TRPL
AI Photo Station

Producing an AI-Driven Interactive Experience for the Theodore Roosevelt Presidential Library

Role

Interactive Production Coordinator (Lead)

Timeline

May 2024 – March 2025

Overview

The AI Photo Station is an interactive exhibit designed for the Theodore Roosevelt Presidential Library (TRPL) that allows visitors to insert themselves into historically inspired scenes from Roosevelt’s life. Using AI-generated transformations, visitors take a selfie and see themselves reimagined across different narrative contexts—ranging from the Elkhorn Ranch to Roosevelt’s time in the Navy and the White House.

As the Interactive Production Coordinator and project lead for this system, I was responsible for shaping the experience from early concept through production—defining interaction flows, structuring content systems, coordinating across teams, and translating client feedback into a cohesive, functional exhibit.

My Role

  • Led production of the AI Photo Station as a core interactive within the TRPL experience
  • Defined interaction logic and user flow from attract loop to final output
  • Designed and structured the content system (background sets, categories, and asset logic)
  • Coordinated across design, programming, and content teams
  • Synthesized client feedback into actionable design and system changes
  • Directed testing and iteration, focusing on usability, realism, and narrative clarity

Ongoing Research

The project required balancing three competing priorities:

Historical Authenticity vs User Delight

The experience needed to feel grounded in Roosevelt’s world while still being engaging and playful.

AI Realism vs System Constraints

A central requirement was that users should feel like the output “actually looks like them”—a difficult challenge given variability in AI-generated results.

Narrative Structure vs Technical Simplicity

The system needed to support multiple thematic environments while remaining manageable from a content and CMS perspective.

Approach

Designing the End-to-End Interaction Flow

I developed a clear interaction structure to guide users through the experience:

Step 01

Attract Loop

draws visitors in with dynamic visuals and messaging

Step 02

Capture Moment

user takes a photo

Step 03

Transformation Phase

AI applies historical styling

Step 04

Result Display

final composited image within a themed environment

Special attention was given to transitions—ensuring that movement between stages felt smooth, intuitive, and visually cohesive within a public exhibit context.

Structuring the Content System

To support both narrative variety and system scalability, I organized the experience into distinct thematic categories, each tied to a period or identity from Roosevelt’s life:

Elkhorn Ranch
(nature-focused environments)

Navy
(nautical settings)

New York Elite
(formal, high-society styling)

Rough Riders
(camp and military-inspired scenes)

Amazon Expedition
(jungle environments, adjusted to remove weapons)

White House
(interior and exterior presidential settings)

Each category required:

Curated background assets

Consistent visual tone

Alignment with historical and client expectations

I worked closely with the team to refine these sets based on feedback—for example:

Expanding nature-focused backgrounds
 for Elkhorn

Increasing
nautical visual elements
 for Navy

Removing historically sensitive elements (e.g., weapons) from the
 Amazon set

Translating Client Feedback into System Decisions

A key part of my role was bridging client vision and technical implementation. Rather
 than treating feedback as surface-level changes, I translated it into structured updates:

  • “Does it look like the user?” → Adjusted evaluation criteria and testing focus
  • “More variety in backgrounds” → Expanded asset pools within categories
  • “Better thematic clarity” → Strengthened visual distinction between sets

This required constant iteration across:

Content selection

Prompting and output evaluation

Interaction pacing and clarity

Designing for Public-Space Usability

Unlike personal apps, the experience had to function in a high-traffic museum environment. I accounted for:

Quick comprehension

(users must understand the experience instantly)

Short interaction cycles

(to support throughput)

Clear visual hierarchy

for instructions and outputs

Accessibility considerations,

including readability and interaction clarity

Iteration and Testing

I led ongoing testing to evaluate:

  • Quality and consistency of AI outputs
  • User recognition (“does this feel like me?”)
  • Clarity of interaction steps
  • Engagement with different thematic environments

This required constant iteration across:

Background selection

Transition timing

Output presentation

System flow

Outcome

The AI Photo Station evolved into a multi-scenario interactive system that:

  • Integrates AI-generated imagery into a narrative museum context
  • Balances historical storytelling with personalized engagement
  • Supports multiple themed environments through a structured content system
  • Translates complex technical processes into a simple, intuitive user experience

Key Contributions

  • Defined the core interaction model for the experience
  • Built a scalable content and category system for themed outputs
  • Led cross-functional coordination across design, engineering, and content
  • Transformed client feedback into clear system-level decisions
  • Guided testing and iteration to improve realism, usability, and engagement

Reflection

This project reinforced how designing AI-driven experiences is not just about the technology—it’s about perception, trust, and interpretation.

The most critical challenge was not generating images, but ensuring that users:

  • Recognize themselves in the output
  • Understand the transformation
  • Feel meaningfully connected to the experience

It also highlighted the importance of structuring content systems early, especially when working with dynamic, multi-scenario interactions.

Scroll to top