Process Mining is the “hot thing” for a couple of years now, but it feels as if the momentum has stalled a bit and the focus is now on rolling it out into organizations—Gartner classified it as “early mainstream” last year and from my experience when talking about it (and I do a show of hands at the beginning of each talk) that seems to be right when only about 20% confirm that they actually do process mining projects in their organizations.
But there appears to be a new twist and evolution coming to process mining—Causal Process Mining. An approach where you don’t focus exclusively on the event sequence based on timestamps but analyze the process on the cause and effect relations in the activities. The idea is to not only to discover the “what is happening” but the “why it is happening”—a very useful perspective for any business analyst.


Lukas and Jan are the co-founders of Noreja, a company that creates a causal process mining tool that gives not only a new spin on the traditional process mining (be it case-centric or object-centric), but also uses AI to get the insights into the process with different perspectives and thereby eliminates the need to handcraft event logs but rather adds more context information to the analysis, such as documents.
Jan is a professor at the Humboldt University in Berlin (and before that in Vienna) and was also co-founder and chair of the Berliner BPM Initiative and started out as a Java developer and risk analyst at a Big 4 company way-back-when.
Lukas started at the German postal services as a business analyst and IT system designer, moved on to a (different) Big 4 company before going into academia and worked with Jan before they—and their third co-founder, Philipp—founded Noreja where his current role is the CEO.
In this episode of the podcast, we talk about:
- The Evolution of Process Mining: Traditional process mining began in the 1990s and grew around 2000 with Wil van der Aalst’s integration of Petri nets and concurrency concepts, effectively handling simple, linear processes with single objects.
- The Complexity Gap: Classical techniques struggle with complex enterprise processes (like order-to-cash) where initial objects converge, diverge, and break up into line items, production orders, or separate warehouse deliveries.
- Moving Beyond Correlation: While standard process mining looks at what happens sequentially (often mistaking correlation for causation), a causal approach uses a priori information to validate actual triggers against observed data.
- Multi-Source Causal Modeling: Norea builds its foundational causal models using three primary sources: database schemata (foreign/primary keys and cardinalities), AI/LLMs to validate logical sequences, and organizational domain knowledge.
- Skipping the Flat-File Headache: Rejecting traditional CSV and event log exports prevents companies from discarding critical database semantics and spending excessive time on complex ETL data-cleaning pipelines.
- Streamlined Technical Onboarding: By connecting directly to source systems via databases or APIs, users can select core tables (e.g., 20 standard SAP tables for order-to-cash) to automatically generate an initial entity graph.
- Perspectives Over Rigid Views: Users can define different “perspectives” by picking specific data objects or targets as starting points, allowing AI agents to analyze scoped variations and pinpoint root causes efficiently.
- Lowering Barriers to Entry: Instead of getting bogged down by perceived bad data, companies can use the tool to instantly visualize where data quality issues reside in the process and measure their exact impact.
- AI Context and Token Efficiency: Modern enterprise context windows are too large for standard LLMs; structuring processes through graph databases and causal paths helps AI consume tokens intelligently and avoid messy “spaghetti diagrams”.
- Maintaining Dynamic Context: To prevent context from becoming outdated, organizations can connect operating procedures and procedural versions to the analysis tool to verify compliance with newly introduced rules.
- Process Frontier Agents: Norea is developing autonomous background agents (such as context managers) to scan SharePoint repositories, archive outdated documentation, and streamline organizational knowledge.
- Cost-Efficient Local LLMs: Leveraging local open-weight models (like Gemma) for heavy background tasks reduces token API costs, utilizing off-hours computing power effectively.
- Compressing the BPM Lifecycle: By combining AI-driven insights, automated requirement definition, and rapid artifact generation, the goal is to shrink the traditional 12-month BPM lifecycle down into a concentrated 2-to-3-day workshop.
- Proving ROI on the Balance Sheet: Historical SAP process improvement research demonstrates an average return ratio of 1 to 20 on optimization projects, which accelerates drastically when decision velocity and accuracy increase.
- Next Steps & Resources: Listeners can explore Norea’s whitepapers, blog posts, and request a demo environment by visiting their official website or connecting with Jan and Lukas via LinkedIn.
Lukas and Jan can be reached via Linked in and their website: https://www.linkedin.com/in/lukas-pfahlsberger/ and https://www.linkedin.com/in/janmendling/.
Please reach out to us by either sending an email to hello@whatsyourbaseline.com or signing up for our newsletter and reading articles about process and architecture on our Substack… Go and subscribe at whatsyourbaseline.substack.com.
And if you like to support “the little podcast that could,” become a Patron at https://www.patreon.com/c/whatsyourbaseline. We appreciate you!
Additional information
- Here is the White Paper that Lukas mentioned in the episode: https://144242473.fs1.hubspotusercontent-eu1.net/hubfs/144242473/website_files/Noreja%20Causal%20Hypothesis%20Whitepaper.pdf
- The graphic below shows the evolution of process mining

Credits
Music by Jeremy Voltz, www.jeremyvoltzmusic.com
Roland Woldt has spent nearly 30 years helping organizations cut through complexity — transforming how they think about their architecture, processes, and strategy. A former officer in the German Armed Forces, he has held executive and leadership roles at Accenture, Software AG/IDS Scheer, KPMG, and iGrafx, leading engagements across the full project lifecycle from solution design to rollout.
Roland is the author of Successful Architecture Implementation and Successful Process Mining Projects, and co-hosts the podcast What’s Your Baseline?, where he demystifies Enterprise Architecture and BPM for practitioners worldwide. Learn more at woldt.de.
