Camunda, Barclays Rework Processes for Agentic AI

Camunda and Barclays presented enterprise examples at CamundaCon in Amsterdam, held May 19–21 with about 1,100 attendees, that illustrate reengineering business processes for agentic AI. Camunda introduced a platform called ProcessOS and reported that a redesigned quote-to-cash flow freed 6,000 person-hours.

Camunda’s CEO Jakob Freund told the conference that agents handling customer data must include strong human approval steps and predictable behavior, a combination he labeled “the power of agentic orchestration.” He also pointed out that many corporate processes were designed before current AI capabilities existed.

Chief technology officer Daniel Meyer described ProcessOS as “an agentic operating system” that reengineers and continuously optimizes business processes for AI-driven tasks. Camunda used the platform to overhaul its quote-to-cash workflow, which the company said improved cycle time, raised efficiency and cut error rates. Chief financial officer Clemens Morgenroth reported the new workflow freed 6,000 person-hours, noting the prior process took about five hours per deal.

An April analyst report on adaptive process orchestration found the market shifting from task-level automation to platforms that combine process intelligence, modeling, execution and monitoring into a single orchestration layer. The research highlighted vendor focus on governance, auditability and hybrid execution models that mix automated steps with human oversight.

Barclays presented a production use case applying agentic orchestration to wholesale client onboarding and financial crime operations. Lily Wang, chief information officer for wholesale client onboarding and group financial crime, said ProcessOS addresses a barrier to enterprise AI adoption: companies cannot design future processes using only current assumptions.

Gautam Verma, head of financial crime core platforms and client due diligence technology, outlined the bank’s customer due diligence workflow as complex and slow, with many sequential handoffs and manual evidence gathering. Barclays developed three cooperating agents: one for data collection, a data intelligence agent that evaluates applicable policies and procedures, and an agent that executes the necessary policy steps. The agents are coordinated by a deterministic orchestration layer that runs across multiple internal systems and teams.

Both Camunda and Barclays emphasized retaining human oversight in agentic workflows. Camunda’s presentations stressed continuous optimization while preserving governance and deterministic paths. Barclays highlighted deterministic orchestration across hybrid systems to meet compliance and operational requirements.

Vendors at the conference described an industry trend of consolidating separate automation tools into orchestration backbones that provide a data foundation for AI behavior while keeping audit trails and human-in-the-loop controls. Executives framed ProcessOS and similar approaches as practical methods to adjust legacy processes for use with agentic AI, with attention to measurable process metrics and regulatory needs.