An eight-week proof of concept has tested whether autonomous AI agents can operate within enterprise compliance frameworks while meeting governance requirements. IBM and global technology firm e& completed the trial using IBM watsonx Orchestrate integrated with the OpenPages governance platform.

The companies claim the technology represents a shift from question-and-answer chatbots towards AI systems that can take autonomous actions. The system is designed to help employees and auditors access and interpret legal, regulatory and compliance information with what IBM describes as “clear, traceable responses aligned with enterprise governance requirements”.

IBM claims the deployment represents “one of the early enterprise-grade agentic AI implementations in the region”. The companies announced the collaboration at the World Economic Forum in Davos.

IBM’s watsonx Orchestrate provides access to more than 500 tools and customisable agents from IBM and partners.

The system’s architecture allows large language models to run across hybrid environments, including customer-managed infrastructure, while maintaining governance controls. This addresses concerns about data sovereignty and regulatory requirements that affect many organisations considering AI deployment.

The proof of concept tested whether AI agents could reason, orchestrate tasks and integrate with enterprise systems while remaining aligned with e&’s existing governance, risk and compliance framework. Success depends on the technology’s ability to provide explainable decisions that auditors and regulators can verify.

Hatem Dowidar, Group CEO of e&, said the company aims to move “beyond isolated AI use cases toward enterprise-scale agentic AI that is trusted, governed, and deeply integrated into how the organisation operates”.

e& is a global technology and investment conglomerate based in Abu Dhabi, United Arab Emirates

Ana Paula Assis, IBM’s SVP and Chair for Europe, the Middle East, Africa and Asia Pacific, said the proof of concept “intends to demonstrate how agentic AI can be designed and validated for enterprise-scale use, deeply integrated into core systems, governed by design, and trusted to support human-led decisions and outcomes”.