← Selected work

AI experience design · Museum UX · Accessibility

MusePilot

A museum companion that handles what comes next so first-time and occasional visitors can stay immersed in the art and the real environment around them.

My roleConcept originator, product strategy, PRD, interaction model, and early prototyping

TeamFive-person student team

TimelineSpring 2026

ToolsFigma, FigJam, Figma Make

My contribution

I originated the concept, wrote the PRD, defined the product’s core behavior, and contributed to the early prototype. My teammates led much of the final visual system and high-fidelity motion prototype.

Project outcome

Critique and prototype feedback shifted MusePilot from a route-planning tool into a quiet, on-demand companion that supports visitors without competing with the art.

Solution

Help briefly, then return attention to the museum

MusePilot stays quiet while visitors explore. When uncertainty appears, it offers one relevant suggestion, gives only the guidance needed, and returns to Standby when the task is complete.

MusePilot immersive mode in a quiet standby state
01 · Standby keeps the interface unobtrusive during the visit.
MusePilot offering a timely artwork recommendation
02 · One timely recommendation responds to interest without prescribing a route.
MusePilot providing brief guidance to a nearby rest area
03 · Brief guidance ends when the visitor reaches what they need.
MusePilot final interface system showing quiet, time-aware museum guidance
The final interface system across orientation, exploration, recommendations, access support, and recap.

The challenge

Visitors wanted enough guidance to feel confident without turning the visit into another screen experience

First-time and occasional visitors face limited time, uneven energy, and unfamiliar collections. Existing tools often add another rigid layer instead of supporting spontaneous discovery.

Presentation slide summarizing overload, orientation, and rigid museum tools
We framed the problem around three recurring breakdowns: overload, orientation, and tools that cannot adapt.

Research evidence

Guidance mattered most at moments of uncertainty

Two interview rounds, museum observation, critique, and later qualitative prototype feedback with four participants helped us focus on starting choices, pacing, wayfinding, and screen attention.

Three visitor quotes about choosing artworks, reading while walking, and avoiding an intrusive tour app
Visitor language gave us a practical test: the product should help people decide without taking over the visit.
  1. ArrivalUncertainty

    Where should I begin with limited time?

  2. ExplorationFatigue

    How can I pause without feeling that I broke the plan?

  3. OrientationWayfinding

    Where is the next artwork, seat, restroom, or exit?

Product direction

A system designed around restraint

Support exploration Preserve wandering and discovery.

Keep visitors in control Let people choose when and how deeply to engage.

Appear at useful moments Keep assistance brief and infrequent.

MusePilot design principles from the final presentation
The original design principles translated visitor needs into product boundaries.
MusePilot system overview showing five moments from orientation to discovery
Five moments describe when the agent appears, assists, and fades back.

Early exploration

Paper before polish

Paper flows and storyboards helped the team test pacing before committing to high-fidelity screens.

Early MusePilot paper prototype
Paper prototype
Storyboard of a visitor using MusePilot in the museum
In-context storyboard

How feedback changed the product

The guide was becoming the visit

Our early concept organized a complete route and required frequent phone interaction. Critique made the conflict visible: visitors came to look, talk, and explore. Standby became the default and guidance became brief.

Feedback

Visitors were looking at the phone too often.

Earlier design

The product organized and monitored a complete route.

Change

The product offered one next step, then returned to Standby.

The product goal shifted from route completion to visitor confidence, attention, and freedom.

Accessibility and rest

Comfort information became part of the recommendation

Seating, stairs, water, restrooms, distance, and fatigue shape what a visitor can realistically do next. MusePilot brings this context into the flow instead of treating access as a separate feature.

MusePilot offering a contextual recommendation during a museum visit
01 · Understand the visitor

Capture comfort and access needs

Comfort and access needs are captured during setup.

02 · Recommend at the right moment

Respond to the visitor’s context

MusePilot considers distance, available seating, and signs of fatigue before suggesting a break.

03 · Return control

Continue whenever they are ready

Visitors can find a nearby place to rest and return to the experience on their own terms.

Outcome

A quieter product direction

This was a concept project, so the outcome is a tested design direction rather than a launched-product metric.

Reflection

Designing restraint

We initially treated better guidance as more guidance. MusePilot became more useful when it supported specific moments of uncertainty and then moved out of the way.

What I would explore next

  1. Test the concept during a real museum visit rather than only through a remote prototype.
  2. Measure whether the experience reduces screen attention while improving visitor confidence.

Interactive prototype

Explore MusePilot

Move through the final concept as a visitor. The prototype demonstrates quiet guidance, rest support, and the end-of-visit recap.

Open the live Figma prototype ↗
Next projectReWA