Mainframes still power many of the world’s most critical systems — but working with them often feels very different, especially for developers used to modern tools and workflows. While today’s development relies on Git, CI/CD pipelines, and AI assistants, these practices don’t always translate easily to the mainframe world.
At the same time, AI-powered tools like Github Copilot and ChatGPT have become part of everyday workflows, yet their understanding of legacy languages and mainframe environments is still limited. This makes it harder to achieve the same level of productivity and automation.
That’s where IBM Project Bob comes in. When I first heard about it, I was curious:
can it actually improve the day-to-day experience of working with mainframe systems?
What is IBM Project Bob?
IBM Project Bob is positioned as an AI-driven development assistant for mainframe environments, aiming to support developers across the entire software lifecycle—from planning to deployment.
Rather than focusing only on code generation, it acts more like a companion throughout the development process. Instead of switching between multiple tools and steps, the work happens directly in the IDE, with AI support integrated throughout—from writing and reviewing code to identifying issues early.
In practice, this shifts development from a sequence of isolated tasks to a more continuous workflow of writing, checking, and improving code. It helps developers interact with tools like Git or CI/CD pipelines more effectively, for example by assisting with code understanding, reviews, and the design of build and test workflows. In this way, it helps bridge the gap between mainframe development and modern engineering practices.
How You Interact with Bob (Modes)
Project Bob operates through different modes, each designed for a specific type of task:
| Ask | for explanations, questions, and quick insights |
| Plan | for structuring tasks and designing changes |
| Code | for implementation and code generation |
| Advanced | for deeper analysis and complex tasks |
| Orchestrate | for multi-step workflows combining multiple actions |
These modes can be configured to guide how Bob behaves and approaches problems.
First Impressions of Project Bob
After experimenting with Project Bob across different modes, the overall impression is that it provides solid support throughout the development workflow.
For understanding and review tasks, Bob performs really well. It can explain codebases, highlight common issues, and even outline system or business architecture in a clear and structured way.
Planning is another strong area, where Bob is able to break down changes into well-organized steps and consider impacts across the system. In implementation tasks, it can generate working code and guide testing, although some steps still require explicit instructions.
One interesting aspect is that Bob doesn’t just generate code and stop—it can detect issues during execution and adjust its approach, making the process more iterative and interactive.
In more complex scenarios, Bob shows its ability to coordinate multi-step workflows and switch between different modes to complete a task. At the same time, it also depends on available integrations and external tooling (for example through MCP-based extensions).
Overall, Project Bob feels like a capable assistant that supports the development workflow. It acts like a companion to developers in day-to-day work.
Why This Matters for the Future of Mainframe
For a long time, working with mainframes has required a different mindset, different tools, and often a steep learning curve—especially for new developers.
By introducing AI-assisted workflows and aligning development with more familiar practices like automation and continuous feedback, Project Bob helps reduce that gap. This can make mainframe development more accessible and easier to integrate into modern engineering workflows.
At the same time, it doesn’t remove the inherent complexity of mainframe systems. These environments remain critical and specialized — but tools like Bob suggest a direction where working with them becomes more approachable.
Final Thoughts
From my initial experience, it works well as a supportive layer throughout the development process—especially when it comes to understanding code, planning changes, and guiding implementation.
What stands out is not just what Bob can do today, but what it represents. It points toward a future where working with mainframe systems becomes more integrated, more accessible.
If you’re exploring how to introduce IBM Project Bob into your organization, we’re happy to share our experience and help you get started — from initial evaluation to integrating it into your existing workflows.
