NIKO the Fox
Digital Assistant
Service Design
User Research
3-4 min read
Team:
Group of 4
Tools:
Figma, FigJam
Duration:
4 Months
NIKO the Fox AI Chatbot is an integrated feature within the NYCU AR App. The mission was to define the university mascot's personality and design a personalized conversational experience to increase user engagement.
I served as a UX Design Researcher on a team of 4, leading qualitative research, workshop facilitation, and personality framework development.
The university mascot lacked a defined personality and interaction style, resulting in low user engagement and an inability to meet student needs efficiently or empathetically.
Offer simultaneous functional solutions and emotional support, addressing user anxiety or frustration before providing practical troubleshooting.
Consolidate solutions and guidance into a dialogue to reduce unnecessary detours or switching platforms.
The final persona was defined as a reliable senior who excels at analyzing situations and offering rational advice, while exhibiting patience and empathetic support.
We developed prompt guidelines to ensure NIKO delivers a consistent conversational performance:
1. Emotion-First Support: First provide emotional validation if needed, then present the plan.
2. Needs Confirmation: Actively confirm user preferences to ensure precise suggestions.
3. Multi-Option Guidance: Offer 2-3 actionable solutions and guide the user towards decision-making.
4. Reliability & Efficiency: Responses must be structured and concise, focusing on high-value advice.
5. Empathy & Warm Tone: Use natural interjections and friendly language to increase approachability.
Evaluating NIKO's responses required first defining what "appropriate" meant in context. Through co-creation workshops and student interviews, I found that appropriateness wasn't just about accuracy, it was about timing: emotional validation must precede practical guidance.
This framing shaped how we assessed the prototype: we weren't asking "is this answer correct?" We were asking "does this response serve this student, in this moment, in the right way?"
We utilized a ChatGPT prototype by inputting the prompts and testing the system with scenario-based dialogue.
Conducted a Service Safari, stakeholder interviews, and student walkthroughs to map campus touchpoints and define a service-driven persona that balances functional efficiency with emotional support.
Campus data is scattered across multiple platforms, creating high friction. Our Laddering Analysis shows that students prioritize "maximizing time" and "efficient commuting" to fulfill their core value of Achievement.
Standard navigation tools (e.g., Google Maps) fail to account for campus shortcuts and facility details, such as specific building services. Students prioritize "lived-experience" data over generic information to maximize efficiency. This preference is driven by a motivation for Better Control of Time.
High-stress scenarios, such as bus delays or unclear service protocols, lower a student’s ability to process technical instructions. Providing emotional validation before offering practical solutions is essential to reduce user anxiety and sustain engagement.
"I want to explore diverse interests as long as they don't interfere with my professional studies."
JOBS-TO-BE-DONE
Quickly find reliable information and maximize available time; seek non-judgmental guidance and validation during moments of stress or confusion.
Pain points
Feels restricted by a homogenous social circle and requires curated "summary packs" to avoid wasting time.
Ran a 90-minute co-creation workshop with six participants to defined NIKO's personality, tone, and interaction style through direct user collaboration.
Relationship Mapping
The Concentric Circles activity guided participants to reflect on the most supportive figures in their lives. This step provided essential data on desired persona traits.
Trait Definition
Participants role-played scenarios and identified critical interaction needs. The overlapping traits voted were reliable, thoughtful, observant, and insightful.
Archetype & Final Voting
Participants mapped their final traits to established archetypes. The resulting archetypes were:
Scope Limitation
The prototype was focused solely on conversational flows and prompt logic, full multimodal integration (AR, voice, and visual UI) remains a challenge for future development.
Privacy & Personalization
While ideal to have personalized settings and memory, data privacy and consent make the customized dialogue difficult.
Evaluation Scope
Response quality was validated through scenario-based testing. A natural next step would be to develop behaviorally anchored evaluation criteria to make future assessments more consistent and scalable.
Conduct usability studies to refine the chatbot interactions and optimize the dialogue scripts.
Customizing mood-based responses based on user preferences to improve personalized and dynamic interaction.
