Camunda CEO: Enterprise-grade agents need enterprise-grade orchestration
Automation accelerates.
We know that automation services that stem from artificial intelligence services spanning predictive, generative and agentic AI are now being applied to the working coalface of businesses every day.
Although the actual deployment surface is said to still be comparatively small – magical analyst house Gartner estimates that real world deployment may be only seen in somewhere around 17% of businesses and the organisation agrees that this could be as low as 5% in real terms – automation is progressing forward and the goal has been to create more time for humans to focus on what matters most.
Camunda CEO Jakob Freund says that with agentic AI, it’s possible to automate work that has always been out of reach.
“Work that required human judgment, context, or deep knowledge. Think about onboarding a client at a global bank, processing trade exceptions, or handling sensitive healthcare cases. These processes involve endless variations and decisions. Until now, they couldn’t be fully automated,” wrote the CEO, on his own blog.
Agentic mushrooms
Freund says that AI agents are everywhere right now, whether they’re chat assistants, copilots, or embedded bots in apps.
“They’re useful, but most are isolated like mushrooms popping up after a rainstorm. Most often, these agents are task-focused and not built to handle the work enterprises care about most,” said Freund. “That’s why so many projects stall. Research [suggests] that nearly 40% of agentic AI projects will be scrapped by 2027 because they fail to create business value. 77% of leaders are worried about generative AI risks and 44% have already seen AI produce the wrong output. Despite the hype, only 8% of companies are scaling AI across the enterprise.”
Unsurprisingly, perhaps (Freund’s company sells agentic orchestration services and process management technologies), he thinks that these “mushroom” agents lack orchestration. They operate in silos, without the guardrails or transparency needed to complete mission-critical processes successfully.
“Traditional process automation has always involved deterministic orchestration, with well-defined processes, predictable outcomes and compliance guardrails built in. AI introduces the power of dynamic orchestration, or agents making situational decisions in real time. But by themselves, neither is enough to make enterprise-grade agents work,” said Freund.
What is agentic orchestration?
The so-called “breakthrough” Camunda heralds is agentic orchestration i.e. a blend of deterministic guardrails with dynamic autonomy.
“Agentic orchestration makes it possible to trust agents with high-value, high-stakes work. You can dial an agent’s autonomy up or down, add human-in-the-loop controls, enforce governance and still let AI handle the unexpected,” said Freund. “With this foundation, enterprises can finally break through the ceiling and bring AI into the very heart of their operations.”
CEO Freund says that his team believes AI agents are the biggest opportunity in business since the rise of the computer.
He concludes by saying that automation has always been about progress i.e. moving work from human hands to reliable systems.
Agentic orchestration is, therefore, the next step as it is focused on creating the conditions for people and agents can achieve more together than either could alone.
