Goal

Design an accessible website to provide people living with low vision with information about ophthalmic conditions

Outcome

User research reports

Design guidelines

High-fidelity prototype

Background

The results of my work at Roche Oy are reported in my Master’s thesis “Leveraging Conversational User Interfaces and Digital Humans to Provide an Accessible and Supportive User Experience on an Ophthalmology Service”

To improve the adoption rates of the website, Roche wanted to design a better user experience that featured a digital human as a conversational agent

My role

🎤 Semi-structured interviews

🤝 Collaborative activities

🚶 Cognitive walkthrough

👀 Participant observation

🗣️ Focus groups

📋 WCAG 2.1-based evaluation

🤖 Low- and high-fidelity prototyping

My process

The research question behind this work was very interesting:

<aside> 🗣️ “Can a digital human provide a better user experience to people living with low vision through conversational interaction?”

</aside>

My design process involved exploratory research to define the problem space and the context of use. These insights informed the construction of the proof of concept. I was involved in prototyping both the UI and the conversational interaction for the PoC.I then tested the PoC through expert evaluation and a user-based evaluation session, which validated the concept.

🔎 Explorative research focused on end-users’ needs, requirements, preferences and opinions.

🎓 Evaluative research focused on design, usability and accessibility.

Results

Personas

I created three personas based on the research that was previously conducted. The personas were useful to keep everyone aligned on the goals and needs of the user group that we were designing for.

Exploratory user research

We conducted exploratory research through semi-structured interviews paired with collaborative activities on Miro. I wanted to understand people’s reactions to digital humans and expectations for its personality. The results informed the digital human’s character design.

Interface

In the prototype, only part of the envisioned content was available. We focused on visual accessibility and clarity.

Evaluation

The evaluative research was conducted through two methodologies: expert evaluation and user-based testing.

Expert evaluation

🚶 Cognitive walkthrough to investigate usability

📋 WCAG 2.1-based evaluation to investigate accessibility

User-based testing

👀 Participant observation with 14 users (caregivers and people living with mild and severe low vision)

🎤 Follow-up interviews and focus groups

During the exploratory research we discovered that the digital human did not look very friendly to the users. We strove to make her more friendly and when one of the participants came back to do the evaluation they said “Oh, look at her, she’s smiling now!”