IMIXS-WORKFLOW Blog

With the latest version of Imixs-AI we just shipped a new feature extending BPMN with AI: Tool Calling support for the OpenAI API. This is a powerfull new feature introducing a fundamental shift in what an LLM can do inside a workflow application.

From answering questions to starting business processes

Until now, an LLM integrated into Imixs could analyze data, generate text, evaluate conditions, and assist users with information. All of that is great. But the LLM was always just talking. It could tell you “you should start a vacation request” — but it couldn’t actually do it. Tool Calling changes this completely.

With Tool Calling, the LLM no longer just responds with text. It can respond with a structured request to execute a function in your application. Consider this scenario:

A user types into the workflow application:

"I need next week off."

Imixs-AI now doesn’t explain how to submit a vacation request. It doesn’t ask for clarification. It fires a tool call:

{
  "finish_reason": "tool_calls",
  "tool_calls": [{
    "function": {
      "name": "load_skill",
      "arguments": "{\"process_id\": \"urlaubsantrag\"}"
    }
  }]
}

Now a business application based on Imixs-AI is able to handle this respons and start the correct BPMN process — all from a single natural language sentence. No menus, no navigation, no searching. Just say what you want and it happens.

Continue reading “Imixs-AI – LLM Tool Calling”

Theia IDE is a modern, AI-native IDE which can be run on Desktop as also in a Docker container as a Web app. With the Docker image you can deploy Theia IDE in a Kubernetes cluster. The deployment is straightforward — until you try to use extensions that open embedded browser windows. This post explains the problem and how to solve it using wildcard certificates with DNS-01 challenges.

Continue reading “Running Theia IDE in Kubernetes”

If you work on modern enterprise IT systems, “AI Agents” are the new elephant in the room. Every day, new YouTube videos appear where someone talks enthusiastically about AI Agents and Agentic AI. The base line – software systems not using AI Agents are outdated and behind the times. But what does it really mean to integrate an AI agent into critical business applications?

Imixs-Workflow is an open-source project for building transactional, secure, and transparent business applications based on the BPMN 2.0 modeling standard. The idea is simple: you describe a business process in a model, then use a workflow engine to execute that model. The concept of a workflow engine is not new, but it remains a well-grounded approach to executing predefined business processes in a transparent, comprehensible, and transactional context. This simply means that the workflow engine guarantees a task is carried out exactly as previously defined. Execution can be performed by both IT systems and humans.

Let’s have a look how this concept is related to AI Agents.

Continue reading “AI Agents vs. AI-Augmented Workflows”

With our latest version of Imixs Workflow, we introduce a game-changing approach to business process automation, enabling domain-specific handling of BPMN 2.0 conditions.

The Challenge

In previous versions, conditional sequence flows in the Imixs Workflow engine —as in most traditional BPMN engines — could only be defined using script functions. While this scripting-based approach offered flexibility, it also introduced several limitations:

• Domain-specific languages could not be supported
• Conditions were tightly coupled to technical implementation
• Adding new evaluation logic required modifying or extending engine code
• Complex business rules were difficult to express and maintain

These constraints made it challenging to adapt and extend business logic across different domains.

The Solution: CDI Event-Driven Condition Evaluation

Imixs Workflow now introduces a revolutionary and extensible architecture for condition evaluation. Leveraging Jakarta EE’s CDI (Contexts and Dependency Injection) events, conditions can be transformed and interpreted by specialized observers before evaluation — all without modifying the core engine.

This new approach not only supports functional scripting, but also enables fully domain-specific evaluation of BPMN 2.0 conditions, such as SQL-based queries, external API calls, or even AI-driven logic.

How It Works

When a sequence flow condition is encountered, the system fires a CDI event that any registered observer can intercept and handle. Observers can:

• Translate domain-specific languages into executable code
• Implement business rule languages tailored to specific domains
• Add validation, monitoring, and auditing layers
• Integrate external services and rule engines

Real-World Example: Domain-Specific Conditions

Instead of writing complex JavaScript like this :

// Before: Hard-coded and difficult to maintain
!['AT', 'BE', 'BG', 'HR', ...].includes(workitem.getItemValueString('country'))

You can now define business rules in a domain-specific language:

SQL: SELECT COUNT(*) FROM eu_countries WHERE country_code = ? > 0

A specialized domain observer intercepts this rule and translates it into the appropriate JavaScript. The beauty? The RuleEngine doesn’t need to know anything about domain-specific languages — it simply evaluates the resulting JavaScript as it always has.

Implementation Example

The new event-driven design of Imixs Workflow is built on Jakarta EE’s CDI (Contexts and Dependency Injection) event mechanism — a core implementation of the Observer Pattern. This means that custom components can observe and react to condition evaluation events without depending on the workflow engine itself.

Creating your own domain-specific observer is straightforward. Simply extend the new ConditionalExpressionEvent class and implement your logic as a CDI observer method:

@ApplicationScoped
public class DepartmentSQLRule {
    
    public void onConditionEvaluation(@Observes ConditionalExpressionEvent event) {
        String condition = event.getCondition();
        
        if (condition.startsWith("SELECT ")) {
            // Transform domain-specific rule into JavaScript
            String javascript = translateSQLRule(condition, event.getWorkitem());
            event.setCondition(javascript);
        }
    }
}

In this example, the observer listens for every ConditionalExpressionEvent. If the condition starts with an SQL statement, it automatically transforms the domain-specific rule into executable JavaScript. Because this is handled entirely via CDI events, the core RuleEngine remains untouched — ensuring maximum flexibility and extensibility.

Opening Doors for Enterprise Development

This new observer-based architecture fundamentally changes how enterprise workflow systems are developed and evolved. By decoupling condition evaluation from the core engine, organizations gain unprecedented flexibility in how business rules are expressed, validated, and executed. Teams can now create tailored domain-specific languages that speak the language of their business domain—whether that’s banking regulations, insurance underwriting, supply chain logistics, or healthcare compliance.

The plugin-based approach means that new capabilities can be added without touching the workflow engine itself. Observers can be developed independently, tested in isolation, and deployed without impacting existing systems. This enables enterprises to build sophisticated, multi-layered workflow systems where each layer serves a specific purpose: one observer might handle compliance validation, another translates business rules into executable code, a third provides audit logging—all working seamlessly together.

The possibilities are as diverse as the challenges enterprises face. Complex routing logic that once required extensive JavaScript development can now be expressed declaratively. Integration with external rule engines and services becomes straightforward. Most importantly, subject matter experts and business analysts can participate directly in defining and maintaining business rules, reducing the gap between business requirements and technical implementation.