Camunda, a leader in process orchestration and automation, has detailed how its agentic orchestration capabilities are being utilized by organizations to integrate artificial intelligence (AI) at scale into their business processes while maintaining transparency, compliance, and control.
At CamundaCon 2025 Amsterdam, several customers presented their applications of Camunda to support AI initiatives. Among these, Halkbank Turkiye, a state-owned Turkish bank, is leveraging Camunda alongside AI to enhance its high-volume money transfer operations. In this process, customer orders, often submitted in free formats such as scanned letters, are processed using optical character recognition (OCR) and an on-premise AI model to extract transaction data. This data undergoes human review before execution. Camunda orchestrates the workflow, integrating AI dynamically within defined guardrails. This implementation reduced total processing time for a money transfer order from 54 seconds to 9 seconds, halved errors, and enabled 63% of transactions to be completed without manual correction.
Gülefşan Yıldırım Başaran, IT Manager at Halkbank Turkiye, stated, “Camunda is delivering measurable AI value today. The company is solving complex problems, orchestrating processes across legacy systems and modern AI tools, while doing so in a scalable, transparent, and compliant manner.”
Camunda’s partners also demonstrated agentic AI solutions built on its platform. EY showcased an agentic trade exception management solution (available on Camunda Marketplace). EY noted that many of its global financial services clients possess mature AI models and internal large language models (LLMs) but often lack an orchestration layer to translate these into tangible business outcomes. With Camunda, EY assists these organizations in embedding existing AI assets into structured, auditable workflows that comply with operational and regulatory standards. Camunda’s platform enables clients to connect their preferred AI models, whether cloud-hosted or internal via EY labs, and apply deterministic guardrails to ensure AI activation is appropriate. This approach allows clients to avoid rebuilding AI, focusing instead on governance, visibility, and scalable deployment. In one capital markets implementation, EY reported an 86% reduction in manual effort, a 98% reduction in T+1 delays, and an increase in analyst productivity from 6–10 cases per day to 41–64 cases per day, representing a sevenfold improvement.
Cognizant discussed the legal and regulatory aspects of deploying agentic AI in the European market, particularly concerning GDPR and labor laws that define limits on data usage and full automation. Cognizant has developed workflows within Camunda that incorporate mandatory human review steps, allowing AI to suggest actions while requiring manual approval before execution. This balance enables organizations to leverage AI-powered insights while adhering to regional regulations. Audit trails, escalation paths, and process visibility are integrated into the Business Process Model and Notation (BPMN) model, assisting organizations in demonstrating control over each agentic interaction. This resulted in the quality audit process being reduced from 138 minutes to 7–10 minutes, increasing auditor productivity by 20–30%, and cutting costs by 30–50%. All activities are traceable through embedded audit trails and escalation paths in BPMN.
Incentro presented how it integrated Camunda’s agentic orchestration into customer service workflows. Instead of standard auto-responses, Incentro developed an AI agent utilizing an LLM to analyze queries and draft real-time responses. The agent accesses a comprehensive FAQ and documentation set for specific answers. Camunda’s BPMN model structures the logic, with the agent dynamically selecting the optimal response path via an ad-hoc sub-process. When implemented with Payter, a Dutch payment technology company, Incentro reduced handling time per inquiry from 24 minutes to 12 minutes, with lead time cut by 58%. This improvement contributed to increased customer Net Promoter Score (NPS) and agent satisfaction without increasing headcount.
BP3 shared its approach to integrating agentic AI into decision-intensive workflows within regulated industries such as banking, financial services, and insurance (BFSI), pharmaceuticals, healthcare, and utilities. Their method employs LLMs alongside Decision Model and Notation (DMN) tables to generate “accept, reject, or refer” outcomes. In ambiguous scenarios, decisions are escalated to a human, allowing the AI to learn from real-world feedback.
Daniel Meyer, CTO of Camunda, commented, “While there is considerable discussion surrounding AI, many companies still seek a practical method for its safe and effective application within their core operations. Our approach provides organizations a means to integrate AI into their business processes without relinquishing control. Rather than yielding everything to a black box, AI is utilized where it provides the most value, orchestrated alongside human input and deterministic automation. This makes agentic orchestration flexible, observable, and designed for scalability.”
CamundaCon 2025 Amsterdam was Camunda’s largest event to date, with over 1,000 attendees and speakers from organizations including BT Group, HSBC, and T-Systems. Camunda enables organizations to orchestrate and automate processes across people, systems, and devices to overcome complexity, increase efficiency, and fully operationalize AI. Serving over 700 companies across various industries, including Atlassian, ING, and Vodafone, Camunda’s platform supports the design, orchestration, automation, and improvement of business-critical processes to accelerate digital transformation.
