Understanding Business Process Continuums

In a recent engagement with a healthcare provider, we observed the same case management scenario modeled in three different ways—each using a different notation, but only one captured the actual behavior of frontline staff. That discrepancy wasn’t due to poor execution. It was a symptom of misaligned modeling tools.

Every process isn’t born equal. Some are predictable, repeatable, and rule-based. Others are dynamic, knowledge-intensive, and shaped by ongoing discovery. The key insight? There’s no single best tool for every kind of work. Instead, we operate on a continuum.

This chapter introduces the process modeling continuum, a practical framework that helps you place any business activity along a spectrum—from highly structured to fully adaptive. By understanding where your process falls, you avoid forcing workflows into rigid structures or under-designing complex cases.

You’ll learn how to identify whether a process leans toward structured vs unstructured processes, when to use BPMN, when to reach for CMMN, and how to blend both for true business agility. This is not theory—it’s what I’ve refined over 20 years of modeling in banks, insurers, and public services.

Mapping the Process Modeling Continuum

Imagine a horizontal line. On the left: workflows that are repeatable, predictable, and governed by fixed rules. On the right: cases built on discovery, intuition, and evolving knowledge. The space between them? A spectrum of complexity where most real-world processes live.

This continuum isn’t just conceptual. It governs not just how we model, but how we execute, measure, and evolve processes.

Left Side: Structured, Predictable, Rule-Driven Work

These are workflows where the steps, sequence, and outcome are known in advance. Think invoice processing, onboarding new employees, or manufacturing assembly lines.

They thrive under BPMN. The notation excels when you can define:

• Clear start and end events
• Fixed decision points (e.g., “Is the document signed?”)
• Repeatable task sequences
• Automated transitions based on data and conditions

When the path is fixed and the data is consistent, BPMN becomes a powerful blueprint for automation and compliance.

Right Side: Adaptive, Knowledge-Driven, Discovery-Based Work

Here, the steps aren’t known upfront. The outcome depends on investigation, judgment, and new information. Examples include insurance claims, clinical diagnostics, or fraud investigations.

These scenarios demand CMMN. It’s built for situations where:

• Tasks are triggered by events, not sequence
• Staff decide what to do next based on real-time evidence
• Tasks may be revisited, reordered, or skipped
• Case files evolve as more data emerges

CMMN doesn’t assume the path. It enables it.

Center: The Gray Zone – Where Hybrid Modeling Begins

Most business activities live somewhere in the middle. They have structure—but also room for adaptation.

Consider a customer complaint resolution process:

• Standard steps: log, acknowledge, assign, respond.
• But if the complaint reveals a product defect, new tasks like “notify engineering” or “initiate recall review” may emerge.
• These aren’t part of the original plan—they’re discovered during execution.

This is where the continuum becomes actionable. The model must reflect not just what’s known, but what might be.

Using the Continuum: A Practical Decision Framework

When evaluating a process, ask these questions—each pointing you toward the right notation.

Ask: Is the path predictable and repeatable?

• Yes → Consider BPMN. The process benefits from clear control flow, automated routing, and measurable performance metrics.
• No → Consider CMMN. The process likely requires dynamic task assignment and event-driven progression.

Ask: Is the decision-making knowledge-intensive?

• Yes → CMMN wins. Frontline staff are the decision-makers, not a static workflow.
• No → BPMN fits better. Decisions are based on data and rules, often automated.

Ask: Can the outcome change based on new information?

• Yes → Lean toward CMMN. The case must adapt as facts emerge.
• No → BPMN is suitable. The outcome is fixed by design.

This isn’t a binary choice. It’s a spectrum-based alignment. The goal isn’t to pick one model over the other—but to match the tool to the work.

Real-World Examples: From Continuum to Model

Let’s walk through how the same scenario shifts across the spectrum.

Example 1: Loan Approval

At a bank, a loan application starts with structured steps: verify income, check credit score, assess debt-to-income ratio. These are BPMN territory.

But if credit is low, the process branches into a deeper investigation: “Request additional documentation,” “Contact employer,” “Review employment history.” These are triggered not by sequence, but by events—e.g., “When credit score < 600.”

Here, the process evolves into a case. The initial BPMN model becomes a high-level orchestration—a CMMN case plan with subtasks like “Investigate credit risk” and “Schedule customer meeting.”

This hybrid model reflects reality: structured entry, adaptive execution.

Example 2: Patient Diagnosis

A doctor sees a patient with unexplained symptoms. The diagnosis path isn’t fixed. It’s driven by:

• Lab results
• Imaging scans
• Physical exams
• Emerging hypotheses

No two patients follow the same path. This is classic CMMN use: a case plan with stages like “Initial Assessment,” “Diagnostic Testing,” “Symptom Review,” and “Final Diagnosis.” Tasks appear based on evidence, not pre-defined sequence.

Even if the doctor uses a checklist (a structured tool), the decision-making is adaptive. The model must reflect the uncertainty, not pretend it’s known.

Choosing Your Tool: Beyond the Binary

Many still see BPMN and CMMN as competitors. But they’re not. They’re complementary.

Think of BPMN as the choreographer of predictable dances. CMMN is the jazz improviser—structured in form, free in expression.

When you model the full lifecycle of a customer incident:

• Use BPMN for the standard routing: assign, escalate, update status.
• Use CMMN for the investigation: tasks like “Analyze logs,” “Interview user,” “Verify system impact” appear dynamically based on detected events.

One controls the flow. The other adapts to discovery. Both are needed.

Guidelines for Hybrid Modeling

When blending BPMN and CMMN, follow these principles:

1. Use BPMN for orchestration: control the high-level process, such as handling, escalation, and routing.
2. Use CMMN for case execution: model the investigative, adaptive core where decisions emerge during work.
3. Link them via subprocesses: embed a CMMN case as a subprocess within a BPMN process, or call a BPMN sub-process from a CMMN stage.
4. Keep boundaries clear: avoid mixing rules and adaptability in one diagram. Let each notation do what it does best.
5. Use shared data: ensure both models reference the same case files and data objects to maintain consistency.

These aren’t rules—just habits I’ve seen successful teams adopt. The goal is clarity, not complexity.

Frequently Asked Questions

What is the process modeling continuum?

It’s a spectrum that ranges from highly structured, predictable workflows (BPMN) to adaptive, knowledge-driven processes (CMMN). Most work falls in between, requiring hybrid modeling.

When should I use BPMN over CMMN?

Use BPMN when the process path is fixed, repeatable, and governed by rules. Ideal for approvals, manufacturing, and compliance workflows.

When are adaptive processes best modeled with CMMN?

CMMN excels for cases where the next step depends on discovery, judgment, or events. Examples: insurance claims, fraud investigations, or clinical diagnoses.

Can one model use both BPMN and CMMN?

Absolutely. In practice, a BPMN process can include a CMMN subprocess for the investigative phase—like handling a customer incident. The BPMN manages routing; the CMMN enables adaptability.

How do I avoid overusing CMMN?

Ask: “Is this truly adaptive?” If the steps are fixed and repeatable, CMMN introduces unnecessary complexity. Use it only when flexibility is essential.

Should I use BPMN for all predictable processes?

Yes, but only if the process is truly predictable. If it evolves with new information—e.g., contract negotiation—consider a hybrid approach. Over-structuring leads to brittle models.

Do I need to know both BPMN and CMMN to be effective?

Yes. The most effective analysts think in both. Understanding the continuum helps you choose the right tool—without forcing a solution into a wrong language.