← All workConcept / Screens

OASIS.

Making a reflective journaling flow easier to follow—and its outputs easier to question.

RoleUX researcher · Research, personas and usability testing
TeamFive-person team; prototype and visual-design ownership assigned separately
Timeline2026 · Concept research and usability review
StatusTeam prototype · Human usability study and AI-assisted evaluation
OASIS home prototype with a personalized welcome, Record, Diary and Insight actions, and example activity counts.
Supplied home prototype: Record, Diary and Insight form the three main entry points. Counts are illustrative.
Original project artifact

Supplied home prototype: Record, Diary and Insight form the three main entry points. Counts are illustrative.

The short version

Human usability sessions turned broad reflection goals into concrete questions about recording feedback, editing and interpretation.

Human study
Four participant records · five tasks
My focus
Research and usability testing
Iteration
V2 informed by AI feedback; no documented human retest

What I contributed

  • Conducted user research and developed personas for the team concept.
  • Facilitated usability tests; the supplied team report credits Jiajing Lin as UX researcher.
  • Contributed to research synthesis around recording feedback, navigation and interpretation.

Methods & making

Moderated think-aloud testingFive core tasksAffinity synthesisAI-assisted evaluationFigma prototype review

Five-person team; prototype and visual-design ownership assigned separately. Individual contributions are described separately from collective work.

Inside the project

Watch the flow. Check the meaning.

The moderated sessions covered account setup, onboarding, voice journaling, AI insights and emotional visualization. The important evidence was what people hesitated over and could explain after completing an action.

4
Participant recordsAnonymized task notes; a small formative study.
5
Core tasksFrom account setup to interpreting a visualization.
Evidence into a choice

Human observation

Participants looked for confirmation that recording had started, struggled to edit the transcript and questioned the meaning of chart values.

Source: Usability Test Note, pp2–9.

Inside the project

“Wait… is it recording right now?”

Evidence

A participant searched for confirmation after pressing Record. Other notes describe an immediate start, a cramped transcript editor and uncertainty about finishing.

Decision

Prioritize a visible start state, distinct review actions and an explicit saved confirmation. Treat editing space and the recording countdown as remaining testable requirements.

Source: human task notes, pp2,4,6,8; recommendations in the usability report, pp15–16. These observations do not establish the effectiveness of the revisions.
Inside the project

Inspect the recording decisions.

The flat mode-choice export and V2 wireframes expose different parts of the proposed flow. They remain labeled by fidelity.

Choose a mode

OASIS Record prototype offering Voice Recording and Text Entry, with a Home button.
Choose a mode

The supplied screen spells out Voice Recording and Text Entry, with a short explanation of each.

Supplied record prototype: voice and text are named with short explanations.
Choose a mode

Supplied record prototype: voice and text are named with short explanations.

Inside the project

Make the color mean something.

Participants asked what numbers meant and found some colors difficult to distinguish. The supplied calendar exports make that interpretation problem tangible.

Evidence into a choice

V2 calendar

Colored dates appear without an explanation in the screen. The V2 source itself notes the missing legend.

OASIS V2 diary wireframe with a month grid and colored entries.
V2 wireframe: the calendar uses color, while the accompanying source still identifies a missing legend.
Original project artifact

V2 wireframe: the calendar uses color, while the accompanying source still identifies a missing legend.

Inside the project

What the study changed for me.

Reaching an insight screen did not mean understanding it. I would next test recording feedback, transcript editing and plain-language explanations of how an insight relates to the entry, while preserving control over personal content.

OASIS is an exploratory reflection concept, with no diagnostic or treatment claim. AI-generated evaluations are separate from human evidence; prototype counts are illustrative.

Another context. Another question.

Airbnb