Hello, I need to convert my BPMN process models inside my project into code that can be converted and transformed by my ML model, this model is going to be used along with other data from different sources.
These process models will work as building blocks for my ML model.
Has anyone worked with this approach?
Thank you!
Hi @christianbg92, welcome to the forum! Interesting question! What kind of “code” do you mean? The models are already XML that adhere to the BPMN spec from OMG; Camunda uses the XML to train our own AI models.
Hello!
Is it possible to extract the code into CSV or JSON, or another type compatible with Google Cloud or n8n models.
Or in another case, is it possible to train an AI Agent model from Camunda and not use other platforms?
Let’s say this model is for automating simple requests from Customer Service
Thanks!
@christianbg92 - BPMN models are not code, so not really. You could maybe find an XML to JSON convertor that gets you something but you will lose much of the semantic meaning and the structure of the models.
Got it!
So overall, you wouldn´t recommend this approach to train AI agents?
@christianbg92 - lots of people have trained models on BPMN. It really depends a lot on what your goals with the model are, what the capabilities and knowledge of the base model is (if you’re using one), and how you’re training it. I’m not a data scientist so I can’t help much beyond that, though.