFIDENTIAL For Educational Use Only

MAN6910 – Perth Health Emergency BPM Case Study 2025 Page 19

ECU Internal Information

for it. Your model should show new data flows that do not exist in the As-Is model and

explain what decisions become possible as a result.

What decisions are currently impossible because information does not exist, is not shared, or

arrives too late?

4 Model the system across its full-time horizon

The scope of option B’s could operate ahead of real time, predicting demand, anticipating

bottlenecks, triggering actions before problems materialise. Your model could show how the

system behaves across different time horizons: 48 hours ahead, 4 hours ahead, and in the

moment. This temporal dimension is what separates strategic AI from operational

automation.

Which process steps in your model are reactive and which are proactive? What is the ratio,

and is it better than the As-Is?

5 Address the patient record problem explicitly

One of the most significant root causes in the As-Is model is that patient information does

not follow the patient across sites. Your To-Be model must show a specific solution to this

and be honest about its constraints. A full EPR replacement is a decade-long project. What

is a realistic, staged approach that delivers meaningful clinical benefit within your

implementation horizon?

What is the minimum viable patient record capability that would eliminate the most critical

information gaps in the As-Is model?

MODELLING REMINDERS

Option B may introduce new swim lane actors, AI platforms, centralised operations, community

care coordinators. Justify each new actor: what role do they play and why does the current

system have no equivalent?

Show where AI generates an output and a human acts on it, these are two distinct process

steps and must be modelled as such.

Do not assume the AI is always right. Your model should include exception handling: what

happens when AI output is wrong or unavailable?

Clearly distinguish which parts of Option B build on Option A foundations and which require

entirely new infrastructure.

QUESTIONS YOUR MODEL MUST BE ABLE TO ANSWER

Which root causes does this option resolve and which does it only partially address?

Be honest about limitations. No option resolves everything.

MAN6910 BPM Case Study | Perth Health System Crisis CONFIDENTIAL For Educational Use Only

MAN6910 – Perth Health Emergency BPM Case Study 2025 Page 20

ECU Internal Information

Where does AI create new risks that did not exist in the As-Is model? How are they

mitigated?

Consider: algorithmic bias, system failure, data privacy, deskilling.

How does the system respond to a 30% surge in demand under Option B trace the

process from prediction to resource deployment.

This tests whether your model is genuinely predictive or just reactive with better tools.

What does Option B require that Option A does not – in terms of data, infrastructure,

governance and culture?

Honest assessment of requirements is more persuasive to executives than optimism.

Guidance that applies to both options

Anchor every claim in the

As-Is

Your As-Is model is your

baseline. Every improvement

claim in your To-Be must

reference a specific problem

identified in the As-Is not a

generic aspiration about what

technology can do.

Separate process from

technology

Be precise about which

changes are process redesign,

which are technology

enablement, and which require

both. Conflating the two

produces models that are hard

to evaluate and harder to

implement.

Account for the resistant

stakeholder

Consider the stakeholder most

likely to resist each change.

How does your model account

for their concerns? A model

that ignores implementation

reality is not a credible model.

WRITE MY PAPER