Build and orchestrate AI agents inside end-to-end business processes, with governance and auditability built in.
Every Camunda agent combines an LLM, tools, a system prompt, and structural guardrails inside a BPMN process. The model can reason and choose what to do next, while orchestration governs how each action executes.
Agent
Set clear boundaries and governance rules for the agent, defining the tools it can use and permitted actions.
Verify
AI agents need reliable validation, but traditional testing fails when responses vary. Camunda Process Test (CPT) runs your process in an isolated Camunda runtime inside your CI pipeline, so you can assert routing paths, guardrails, agent escalations, and other process behavior before production.
CPT generates a test coverage report showing which paths in your process were exercised and which weren’t. Judge assertions score agent reasoning against clear criteria, while non-technical teams validate workflows visually using Play. Verify that your AI decisions and business workflows consistently generate the correct outputs across every path.
Dashboard
▼ AgentLoanTest
Process: AgentProcess_LoanOrigination
Coverage: 100.0%
From simple task agents to complex multi-agent systems, use the same orchestration foundation to put agents to work across enterprise processes.
Task agent
Acts on systems
Queries databases, calls APIs, updates records, sends communications. The LLM reasons about which tool to invoke and when. Tools are external system connections: any system Camunda can reach.
Tools are
APIs, databases, services, messaging systems
Seen in
Fraud detection, order processing, customer data management
Seed export compliance agent
One agent subprocess, three real tool calls — a SQL lookup, a GraphQL API, and a REST calculation. It reads free-form shipment notes, then clears compliant shipments automatically or hands the rest to a human for review.
Enterprises running these agent patterns today.
Banking · Human-in-loop agent
Agents that investigate fraud alerts end to end: querying customer data, cross-referencing transaction patterns, selecting the right response path, and escalating to human investigators when the case warrants it. Running inside a structured process with a full audit trail.
Banking · Document processing agent
Agents that read free-format customer orders — scanned letters, unstructured text — extract transaction data using OCR and on-prem AI, route to human review when confidence is low, and execute automatically when it’s high. 50,000 transactions daily.
Telco · Conversational agent
Agents that guide customer case journeys — populating context from multiple systems, enriching cases with data before human agents pick them up, and escalating sensitively for vulnerability or bereavement. GenAI inside a structured orchestration layer.
Financial services · Orchestrator agent
Agents that handle trade exceptions — identifying discrepancies, querying counterparties, applying resolution logic, and escalating edge cases. Deterministic rules handle the predictable exceptions; agents handle the ones no rule anticipated.
Coordinate AI agents alongside people, decisions, applications, and core systems in one end-to-end process, with visibility and governance from the first request to the final outcome.
Give agents room to reason while Camunda provides the orchestration, governance, and resilience that mission-critical processes demand.
Enforce approvals, policy checks, and escalation paths structurally in BPMN so agents cannot bypass critical controls. Route exceptions, low-confidence decisions, and sensitive reviews directly to human colleagues within the same governed workflow.
Coordinate the complete end-to-end process journey around your agents while governing how they make decisions and use individual tools. Manage overall workflow progress and granular agent actions in a unified framework.
Use Camunda as the orchestration layer across any agent frameworks and use any AI model you choose, ensuring core business process logic remains completely portable and independent of a single provider.
Preserve process state through long-running work, failures, retries, timers, and external events so agentic processes can keep moving.
Test individual integration tasks and agentic subprocesses in Modeler before running end-to-end processes.
Run deterministic workflows, AI agents, and human tasks in one BPMN model with visibility across the end-to-end business process.
Zoom out from single execution races to monitor company-wide trends. Camunda Optimize provides an agentic control plane dashboard to manage operational performance and reliability.
Track trends for token usage across processes to prevent cost overruns and agent drift.
Detect unexpected agent behaviour changes before costs overrun or systems are disrupted.
Measure cycle time, escalation rates, and SLA adherence to know where agents succeed.
Camunda orchestrates your operations independently of AI vendors. Combine different LLMs and frameworks across a single process or switch providers anytime without rewriting core process logic. Every agent action, system call, and MCP connection remains fully governed and auditable in one central platform.
Select any model, including Anthropic, Amazon Bedrock, or custom LLMs. Coordinate frameworks like LangChain, CrewAI, and AutoGen in one unified process.
Camunda connects to Model Context Protocol (MCP) in both directions. Agents inside Camunda can invoke external MCP tools, while external AI systems can call Camunda’s MCP server to initiate end-to-end business processes, while keeping every step auditable and governed.
Connect enterprise applications, APIs, and systems through the integration layer and make them available to agents as tools.
Prompts can tell an agent what it should do. Orchestration can enforce what it must do. Model approvals, policy checks, human review, and escalation paths are mapped directly into the process so critical controls cannot simply be skipped.