IBM Moves Its Bob Coding Agent Inside the Firewall
Plenty of enterprises want AI coding agents. Fewer are willing to send their source code to someone else’s cloud to get them. That tension has slowed adoption at banks, insurers, government agencies and other organizations that build software under strict rules about where code and data can live. IBM is betting a self-hosted option will […]
Enterprise software builders often resist sending their source code to third-party cloud providers for AI coding assistance due to strict regulations governing code and data storage. To address this challenge, IBM has launched its AI coding agent, Bob, with a self-hosted version that can operate within private networks, sovereign clouds, or air-gapped environments. Introduced as a software-as-a-service offering in April, the on-premises deployment option has now been made generally available.
Unlike traditional coding agents, Bob functions as a comprehensive partner throughout the software development lifecycle. It coordinates specialized agents for code generation, testing, documentation, and pipeline management, utilizing various models, including IBM's own Granite. Bob incorporates approval checkpoints and security measures to monitor sensitive data usage and enforce organizational policies.
In a pilot with over 80,000 IBM employees since June 2025, Bob has reportedly increased developer productivity by an average of 45%. This success prompted IBM to expand Bob's capabilities to support Z mainframe applications, allowing organizations to retain control over sensitive code while reaping the benefits of AI coding assistance.
IBM's strategy hinges on the growing demand for AI solutions that comply with data sovereignty regulations. A June report by IBM's Institute for Business Value revealed that 68% of surveyed executives face difficulties meeting data residency and sovereignty requirements. With hybrid and edge deployments projected to capture 44% of the AI infrastructure market by 2030, IBM aims to provide a solution that bridges the gap between regulatory compliance and the adoption of AI coding agents.
However, self-hosting Bob introduces new complexities, such as managing GPU resources, updating models, monitoring agent behavior, and ensuring compliance with regulatory requirements. While self-hosting keeps code within the organization's control, it also necessitates the configuration of approval checkpoints, policy enforcement, and traceability mechanisms.
Despite these challenges, the move towards self-hosted AI coding agents like Bob represents a significant shift in the enterprise AI landscape, catering to the stringent needs of regulated industries.
Written by urgent.news from DevOps.com's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.