MAiA ( My Ai Assistant)

Medium Fidelity Prototyping

​At the medium-fidelity stage, I translated early concepts into interactive flows that could be tested for clarity, usability, and alignment with user needs. The focus was on structure and functionality over visual polish, allowing stakeholders to evaluate the core user journey without distraction from final UI details.

For Maia’s Ask ELISA experience, this phase mapped out:

• End-to-end task flows, from initial query to tailored AI responses.
• Response variations, showing both short and long outputs, and how users could drill deeper into information.
• Integration touchpoints, such as opening related resources in new tabs or embedded modules.
• Editing and iteration paths, enabling users to refine their queries without losing context.

Medium-fidelity prototypes were critical in validating interaction patterns, navigation logic, and content delivery before committing to high-fidelity UI design. This ensured the product would not only function seamlessly but also deliver a smooth, intuitive experience for end-users.

High-Fidelity Design

​In the high-fidelity stage, the Maia interface evolved from functional wireframes into a polished, branded product ready for stakeholder review and near-production validation. This phase brought together the visual identity, refined typography, colour palette, and micro-interactions to create a consistent and engaging user experience.

Key refinements included:

• Branded UI components aligned with Proximus’ digital design system for seamless integration.
• Onboarding flows enhanced with clear visuals and microcopy to reduce friction for first-time users.
• Chat interaction states (short and long responses, loading animations, and error states) visually differentiated to guide users through their tasks without cognitive overload.
• Contextual integration points with existing platforms, ensuring Maia’s responses could be accessed in a familiar environment without breaking workflow.
• Accessibility checks to maintain compliance and inclusivity across all touchpoints.

By the end of this stage, Maia not only looked like a finished product — it delivered a user experience that felt intuitive, supportive, and aligned with the needs of its diverse internal audience. The high-fidelity prototype served as the foundation for development, ensuring both design clarity and implementation accuracy.

‍

‍

‍

Launch Demo

What I learned

Working on Maia was a reminder that designing an AI-powered product is as much about people as it is about technology. At the start, I was eager to explore the possibilities of AI, but I quickly realised that the real challenge lay in making those capabilities feel approachable and trustworthy. Every interaction had to be intuitive, every step in the flow clear, so that users could focus on their task rather than the tool itself.

Integrating Maia into Proximus’ existing systems taught me to think beyond isolated features and design for the reality of established workflows. It wasn’t just about creating something new it was about making sure it fit seamlessly into an environment people already relied on. This meant constant conversations with developers, AI specialists, and stakeholders, ensuring that every idea was not only desirable but also technically feasible.

Looking back, the most valuable lesson was that innovation doesn’t always come from adding more, sometimes it’s about stripping away complexity until only what truly serves the user remains. Maia showed me that when technology is paired with empathy and clarity, it can become not just a tool, but a trusted companion in someone’s daily work.

‍