Human-Centered AI · Behavioural Products · Explainability
Sound Balance — Making AI-Supported Guidance Understandable
Context
Behavioural products can combine multiple signals to provide personalised recommendations, but those recommendations need to remain understandable rather than becoming opaque algorithmic outputs.
Scope
Exploratory case using a public behavioural and music dataset.
Problem
The challenge was exploring relationships between listening behaviour, focus and work context while keeping the resulting experience interpretable, transparent and non-intrusive.
My Role
Behavioural analytics & UX-led concept design combining:
Behavioural insight · Human-centered AI · Information architecture · Recommendation UX
Key Decisions
Designed explanations around visible behavioural relationships instead of black-box outputs
Organised guidance around different work moments and user situations
Framed relationships as signals for exploration rather than deterministic predictions
Outcome
Created a human-centered concept translating behavioural relationships into more understandable and contextual guidance.
The experience explores how AI-supported recommendations can remain useful without requiring users to blindly trust an algorithm.
Decision Impact Summary
Key Insight
Recommendations are easier to trust and use when people can understand why they are being shown and how they relate to their context.
Decision Supported
Help users explore behavioural patterns and decide whether a recommendation is relevant to the current situation.
Intended Decision Value
Demonstrate a more transparent approach to behavioural recommendation experiences.
UX Section
Recommendation Experience & Information Architecture
Designed an enterprise-grade AI dashboard structured by stakeholder needs.
Overview — Understand
Focus
Mood
Energy
Behavioural patterns
Analysis — Explore
Listening relationships
Focus patterns
Context
Guidance — Apply
Contextual recommendations
Work moments
User-controlled choices
Why it matters
Demonstrates how behavioural insight and UX can translate AI-supported recommendations into experiences people can understand rather than simply accept.
Applicable contexts
Human-Centered AI · Explainability · Behavioural Products · Digital Wellbeing
Extended case details available upon request.