The rise of agentic AI has exposed a critical gap that most enterprises are only beginning to reckon with: AI systems that make autonomous decisions but have no reliable memory of what they did, why they did it, or how to undo it. In this VMblog Expert Interview, we sit down with Jessica Reeves, CEO and Board Member at Axoniq, and Allard Buijze, Founder and CTO, to discuss the company’s major new announcement — the Axoniq Framework — and why the problem of explainability and traceability in AI workflows has never been more urgent.
The Problem: AI Systems With Amnesia
Buijze frames the core challenge bluntly: most applications are built to forget. That has long been an accepted norm in software development, but in an era of autonomous agents making real business decisions, that architectural blind spot is no longer just a technical inconvenience — it is a liability. He points to a now-infamous example of an agent that deleted an entire database because it calculated a performance improvement, with no way to reverse the damage after the fact.
“It’s not only painful, it’s also irresponsible!” Without a system that captures every decision as it happens, there is no auditability, no traceability, and no path to remediation when something goes wrong.
Reeves echoes that sentiment from the business side, describing the current state of most enterprise AI implementations as “systems with amnesia” producing what she calls “a spaghetti mess” of disconnected data stores, custom glue code, and zero accountability. “No auditability, no traceability,” she says, “and very slow given it’s a spaghetti mess.”
15 Years in the Making
What makes Axoniq’s position distinctive is the depth of its history. The first line of code for the open-source Axon Framework was committed 15 years ago, giving the company a foundation that most competitors simply cannot replicate. The framework has been quietly powering event-driven, event-sourced architectures across some of the world’s largest organizations — including, according to Reeves, 80% of the Fortune 100 and 60% of the Fortune 500, with over 70,000 companies using the technology in some form.
That track record is now being channeled into something new.
Introducing the Axoniq Framework
The Axoniq Framework is the commercial evolution of the open-source Axon Framework, designed to meet the needs of enterprises at scale while preserving the open-source community that has grown around the project over the past decade and a half.
Key additions in the commercial offering include workflow support, agentic workflow capabilities, dynamic consistency boundaries, and a range of performance and efficiency enhancements. Buijze is careful to note that the programming model change required of developers is not radical — the shift is philosophical more than technical. Instead of designing systems around data, developers are encouraged to start with events: the decisions a system makes. From that foundation, everything else — auditability, traceability, AI training data — flows naturally.
“It’s a very small change, but it is a very fundamental change,” Buijze explains.
Reeves is equally direct about the competitive landscape. While Confluent and Kafka are names that come up in conversations around event-driven architecture, she draws a sharp distinction: moving data is not the same as understanding what that data means in a business context or remembering why it changed. The real competition, she argues, is not a single vendor — it is the DIY approach of stitching together five to seven different tools with custom integrations and hoping it holds.
Balancing Open Source and Commercial
The question of how to commercialize a product with a deeply loyal open-source community is one Axoniq has clearly thought through carefully. Buijze acknowledges that some features previously available in the open-source version are moving to the commercial tier, but he is unequivocal that the company is being transparent about it rather than engineering a bait-and-switch.
“Open source is not free,” he says plainly. “Somebody is paying for it.” The goal is to give the community a clear, honest picture of what remains open and what is now commercial, while ensuring that enterprises who need more — and are willing to pay for it — have a path to get there.
What Comes Next
Looking ahead, Reeves points to a “brownfield” capability in development that would allow organizations with legacy systems to gain visibility into what those systems are doing and begin an incremental transformation journey — without requiring a full re-architecture from day one. The message is that you do not have to be event-sourced at the core to start benefiting from what Axoniq offers.
Buijze, meanwhile, sees the Axon programming model as uniquely well-suited to AI-assisted development. Because the framework enforces a disciplined approach to capturing decisions and history, it leaves, as he puts it, “very little room for hallucinations” when AI coding assistants are involved. He envisions a future where developers can describe what they need in natural language and the framework handles the event-sourced implementation underneath — abstracting away the complexity while still capturing the full decision history.
“We just capture that history for you,” he says. “You don’t really have to worry about how it all works internally.”
Why This Matters Now
The timing of the Axoniq Framework launch is not accidental. As enterprises race to deploy agentic AI systems, the absence of a reliable memory and audit layer is becoming impossible to ignore. Axoniq is positioning itself as the infrastructure layer that makes AI deployable responsibly — not by slowing things down, but by ensuring that when something goes wrong, organizations have the foundation to understand it, explain it, and fix it.
With 15 years of momentum behind it and a commercial product now ready for enterprise scale, Axoniq is making its case that explainable AI is not a nice-to-have. It is the foundation everything else depends on.
Watch the full VMblog Expert Interview with Jessica Reeves and Allard Buijze above, and visit Axoniq’s website to learn more about the Axoniq Framework and how it fits into your architecture.
