ProfileAtlas

Research study · Client work · Education

Role
Mixed-methods UX researcher
When
Aug 2025 to Aug 2026
Result
12 of 12 failures became tickets

PLUS human + AI tutoring

A human-plus-AI tutoring platform for underserved middle school students, from Carnegie Mellon, Stanford and Carnegie Learning.

  • 12usability failures found
  • 12 of 12became development tickets
  • 3workflows: goals, attendance, reflection

AskedAug 2025

Is the interface slowing tutors down?

Does the tutor interface slow tutors down?

LearnedFall 2025

Watch the sessions, not just the numbers.

Session replays in Microsoft Clarity, usage data, leadership interviews and competitive analysis.

Turned2026

Twelve specific failures, not a general slowdown.

Sharpened into 12 specific failures in goal-setting, attendance and reflection.

The modelA broad question, does the interface slow tutors down, became answerable once session replays resolved it into 12 specific failures across three workflows. All 12 became tickets.

  • The brief, as asked
  • The turn
  • Found and made
  • Set aside
  • Evidence, one dot per count

LandedAug 2026

All 12 went to development.

All 12 went to development; the public story was repositioned for funders, schools and tutors.

Details

My role
Mixed-methods UX researcher.
Proof
Session replay audit (Microsoft Clarity) surfaced 12 usability failures across goal-setting, attendance and reflection; all 12 became development tickets. Usage data triangulated with leadership interviews and competitive analysis.
Where it landed
Tickets shipped to development; public narrative repositioned for funders, schools and tutors.

ThreadsAccessTrust in AI

Related