When I started looking into the different governance reports we need for daily operations, I wanted to figure out how many distinct report types actually exist and how to build them efficiently without repeating myself.
To test this, I looked at two common use cases:
- Attribute Integrity Checks: Verifying if mandatory fields (like descriptions) are filled in.
- Regulatory & GxP Compliance Checks: Verifying if high-risk process steps actually have linked Risk or Control objects.
At first, these look like entirely different reports. One checks a simple text property; the other checks relationships across connected objects.
But as I started writing the logic for both, I realized I was re-writing the same setup code every single time—grabbing selected models, setting up output files, looping through canvas symbols, and exporting results.
That's when I realized the pipeline itself never changes. Only the rule evaluation changes.
So I built a simple 5-Step Template:
- Step 1: Capture & Guard: Grab selected models from the canvas.
- Step 2: Initialize Output: Set up table styles and headers.
- Step 3: Canvas Traversal: Map visual symbols back to central repository objects.
- Step 4: Governance Assertion: Evaluate the specific rule.
- Step 5: Export: Build the final report output.
Steps 1–3 and Step 5 are fixed pipeline steps. Step 4 is the only piece you swap out.
How the logic changes at Step 4:
- For Attribute Audits: Step 4 looks at a single object and asks: "Is this attribute filled in?" If yes, pass. If empty, flag a documentation gap.
- For Compliance Audits: Step 4 evaluates relationships across nodes. It asks: "Is this task marked GxP-relevant? If so, does a connection exist to a Control object?" If mapped, pass. If unmapped, flag a compliance issue.
By separating the setup pipeline from the rule evaluation, turning an attribute check into a compliance audit takes minutes instead of writing a new script from scratch.
For anyone managing large process architectures, this keeps your script library clean and makes writing new governance checks much faster.
How do you handle reusable patterns in your ARIS scripts? Curious to hear what approach others take!