Pom

Research / User Testing / VUX / UX Design

Pom is a conversational companion designed for kids, aimed at promoting free-flowing conversation in the home environment. Developed during my internship at Kaliber Interactive, the project is the result of their long-standing partnership with De VoorleesExpress, and serves as an experiment for the company.

Kickoff

Kaliber Interactive, having limited experience in Voice User Interface (VUI) design, assigned the daunting task to two inexperienced interns, but with the support of experienced UX designers and knowledgeable programmers, we were able to overcome the challenge. Our journey began with a comprehensive exploration of the current state and history of VUI design.

1.

Background Research

Gave us a general understanding of what’s out there in the voice user interface world, the trends and projections.

2.

Contextual Product Positioning

Gave us an understanding of what the client wanted and how we should play into the current market.

3.

Experimenting with different voice agents

Figuring out what kind of solution works the best for our purpose.

4.

Prototyping Child – Voice Agent interaction

Seeing how interactions on our desired platform and design requirements connect.

Synthesizing research and business opportunities into a design porposal

How do you create a voice agent that does not listen to children but still allows them to talk freely, while documenting and future proofing for future VUI design?

Dynamic Dialog Structure

After careful consideration, we determined that using Google Assistant with Dialog Flow was the most optimal solution for our use case and programming capabilities. Despite its flexibility, it did not function as intended, requiring us to overcome obstacles and VUI barriers presented by Dialog Flow and develop a different dialog structure.

Case Development

To enhance the child’s connection with the device, we designed a case for the Google Nest Mini inspired by the stuffed animal representing the VoorleesExpress brand. The case was optimized for 3D printing as only a limited number was needed for the potential pilot test.