Inside Axon Ivy 14: The Future of Enterprise Orchestration

What was the biggest challenge behind Version 14?

Marcel Prügel:

The biggest challenge was making the platform future-ready.

On the one hand, we had to modernize the IDE. We wanted to move away from Eclipse and introduce AI capabilities into the whole development process.

On the other hand, we also wanted to integrate AI into orchestration.

Which developments had the biggest impact on Version 14?

Marcel Prügel:

There have been two main drivers in the market over the last two years.

One, of course, is the whole development around AI, with AI capabilities at design time as well as at runtime.

The second was creating a modern developer experience. You want people to be able to implement the solutions they come up with directly in the platform, with Visual Studio Code as well as our Smart Workflow capabilities.

How do you modernize a platform without trying to improve everything at once?

Marcel Prügel:

We are addressing both signals and, of course, pains. But it is always important to really focus on specific topics and not try to improve the platform all over the place.

One of the crucial, and sometimes painful, decisions we made was to really focus on the modernization of the platform and stick with those main topics rather than getting distracted by other things.

That is always the hard part.

What sparked the biggest discussions around AI?

Marcel Prügel:

Our biggest discussions were around AI capabilities and their use cases.

If we jump back 16 months, everybody was talking about AI. But at the same time, everyone was asking: What is the real use case? Where does the benefit come from? Or is it all just a vision and an idea?

Figuring out how much freedom we should give AI, especially at runtime, became a big topic of discussion.

How do you give AI agents enough freedom without losing governance and control?

Marcel Prügel:

Axon Ivy is known as an open platform that enables developers to implement many different things.

At the same time, you don't want AI to be able to do everything in your platform. We want governance, restrictions, and guardrails.

Finding the right degree of freedom for AI at runtime is something we've discussed a lot, and it's also an area where we're looking for feedback from customers using Axon Ivy.

That balance becomes more important as AI agents take on a larger role in enterprise processes. The question is no longer simply whether an organization can use AI, but what an agent should be allowed to do within a real business process and where clear boundaries are needed.

For intelligent process orchestration, those questions belong together. AI agents do not operate in isolation. They become part of processes that also involve existing systems, established process logic, and human decisions.

What changed in Version 14 that users might not immediately see?

Marcel Prügel:

Some of the topics you can directly see when you use the platform, such as the new IDE.

Others happen more under the hood, like a lot of modernization of our workflow engine, scalability, reliability, and cloud readiness.

Not every important platform change comes with a new interface or feature users can immediately point to. Part of Version 14 is about the technical foundation needed for the platform to evolve further.

How did your thinking about AI change while building Version 14?

Marcel Prügel:

At the beginning of Version 14, we thought of AI as just a capability or integration.

But more and more, we realized that AI plays a big role at design time in Visual Studio Code, when we develop processes using agentic engineering.

It also plays a role at runtime, when we want to step away from purely deterministic processes and enable AI agents to orchestrate processes through our Smart Workflow capabilities.

So AI became much more integrated into the core capabilities of the platform. It's not just another feature.

This is perhaps the most fundamental shift behind Version 14. AI is no longer considered only as something that can be added to an individual application or process step. It increasingly affects both how processes are built and how they can operate at runtime.

For enterprises, that also changes the orchestration question. The challenge is to bring deterministic process logic and more adaptive, AI-driven execution together while retaining oversight over the end-to-end process.

What do you hope customers remember about Version 14?

Marcel Prügel:

I hope they are happy that they can finally let go of the Eclipse-based Designer and step into the new generation of the Axon Ivy PRO Designer, based on Visual Studio Code.

It's much nicer. It has many more extensions, you can work with AI, and you can bring your own LLM.

The move to Visual Studio Code therefore reflects more than a visual update to the development environment. It is part of the broader modernization that shaped Version 14 and the way developers can work with AI during development.

Version 14 in one sentence?

Marcel Prügel:

Version 14 modernized Axon Ivy not only on the development side, but also on the operations side.

It's about orchestrating humans, systems, and AI agents, and bringing intelligent orchestration one step closer.

From AI capabilities to intelligent process orchestration

Version 14 represents an important shift in how we think about AI within the Axon Ivy platform.

AI can support developers while they design processes. AI agents can also take on a role when those processes are running. The more autonomy they receive, the more important questions around governance, boundaries, and human oversight become.

That is why the developments Marcel Prügel describes are closely connected. A modern developer experience, AI at design time, AI at runtime, and the technical foundation beneath the platform all contribute to the same direction: intelligent process orchestration across people, systems, and AI.

Version 14 is a foundation for that direction, rather than an endpoint. It moves Axon Ivy closer to an environment where deterministic processes and more flexible AI-driven execution can coexist within end-to-end enterprise processes.