Project Overview
Solution
We designed an interactive onboarding flow for Sky Glass that introduces users to the voice assistant in an inclusive, personalised, and intuitive way.
The flow acknowledges different user need through voiceprints, contextual awareness, and guided prompts, helping users understand and adopt voice control features from the start.
Research and Assumptions
Knowns
Elderly users rely on voice for daily tasks
Gesture control supports users who can’t speak or use remotes steadily
Unknowns
Accessibility barriers within current Sky voice UX
Which support features are most valuable?
How users learn gestures
Assumptions
Users with speech or motor impairments struggle with rigid commands
Reminders could help users with cognitive impairments
Adding gestures could improve accessibility for hearing impaired users
Knowns
Users want voice control to feel natural and responsive
Elderly users rely on voice for daily tasks
Gesture control supports users who can’t speak or use remotes steadily
We ran five interviews to understand habits, frustrations, and expectations around voice and gesture interaction.
Voice use: 4/5 regularly use assistants (Siri, Alexa) for quick tasks.
Expectations: Accuracy and accent flexibility were key; users want tolerance for mispronunciations.
Frustrations: Recognition errors (4/5), over-activation (2/5), shallow responses.
Interaction: Users repeat/rephrase before switching to manual input.
AI vs Voice: 3/5 prefer AI for search due to more relevant, tailored answers.
TV habits: Discovery often happens outside the platform. Users want better search tools (e.g. mood-based, trailers, “similar to X,” lists).
Persona
Ideation
Inclusivity
We explored implementing an interactive assistant that actively acknowledges user input, creating a more natural and inclusive interaction.
Contextual Awareness
Ideas focused on enabling Sky Glass to detect the primary viewer, allowing the assistant to offer contextually relevant suggestions based on who’s watching.
Personalised Experience
We considered allowing users to register their voiceprints, improving recognition accuracy and enabling the system to identify individual speakers.
Voice Control Guidance
Concepts centred on providing clear, integrated guidance to help users discover and learn the full range of voice control capabilities.
Wireframing Workflow

Usability Testing
3 Moderated Tests with children aged 10 ±3 that lasted up to 15 minutes
Issue: Parental Control priorities
Severity: 3
Key Insight: Parents want simple, voice enabled parental set up
Design Implication: Integrate quick parental setup via voice commands
Issue: Voice vs Control Concerns
Severity: 5
Key Insight: Parents want filters on voice and interface
Design Implication: Add content filtering + child- friendly assistant tone
Issue: Setup Complexity
Severity: 4
Key Insight: Children confused by unclear prompts and mid-flow steps
Design Implication: Simplify activation + structure parental steps logically
Issue: Profile Privacy Confusion
Severity: 2
Key Insight: Privacy concerns during personalisation
Design Implication: Include clear privacy reassurance during setup
Outcomes


