Home Web Design Graphic Design Photography

Lawrence Getubig

Creative Technologist in Product Design

Swiss Post AI Tool

Home > Swiss Post AI Tool GUI

Role

  • Product Designer
  • UX/UI Web Designer
  • Illustrator
  • Brand Designer

Framing the Challenge

This project is a scrubbed case study inspired by work I originally did for USPS. While the visual design, UX, and UI are entirely new, the journey and business value mirror the kinds of strategic questions that the LLM‑powered generative AI assistant can help answer.

Framing Question: How can Swiss Post decision‑makers leverage AI tools to quickly generate insights in performance, sales, marketing, and logistics—driving smarter business decisions?

To begin, I conducted comparative market research into Swiss Post’s brand identity, rooted in Helvetica and Swiss design principles, and explored how those traditions of clarity and precision could inform a modern AI interface.

Download PDF Deck (opens a new tab)

Above: Moodboarding and visual design research examples.

Designing the Persona

Drawing from brand research, I sketched concepts for the AI’s visual persona.

We set out to create an AI identity that feels warm and approachable while maintaining reliability and trust. Should the AI avatar lean anthropomorphic, suggesting friendliness? Or remain abstract, signaling neutrality and focus?

Guiding Question: What visual symbol best represents a tool that engages naturally—free from the rigid, mechanical feel often associated with artificial intelligence?

swiss post avatar character exploration set 2
Above: Examples of early sketches and illustrations for AI avatar iconography and branding.

Atomic Design Foundations

The design team adopted an atomic design approach, starting with elemental UI units—buttons, inputs, icons—and progressively assembling them into larger “molecules,” then full interactive components.

This modular system ensured consistency and scalability across the product.

Samples of new visual assets made
Above: Example designs of the atomic UI building blocks. Credit: Thin-line icons from Phorphoricons.com

Building Journeys

With a solid atomic foundation, we combined UI atoms into molecules that support specific user journeys. Like an alphabet forming words and sentences, these molecules created a coherent, extensible GUI, providing clean UX/UI.

examples of more complex UI elements called molecules
Above: Examples of how the UI atoms combine to form more complex UI molecules.

Desktop Experience

We illustrated key scenarios of the AI tool in desktop view:

  • Landing page introducing the AI assistant.
  • Prompt‑response GUI offering multiple options based on user input.
  • Complex query handling with tables, charts, and data manipulation.
  • Feedback modal triggered by a thumbs‑down, allowing users to specify issues (e.g., “inaccurate answer,” “forgot context,” “factually wrong”).

For a detailed desktop view:
Download PDF Deck (opens a new tab)

swiss post avatar character exploration set 2
Above: Swiss Post's (proof-of-concept) landing page interface of the LLM‑powered generative AI assistant.
swiss post avatar character exploration set 2
Above: Screen illustrating how the AI presents pre-populated options based on prior prompts.
swiss post avatar character exploration set 2
Above: Example of the AI tool handling a complex prompt based on previous answers and vsiualizations.
swiss post avatar character exploration set 2
Above: Example of the course-correction journey, providing feedback for an inaccurate AI answer.

Responsive Design

We extended the design system to tablet and mobile, ensuring the AI experience remained seamless across touchpoints.

swiss post avatar character exploration set 2

Discussion and Next Steps

If deployed, the product lifecycle would include:

  • Usability testing and analytics to identify pain points.
  • Iterative improvements based on evidence‑driven design.
  • Amplification of positive interactions uncovered in data.

Future Considerations:

  • Which digital touchpoint—mobile, tablet, laptop, desktop—dominates usage?
  • Prioritize UX/UI optimization accordingly.
  • Investigate low‑engagement touchpoints: are they contextual or design‑related? Refine usability to encourage adoption.

Caveat: This fictitious scenario is based on a real proof of concept. In practice, early design‑thinking research—interviews, pain point mapping, demand assessment to identify if this tool will truly solve existing user problems—would be essential before product design begins.

Back to Home