I wanted a digital business card that made it easy to explore my work and get in touch. Instead of writing the page by hand, I asked IBM Bob to build imran.contact. A few prompts later, I had a working page to review, refine and share.
A real page, not just a code sample
The result brought together a responsive layout, IBM Plex Sans, Carbon design tokens, a live Toronto clock, a sticky header and subtle motion. It also linked to my IBM Spectrum HPC, Symphony, ELM and DevOps portfolios. Bob helped create the page and iterate on it; I guided the direction and decided what belonged in the finished experience.
That distinction matters. The useful test for an AI coding assistant is not whether it can produce an isolated component. It is whether it can help turn an intention into a coherent working product, then make changes as the requirements become clearer.
Where the human work still matters
Prompts can describe the goal, but someone still needs to decide what the page should communicate, which details matter, and whether the implementation works for real visitors. I reviewed the generated page, made adjustments and checked the experience across screen sizes. The assistant accelerated the build; it did not remove the need for judgment.
- Start with a clear outcome and the audience the page serves.
- Ask for a small, usable first version before adding refinements.
- Review the result in the browser, especially on mobile.
- Give specific feedback and iterate on the parts that affect clarity or usability.
- Keep responsibility for the final experience with the person or team shipping it.
A practical way to explore AI-assisted development
A personal contact page is a contained project, but it touches real design and implementation decisions: identity, navigation, responsive behavior, motion and links. That made it a useful place to experiment. I could see what the assistant handled well and where my direction and review improved the result.
For teams evaluating AI-assisted development, the same approach can help: choose a focused task, set clear constraints, inspect the generated changes and measure whether the result reduces effort without lowering quality.
You can see the finished card at imran.contact. IBM describes Bob as an AI development partner that can generate code, refactor, debug and update documentation; this project let me explore those capabilities on a page of my own.
Further reading: IBM Bob documentation.
Newsletter
Imran's Insights
Perspectives to help your team navigate AI, software delivery, engineering and high-performance computing.



