Research for Human–AI Collaboration
[Overview]
The Context
As Lead UX Designer at Sport Autonomy and MA researcher at Falmouth University, I’m leading a qualitative study on how AI insights can support — not replace — human coaching.
This work informs Coachi’s next phase: a new instructor-facing feature that turns learner data into actionable, personalised feedback.
The Context
As Lead UX Designer at Sport Autonomy and MA researcher at Falmouth University, I’m leading a qualitative study on how AI insights can support — not replace — human coaching.
This work informs Coachi’s next phase: a new instructor-facing feature that turns learner data into actionable, personalised feedback.
The Challenge
Instructors need clear learner insights — not prescriptive tools that override judgement.
My challenge was to design AI as a transparent, collaborative partner that surfaces patterns while preserving trust, empathy, and instructor autonomy.
The Challenge
Instructors need clear learner insights — not prescriptive tools that override judgement.
My challenge was to design AI as a transparent, collaborative partner that surfaces patterns while preserving trust, empathy, and instructor autonomy.
Research Insights
Leading end-to-end research bridging academic rigour and product goals.
Defined scope, built a theory-informed process, and applied Human–AI Collaboration principles to uncover actionable design opportunities — all while collaborating cross-functionally with engineering.
Research Insights
Leading end-to-end research bridging academic rigour and product goals.
Defined scope, built a theory-informed process, and applied Human–AI Collaboration principles to uncover actionable design opportunities — all while collaborating cross-functionally with engineering.
Skills Demonstrated
Miro
User interviewing
Contextual enquiry
Research planning
UX research & strategy
Human–AI interaction design
[Impact]
Built a 12-Step Qualitative Research Framework
Planned a multi-phase study using thematic analysis and Human–AI Collaboration theory to uncover how instructors interpret, trust, and act on AI feedback — setting the foundation for explainable, autonomy-supporting interface patterns in Coachi’s next release.
Aligned Academic Rigor with Product Strategy to Maximise Research Impact
Mapped research scope, ethics, and participant criteria to meet both MA-level standards and Coachi’s product roadmap — ensuring findings are directly translatable into interaction models, information hierarchy, and design principles for trusted AI systems.
Created a Research Pipeline, Connecting Design & Data Engineering
Planned a collaborative process to ensure insights flow seamlessly from field interviews into prototype development — bridging disciplines early and positioning UX as the strategic driver of Coachi’s next-stage product evolution.
[My Process]
1. Data Collection
Each participant will be observed in their natural teaching environment, followed by a reflective interview.
This combined approach captures both behavioural practice (what instructors do) and reflective cognition (how they think about it).
1. Data Collection
Each participant will be observed in their natural teaching environment, followed by a reflective interview.
This combined approach captures both behavioural practice (what instructors do) and reflective cognition (how they think about it).
2. Observations
Observing real or simulated ski lessons to explore how instructors:
Collect and interpret learner data
Engage with or disregard AI feedback
Adapt decisions based on context and individual needs
Captured insights through field notes, reflexive memos, and optional visuals (with consent).
2. Observations
Observing real or simulated ski lessons to explore how instructors:
Collect and interpret learner data
Engage with or disregard AI feedback
Adapt decisions based on context and individual needs
Captured insights through field notes, reflexive memos, and optional visuals (with consent).
3. Semi-Structed Interviews
Reflecting on Instructor Decision-Making
Holding post-observation interviews to understand how instructors:
Reflect on their own teaching decisions
Build or lose trust in AI feedback
Integrate technology into their coaching
Sessions were guided by open-ended prompts, recorded, and transcribed for analysis.
3. Semi-Structed Interviews
Reflecting on Instructor Decision-Making
Holding post-observation interviews to understand how instructors:
Reflect on their own teaching decisions
Build or lose trust in AI feedback
Integrate technology into their coaching
Sessions were guided by open-ended prompts, recorded, and transcribed for analysis.
4. Analysis Approach
All interviews and observation notes will be analysed using Reflexive Thematic Analysis (Braun & Clarke, 2006), progressing from explicit surface meanings to deeper interpretive insights.
4. Analysis Approach
All interviews and observation notes will be analysed using Reflexive Thematic Analysis (Braun & Clarke, 2006), progressing from explicit surface meanings to deeper interpretive insights.
5. Semantic Coding
Identify and label meaningful segments that describe explicit actions, statements, or experiences.
Focus on what participants say and do when interacting with AI tools and making instructional decisions.
Create digital sticky notes in Miro to visually map emerging ideas.
5. Semantic Coding
Identify and label meaningful segments that describe explicit actions, statements, or experiences.
Focus on what participants say and do when interacting with AI tools and making instructional decisions.
Create digital sticky notes in Miro to visually map emerging ideas.
6. Semantic Theme Development
Cluster related codes into candidate themes that represent patterned meanings across participants.
Review and refine themes for internal consistency and distinctiveness.
Develop a thematic map showing relationships between emerging concepts.
6. Semantic Theme Development
Cluster related codes into candidate themes that represent patterned meanings across participants.
Review and refine themes for internal consistency and distinctiveness.
Develop a thematic map showing relationships between emerging concepts.
7. Latent Interpretation
Move beyond surface meanings to explore the underlying assumptions, beliefs, and values shaping instructor behaviour.
Interpret how participants conceptualise trust, control, and collaboration with AI.
Use theoretical frameworks such as Human–AI Collaboration, Distributed Cognition, and Learner-Centred Design to deepen interpretation.
7. Latent Interpretation
Move beyond surface meanings to explore the underlying assumptions, beliefs, and values shaping instructor behaviour.
Interpret how participants conceptualise trust, control, and collaboration with AI.
Use theoretical frameworks such as Human–AI Collaboration, Distributed Cognition, and Learner-Centred Design to deepen interpretation.
8. Synthesis
Each theme will be examined for design implications — how the findings should influence interface decisions, interaction models, and information hierarchy.
Insights will feed directly into the next Coachi prototype, focusing on:
Reducing cognitive load
Clarifying AI reasoning
Supporting instructor trust and control
8. Synthesis
Each theme will be examined for design implications — how the findings should influence interface decisions, interaction models, and information hierarchy.
Insights will feed directly into the next Coachi prototype, focusing on:
Reducing cognitive load
Clarifying AI reasoning
Supporting instructor trust and control
[Key Learnings]
Surface Themes Aren’t the Whole Story
Learned that identifying semantic patterns is just the starting point — deeper insights emerge by questioning what’s not being said and exploring underlying assumptions.
Latent Interpretation Requires Reflection
Discovered the importance of iterative analysis and researcher reflexivity in moving beyond obvious themes to uncover hidden meaning and cognitive models.
Theoretical Lenses Add Clarity, Not Complexity
Found that applying frameworks like Distributed Cognition helps make sense of complex behaviour patterns — turning raw qualitative data into design-relevant insights.
[Persona]
Jhon Roberts
Marketing Manager
Content
Age: 29
Location: New York City
Tech Proficiency: Moderate
Gender: Male
[Goal]
Quickly complete purchases without interruptions.
Trust the platform to handle her payment securely.
Access a seamless mobile shopping experience.
[Frustrations]
Long or confusing checkout processes.
Error messages that don’t explain the issue.
Poor mobile optimization that slows her down.
[Overview]
The Context
As Lead UX Designer at Sport Autonomy and MA researcher at Falmouth University, I’m leading a qualitative study on how AI insights can support — not replace — human coaching.
This work informs Coachi’s next phase: a new instructor-facing feature that turns learner data into actionable, personalised feedback.
The Context
As Lead UX Designer at Sport Autonomy and MA researcher at Falmouth University, I’m leading a qualitative study on how AI insights can support — not replace — human coaching.
This work informs Coachi’s next phase: a new instructor-facing feature that turns learner data into actionable, personalised feedback.
The Challenge
Instructors need clear learner insights — not prescriptive tools that override judgement.
My challenge was to design AI as a transparent, collaborative partner that surfaces patterns while preserving trust, empathy, and instructor autonomy.
The Challenge
Instructors need clear learner insights — not prescriptive tools that override judgement.
My challenge was to design AI as a transparent, collaborative partner that surfaces patterns while preserving trust, empathy, and instructor autonomy.
Research Insights
Leading end-to-end research bridging academic rigour and product goals.
Defined scope, built a theory-informed process, and applied Human–AI Collaboration principles to uncover actionable design opportunities — all while collaborating cross-functionally with engineering.
Research Insights
Leading end-to-end research bridging academic rigour and product goals.
Defined scope, built a theory-informed process, and applied Human–AI Collaboration principles to uncover actionable design opportunities — all while collaborating cross-functionally with engineering.
Skills Demonstrated
Miro
Miro
User interviewing
User interviewing
Contextual enquiry
Contextual enquiry
Research planning
Research planning
[My Process]
1. Data Collection
Each participant will be observed in their natural teaching environment, followed by a reflective interview.
This combined approach captures both behavioural practice (what instructors do) and reflective cognition (how they think about it).
1. Data Collection
Each participant will be observed in their natural teaching environment, followed by a reflective interview.
This combined approach captures both behavioural practice (what instructors do) and reflective cognition (how they think about it).
2. Observations
Observing real or simulated ski lessons to explore how instructors:
Collect and interpret learner data
Engage with or disregard AI feedback
Adapt decisions based on context and individual needs
Captured insights through field notes, reflexive memos, and optional visuals (with consent).
2. Observations
Observing real or simulated ski lessons to explore how instructors:
Collect and interpret learner data
Engage with or disregard AI feedback
Adapt decisions based on context and individual needs
Captured insights through field notes, reflexive memos, and optional visuals (with consent).
3. Semi-Structed Interviews
Reflecting on Instructor Decision-Making
Holding post-observation interviews to understand how instructors:
Reflect on their own teaching decisions
Build or lose trust in AI feedback
Integrate technology into their coaching
Sessions were guided by open-ended prompts, recorded, and transcribed for analysis.
3. Semi-Structed Interviews
Reflecting on Instructor Decision-Making
Holding post-observation interviews to understand how instructors:
Reflect on their own teaching decisions
Build or lose trust in AI feedback
Integrate technology into their coaching
Sessions were guided by open-ended prompts, recorded, and transcribed for analysis.
4. Analysis Approach
All interviews and observation notes will be analysed using Reflexive Thematic Analysis (Braun & Clarke, 2006), progressing from explicit surface meanings to deeper interpretive insights.
4. Analysis Approach
All interviews and observation notes will be analysed using Reflexive Thematic Analysis (Braun & Clarke, 2006), progressing from explicit surface meanings to deeper interpretive insights.
5. Semantic Coding
Identify and label meaningful segments that describe explicit actions, statements, or experiences.
Focus on what participants say and do when interacting with AI tools and making instructional decisions.
Create digital sticky notes in Miro to visually map emerging ideas.
5. Semantic Coding
Identify and label meaningful segments that describe explicit actions, statements, or experiences.
Focus on what participants say and do when interacting with AI tools and making instructional decisions.
Create digital sticky notes in Miro to visually map emerging ideas.
6. Semantic Theme Development
Cluster related codes into candidate themes that represent patterned meanings across participants.
Review and refine themes for internal consistency and distinctiveness.
Develop a thematic map showing relationships between emerging concepts.
6. Semantic Theme Development
Cluster related codes into candidate themes that represent patterned meanings across participants.
Review and refine themes for internal consistency and distinctiveness.
Develop a thematic map showing relationships between emerging concepts.
7. Latent Interpretation
Move beyond surface meanings to explore the underlying assumptions, beliefs, and values shaping instructor behaviour.
Interpret how participants conceptualise trust, control, and collaboration with AI.
Use theoretical frameworks such as Human–AI Collaboration, Distributed Cognition, and Learner-Centred Design to deepen interpretation.
7. Latent Interpretation
Move beyond surface meanings to explore the underlying assumptions, beliefs, and values shaping instructor behaviour.
Interpret how participants conceptualise trust, control, and collaboration with AI.
Use theoretical frameworks such as Human–AI Collaboration, Distributed Cognition, and Learner-Centred Design to deepen interpretation.
8. Synthesis
Each theme will be examined for design implications — how the findings should influence interface decisions, interaction models, and information hierarchy.
Insights will feed directly into the next Coachi prototype, focusing on:
Reducing cognitive load
Clarifying AI reasoning
Supporting instructor trust and control
8. Synthesis
Each theme will be examined for design implications — how the findings should influence interface decisions, interaction models, and information hierarchy.
Insights will feed directly into the next Coachi prototype, focusing on:
Reducing cognitive load
Clarifying AI reasoning
Supporting instructor trust and control
[Key Learnings]
Surface Themes Aren’t the Whole Story
Learned that identifying semantic patterns is just the starting point — deeper insights emerge by questioning what’s not being said and exploring underlying assumptions.
Latent Interpretation Requires Reflection
Discovered the importance of iterative analysis and researcher reflexivity in moving beyond obvious themes to uncover hidden meaning and cognitive models.
Theoretical Lenses Add Clarity, Not Complexity
Found that applying frameworks like Distributed Cognition helps make sense of complex behaviour patterns — turning raw qualitative data into design-relevant insights.
[Impact]
Built a 12-Step Qualitative Research Framework
Planned a multi-phase study using thematic analysis and Human–AI Collaboration theory to uncover how instructors interpret, trust, and act on AI feedback — setting the foundation for explainable, autonomy-supporting interface patterns in Coachi’s next release.
Aligned Academic Rigor with Product Strategy to Maximise Research Impact
Mapped research scope, ethics, and participant criteria to meet both MA-level standards and Coachi’s product roadmap — ensuring findings are directly translatable into interaction models, information hierarchy, and design principles for trusted AI systems.
Created a Research Pipeline, Connecting Design & Data Engineering
Planned a collaborative process to ensure insights flow seamlessly from field interviews into prototype development — bridging disciplines early and positioning UX as the strategic driver of Coachi’s next-stage product evolution.