A nutrition chatbot prototype for older adults. An academic group project exploring how guidance, personalization and clearer navigation can support a conversational interface.
Nutrition guidance needs to be understandable and practical. For older adults with different needs and levels of digital confidence, a chatbot also needs to explain how to begin, where to go and how to return.
My contribution
I contributed to research, interface design and development, prototyping, usability evaluation, user testing and revision as part of the group. The revisions presented here came from the team's evaluation of participant feedback.
Project goals
01
Older adults
Different needs, preferences and digital confidence.
02
Personalization
Onboarding collects dietary needs and preferences.
03
Guidance
Help people begin and navigate the conversation.
The prototype
NutriBot / 2025
NutriBot
The actual desktop prototype, rather than a completed mobile application.
02
Design improvements
01 / Getting started
Guidance at the point of use
Several participants completed onboarding but were unsure what to do next. A five-step tutorial added guidance at the point of use.
5 / 5
02 / Making actions recognizable
Making navigation visibly interactive
Some participants did not see menu entries as clickable. The revised design adds button boundaries and combines settings with accessibility options.
Before01
After02
The comparisons document interface changes. A measured usability improvement was not established.
03 / A visible way back
A direct route back to the chat
A direct Back to chat link was added where participants needed it: above the settings screen.
↖
03
Testing & iteration
12older adult participants
Task-based sessions included asking about salt intake, changing preferences, adjusting text size and requesting meal suggestions. Interviews gathered participants' experiences and suggestions.
Ask a nutrition question
Change preferences and text size
Request meal suggestions
These observations informed interface revisions. The available material does not establish a measured before-and-after usability improvement.
Accessibility in practice
The prototype explored adjustable text size, guidance and preference controls. Testing also showed that some longer answers and weekly meal plans were overwhelming to read.
◐Text-size controls and clearer guidance
↹More visible menu buttons and return navigation
↔Combined settings and accessibility options
04
Outcome & reflection
Delivered
The team built a functional high-fidelity prototype with onboarding, personalized chat responses, preference editing, settings and tutorial guidance. It was tested with the target group and revised in response. NutriBot remained an academic prototype.
Next steps
Visible guidance mattered even when the interface appeared simple. Longer responses and formatting still needed work; some accessibility features were only partially implemented, and mobile and tablet layouts were incomplete. Further work would include responsive layouts, response readability and assistive-technology testing.
