There are two ways to adopt AI in a business. You can layer AI on top of existing processes, or you can redesign those processes from scratch with AI at the center.
The first approach makes things about 20% faster. The second turns months of work into days. Most companies are still doing the first — and Camunda called this out directly.
May 20th, CamundaCon. In front of 1,200 enterprise leaders from 25 countries, Camunda's CEO made one declaration: "Every process in your business is legacy."
Why "legacy"?
The core argument from Jakob Freund (Camunda CEO): "Every business process in your organization was designed for a world without AI. Approval chains, exception handling, system handoffs — all of it was built on the assumption that humans do the coordination. That assumption no longer holds."
This isn't just marketing. Camunda's 2026 State of Agentic Orchestration report backs it up.
Of the 71% of organizations that say they "use AI agents," 80% are just running chatbots or summarization tools. Agents that actually move systems or make decisions? Far fewer. That's the reality of enterprise AI today.
ProcessOS aims to fix this: not add AI to existing work, but redesign work itself as AI-native from the ground up. It adds an AI intelligence layer on top of Camunda's battle-tested orchestration platform — already processing millions of concurrent workflow instances for some of the world's largest enterprises. Currently in closed beta.
Why call it an "OS"?
Like a computer's operating system, ProcessOS is the infrastructure layer that coordinates how individual AI agents communicate with each other and with existing systems (ERP, CRM, core banking). It's designed as an orchestration layer on top of existing systems — not a replacement.
20% vs 10x — what's the actual difference?
The most striking example is Danica, a Danish insurance company. After re-engineering their customer onboarding process with Camunda, what used to take months now takes days. They didn't just automate steps — they asked "did these steps need to exist at all?"
Camunda's "Great Re-Engineering" framework makes the structural difference clear.
| AI bolted onto legacy process | AI-native redesigned process | |
|---|---|---|
| Outcome | ~20% speed improvement | Months → Days (10x+) |
| Core question | "How do we automate this step?" | "Should this step exist at all?" |
| Bottlenecks | Human coordination assumption remains, structural bottlenecks persist | AI agents coordinate in real-time, bottlenecks removed |
| Tech debt | Existing complexity + AI layer = more debt | Outcome-focused design simplifies everything |
A Barclays managing director said it best at CamundaCon: "The real reason AI adoption stalls is that we can't build tomorrow's processes using only what we know today."
Camunda CTO Daniel Meyer echoes this: "The same shift happening in software development is coming for business operations. Developers used to write every line of code. Now AI writes more and more of it. Business processes are next."
How it works: the ProcessOS 4-step loop
- Discover
AI analyzes how your processes actually operate — not based on official documentation, but on real operational data. Hidden workarounds, bottlenecks, and unnecessary approval steps surface here. - Re-engineer
You describe the desired outcome and KPIs in natural language. ProcessOS proposes an AI-native process design, asking "should this step exist?" before building anything. Real simplification happens here. - Build & Deploy
The designed process gets deployed to production. AI agents, system integrations, data mappings, and UI forms are auto-generated — but every change requires human review and approval before going live. Audit and compliance coverage is built in. - Continuously Improve
After deployment, the process keeps learning. Successful patterns and exception points get recorded; accuracy increases over time. Human feedback loops accelerate learning. The more processes you run on the platform, the more shared knowledge compounds.
The AWS integration is worth noting too. ProcessOS runs natively on AWS, using Amazon Bedrock and Bedrock AgentCore for foundation models, agent memory, identity, and gateway services. Camunda won "Rising Star Technology Partner of the Year" at the AWS Partner Summit in Hamburg.
