Camunda launches agent orchestration platform for secure, scalable AI

Camunda has unveiled a suite of agentic orchestration capabilities intended to help organisations safely design and deploy AI agents at scale for business-critical processes.

The new release addresses a recognised difficulty in enterprise AI adoption: the lack of architectural guardrails that allow agentic AI deployments to progress reliably beyond the pilot phase. Camunda's architecture focuses on improving deployment simplicity, security and operational flexibility, seeking to offer businesses more control and governance while maintaining the benefits of AI autonomy where required.

Agent memory and contextual decision-making

Central to the new capabilities is agent memory. Camunda's composable short-term memory function gives AI agents the ability to recall conversational context, such as user preferences during a support session, enabling personalised responses and handling follow-up questions and feedback loops without the need for additional coding or performance compromise.

Long-term memory, powered by retrieval-augmented generation, equips agents to remember previous conversations, access domain-specific knowledge and customer data, and deliver context-rich reasoning for more accurate decisions. This feature enables, for example, an agent to recall a customer's ordering history in order to recommend suitable solutions tailored to their needs.

Camunda has also introduced a Copilot equipped with Friendly Enough Expression Language (FEEL) capabilities. This allows generation, validation and explanation of FEEL expressions from natural language, making the integration of AI outputs into business rules straightforward. Businesses can now, for example, transform a phrase like "applying 10% discount if order exceeds £250" into an auditable and executable rule automatically.

Agentic orchestration for business processes

The new functionalities are built to support orchestration of agents across complex end-to-end processes. Through Business Process Model and Notation (BPMN), these features are designed to help organisations coordinate work among specialised agents efficiently and transparently.

An AI agent connector is included to enable agents to perform autonomously within processes while integrating with large language model (LLM) providers, such as Azure OpenAI or AWS Bedrock. This gives organisations the flexibility to select models optimised for their specific use cases, balancing speed and cost considerations across various business processes with an emphasis on transparency and compliance.

With the introduction of a Model Context Protocol (MCP) connector, the platform supports multi-agent communication, allowing agents and processes to discover and invoke enterprise systems with full context. This is intended to unlock more advanced coordination patterns while maintaining necessary governance and transparency.

Industry perspectives

"Most agentic AI projects stall at pilot, not because the models aren't capable, but because there is not yet an architecture available that provides the guardrails to deploy agents to business critical processes without risk," said Daniel Meyer, CTO, Camunda. "Agentic orchestration changes that equation, giving enterprises control over how much autonomy to grant an agent, bringing control where needed and flexibility where AI shines. By blending deterministic flows with dynamic agents organizations achieve productivity gains that were once out of reach."

George Kutnerian, Co-Founder, President and CEO of Wellpointe, commented on the practical impact of Camunda's orchestration technology:

"Camunda's proven strength in orchestration made it possible for us to build an advanced medication order management system that delivers consistent, yet personalized and compassionate customer experiences-at scale. It's that same orchestration strength-now coupled with agentic capabilities within the governance we require-that has us excited to unlock even greater efficiency and innovation as we develop transformative care delivery models."

Deployment, security and governance

Camunda's update consolidates its orchestration stack-including Zeebe, Operate, Tasklist and Identity-into a unified application with Orchestration Cluster, streamlining deployment and making configuration, management and monitoring more straightforward. This approach is designed to support any scaling requirement and ensures operational resilience, including high availability and disaster recovery.

Security and governance enhancements include a new identity service to deliver enterprise-grade authentication and fine-grained authorisation across self-managed and software-as-a-service (SaaS) environments.

Meyer added: "Enterprise adoption of agentic orchestration requires a holistic platform. With our latest release we are truly marrying dynamic reasoning with deterministic guardrails, enabling organizations to design AI agents for trust and autonomy. With this foundation, the vision of the autonomous enterprise, where humans and AI agents work together, is no longer a pipe dream."