AI Business

IBM Brings Bob Development Platform In-House With Self-Hosted Deployment

IBM has unveiled a self-hosted version of Bob, its AI-powered software development platform, enabling enterprises to run the tool on their own infrastructure while maintaining control over sensitive code and regulated data.

·3 min read
IBM allows on-prem deployment of its Bob agentic development platform
IBM allows on-prem deployment of its Bob agentic development platform

IBM Corp. unveiled a self-hosted deployment model for Bob, its agentic software development platform, granting enterprises the ability to harness artificial intelligence for application development and modernization on their own hardware. The deployment option accommodates on-premises, private-cloud, sovereign-cloud and air-gapped environments, addressing the needs of organizations where sensitive source code, regulated data or critical systems prevent reliance on externally hosted AI services.

Bob expands AI capabilities beyond simple code generation to encompass the full software delivery and modernization lifecycle. The self-hosted variant furnishes customers with enhanced authority over data residency, security protocols, governance frameworks and development processes.

Organizations need AI that operates inside environments they have control over, especially when working with sensitive code and regulated data

Neel Sundaresan, IBM's general manager of AI and automation

The move tackles a persistent challenge in enterprise AI adoption: deploying AI technology to applications without transferring sensitive information beyond approved boundaries. IBM referenced findings from its Institute for Business Value released in June, noting that 68% of executives surveyed identified meeting data residency and sovereignty requirements across different regions as a significant challenge.

The platform permits enterprises to execute supported models on local systems or integrate Bob with external model services through hybrid setups, enabling organizations to determine where AI computation takes place based on the demands of individual workloads.

Core Capabilities and Model Support

IBM confirmed that the self-hosted variant is now generally available and preserves all essential Bob features, encompassing its BobShell integrated development environment, parallel tool calling, skills and operating modes. The system assists developers in analyzing applications, designing modifications, executing development tasks and assessing outcomes.

Organizations deploying the self-hosted version must supply their own supported model. Currently, IBM supports Nvidia Corp.'s Nemotron and Poolside Inc.'s Laguna for models running on customer-controlled systems. Hybrid arrangements allow connections to external model services, providing organizations with flexibility for workloads with distinct security or operational considerations.

A self-hosted model configuration ensures that source code, development context and build artifacts stay within the customer's managed environment. Customers retain the option to leverage existing model investments through a bring-your-own-license structure.

Use Cases and Deployment Scenarios

IBM illustrated the platform's versatility with a financial services scenario: developers at a bank could leverage local model processing for core banking systems while routing less sensitive work to an approved external service. Bob maintains a consistent developer experience across these varied configurations.

Supplementary premium packages extend Bob's functionality to Java modernization and IBM i and IBM Z platforms, contingent on applicable licensing and deployment terms. IBM is targeting the offering toward financial services, government, healthcare and critical infrastructure sectors, alongside enterprises prioritizing intellectual property protection.

Market reaction proved positive, with IBM stock climbing more than 3% during late trading on the New York Stock Exchange following the announcement.