Mashreq Deploys IBM AI Agents to Accelerate Bank System Integrations

Mashreq has successfully reduced integration design and troubleshooting duration by up to 70 percent through an early pilot of IBM Bob, an artificial intelligence software engineering platform. The Dubai-headquartered financial institution implemented the technology to address longstanding bottlenecks associated with manual processes and system connectivity.

Streamlining Complex Workflows

According to announcements from IBM, the deployment transformed integration tasks that traditionally required two to three hours of manual effort into operations completed within minutes. The institution maintains an active integration platform managing approximately 650 transactions per second across corporate digital channels, enterprise infrastructure, and payment networks.

Mohamed Abdel Razek, group head of technology, transformation and information at Mashreq, noted that the organization is investing in technologies that empower engineering teams to increase velocity while preserving system resilience and service quality.

Model Context Protocol and Secure Access

To enable the AI software engineering agents to operate effectively across the development lifecycle—including solution planning, development, code review, and production troubleshooting—the bank utilized IBM Context Forge. Built on the Model Context Protocol, the secure gateway provides artificial intelligence agents with controlled access to API specifications, architectural standards, and operational context.

• Handling approximately 4,000 monthly requests from AI agents to connect systems.
• Supporting deployment across banking, payments, and enterprise infrastructure.
• Exploring on-premises deployment options to satisfy local data residency requirements.

Santosh Vaidya, head of payments and digital services at Mashreq, emphasized the objective of establishing an environment where artificial intelligence agents collaborate alongside human engineering personnel.

Overcoming Industry Modernization Obstacles

The integration pilot aligns with findings from PYMNTS Intelligence research, which identified manual-intensive internal processes as the primary impediment to real-time payments modernization for 53 percent of financial institutions. Additional cited obstacles included legacy technology and siloed payment architectures.

The broader technical foundation supporting Mashreq involves collaboration with IBM and Gulf Business Machines, utilizing integration software, Red Hat OpenShift, and Confluent data streaming technology. The bank previously connected its corporate online banking channels to Aani, the United Arab Emirates central bank instant payments platform, to facilitate continuous round-the-clock transfers.