The background

IBM has always offered on the market, within its IBM Process Mining application, a distinctive feature called Multilevel Process Mining, capable of generating models and statistics relating to complex business processes involving heterogeneous assets, such as Procure-To-Pay (P2P) or Order-To-Cash (O2C).

As the entire ecosystem evolves, requiring the integration of ever-increasing amount of data and processing speed, the multilevel mining algorithm has highlighted its limitations, hence the need to understand whether to invest in its evolution or move towards alternative models.

Outside of IBM, the academic world of process mining has long embraced a different model called Object-Centric Process Mining, which has now become the de facto standard in the field and is followed by the majority of our competitors.

IBM decided to evolve its process mining product completing a first concrete and consistent step in this direction.

What did we achieve?

A first version of the new approach was included within the IBM Process Mining product included in version 2.0, a major release with new features that will bring benefits and simplifications for users. 

Now each process is no longer composed only of the event log, but the events are supported by an object structure capable of bringing multiple advantages:

• flexibility in the definition of the schema as the information no longer necessarily has to be spread across the event log but can live in structures (objects) that are separate but connected to each other
• simplification of data preparation, as the data structure of the source supports is preserved
• much more powerful analytics thanks to the flexibility of the underlying E/R model

What is next?

Try the 2.0.2 release of IBM Process Mining, GA available from June 2025. You’ll be able to play with interconnected processes benefiting from simplified data preparation and multi-process aggregate statistics.