JULY 3, 2024

Every week, Artificial Intelligence (AI) makes headlines in both tech and mainstream news. With all the buzz, it can be challenging to distinguish between hype and reality. At Flowable, we remain focused and pragmatic about integrating AI and other technologies.

If you’ve attended our webinars, participated in FlowFest, or read our articles, you know we've been closely monitoring AI developments. Since ChatGPT's public launch, we’ve been experimenting, building prototypes, and sharing our insights with you.

In the past few months, this has culminated into a clear understanding of the strengths and weaknesses of the Generative AI (GenAI) technology; and where it makes sense to integrate with it and – perhaps more important – where it doesn’t make sense.

GenAI excels in use cases such as:
  • Content Creation: Generating initial content based on prompts.

  • Enhancement: Improving or reworking existing material with specific styles.

  • Summarization: Quickly summarizing large datasets.

  • Natural Language Queries: Asking specific questions using natural language as ‘query mechanism’.

With these strengths in mind, we’ve focused on the following areas in the Flowable product line:

The empty canvas

Starting out with a new app, process, case, form or any other model can be a daunting task. With GenAI integration in Flowable Design, AI can provide initial suggestions, helping you jumpstart your project and spark creativity.

Contextual model generation

Getting going is one thing, but the real heavy lifting is when you already have existing models and you want to build new models that utilize what is already there. By automatically providing all of this context to the GenAI system in Flowable Design, creating such integrated and cohesive models becomes a breeze.

Integrate AI functionality directly in processes and cases

Flowable is essence is a platform for building applications. As such, it’s impossible to know or even predict what customers build and will build in the future. To tap into this seemingly infinite creativity, we’ve added the AI task to both the BPMN and CMMN palette, which can be used to have the AI system execute tasks during runtime execution.

Besides the AI task, we’ve also added a new type for the service models: the AI service. When using this new service type, it’s possible to fully encapsulate the AI behavior into a reusable component for modeling users that don’t have to worry about the nitty-gritty implementation details (that are needed when using the AI task).


In many real-life automations, documents play a crucial role in sharing of information or decision making. So, it comes with no surprise we made documents a first-class citizen when it comes to AI. Right in the Flowable Work UI, it’s possible to summarize, translate, execute sentiment analysis or simply ask any question in the context of the documents.

Conversational orchestrations

Speaking of asking questions, in Flowable Work it’ll be possible to ask directed questions within the context of a process or case. Behind the scenes, we automatically pass relevant information such as instance variables, people, historical information, model structure and execution history to the AI system.

With this context and the default large language models, it’s now possible to use natural language to explain exactly what piece of information is needed. Or, for example, ask for a short update of where things are. Instead of having to click and search through various UI’s to find an answer (taking into account not all information is always shown), just ask the question and get a reply with the result immediately.

Technically this is built using the typical Flowable architectural patterns. The AI module is opt-in both in front- and backend. This module contains the context creators which are used during prompt creation and enhancing.

Important of note here is that all of this is implemented in a vendor-neutral way, where we’ve added programming interfaces and relevant abstractions where needed. As such, switching between OpenAI’s ChatGPT, AWS Bedrock, Llama from Meta, MistralAI and many others becomes a matter of configuration.

All the above was demonstrated live during our recent webinar “Smart Modeling: Leveraging AI Assistance”. If you missed it and want to rewatch it, find it here.

Looking forward, this is but a first step in a most likely long journey. The usage of GenAI is surely exciting, but it’s important to keep in mind that GenAI is only one type of AI. There are many other AI technologies that could complement Flowable, and we’ll keep you updated on these developments.

Joram Barrez_MG 7807

Joram Barrez

Principal Software Architect

A core developer at Flowable with over a decade of experience in open source software and of building scalable process engines. He co-founded the Activiti project (on which Flowable is based) and was part of the JBoss jBPM team before that.

Share this Blog post
Engineering | FEBRUARY 19, 2024
The Value of AI in Modeling

As AI gains prominence as a pivotal technology and enterprises increasingly seek to leverage its capabilities, we are actively exploring diverse avenues for integrating AI into process automation.

Engineering | OCTOBER 3, 2023
Low-code, High Impact: Flowable & CMMN for Complex Use Cases

The key to managing complexity is to combine different and multiple tools leads to better, faster, and more maintainable solutions. For example, combining BPMN with CMMN.

Engineering | OCTOBER 2, 2023
The New Flowable eLearning Platform

Discover the reasons behind our brand-new Flowable eLearning platform and explore its features by registering for our inaugural free course.