Since Inverbis was just a proto-company, our thinking around process mining has focused on understanding how data should be captured and what data models best support each process—so we can automate insights for improvement.
Let us explain: process mining inevitably relies on a basic data package (without it, there is no mining). This typically includes the case ID, the description of the executed activity, and a timestamp. But there’s an often-overlooked area that adds real depth to the analysis: attributes—which are optional but essential.
What we call attributes includes all the elements that allow us to explore root causes of deviations and isolate specific problems within a process.
In every project we lead, we place strong emphasis on building this layer in collaboration with our clients. And it leads to fascinating discoveries.
But… it’s not enough.
And it’s not enough for the discipline of process intelligence as a whole. In an ideal (but not impossible) world, a well-defined breakdown of activities and a rich attribute model would allow us to diagnose process issues in real time—and understand their relationships with upstream or downstream processes. That could be built into system design. And system design can be changed.
This is still an emerging discipline. Some organizations have robust, high-quality data. Others, not so much. But none of them originally designed their data models or capture systems with mining in mind.
That’s why we decided long ago that one of our R&D priorities must be to make it easier to capture full traces (not just scattered data points) and to develop attribute data models specifically designed for continuous improvement.
Recently, we reached an agreement with Cerner to become a solution partner for ISH-Med, the “SAP for healthcare organizations.” Its architecture already includes many process-based elements—but we want to go further. We’re working on defining recommended attribute models for analyzing hospital processes, so we can offer them by default to our clients. We’re also advancing a separate workstream on how to generate those attributes effectively.
So we ask: Anyone out there with ideas they’d like to share?
We’re listening—and grateful.
Check some of our videos:
- New ➽ ICU discharges: where time slips away
- Fast Track vs Standard Path: Which is Better for Diagnosing Colorectal Cancer
- What causes discards in parenteral nutrition preparation? Key factors revealed!
- How Process Mining Uncovers Medication Administration Issues
The Prompter.io is our open project to share our experience—and that of others—in integrating language models and data-to-text techniques into process intelligence. Don’t miss it!
