The Challenge

Thousands of images were ready. Meeting digital standards was taking days.

The Smithsonian's National Museum of African American History and Culture(NMAAHC) publishes thousands of images as part of its digital exhibition work. Every one of them requires alt text so visitors who rely on screen readers can access the content.

Writing those descriptions manually took days with multiple team members involved and end of cycle bottlenecks. Curators and content teams were spending days on manual descriptions instead of the interpretive work that makes exhibitions meaningful.

The team was constantly choosing between speed and doing it right. They needed a faster way to meet the standard without sacrificing quality.

What We Built

Layer in AI to do the heavy lifting

Fearless built a lightweight tool that generates draft alt text from folders of images. Large image sets that once took days to process now return a structured spreadsheet of descriptions ready for human review in a single session.

Keeping people in control

Our tool removes the burden of starting from scratch. It does not replace editorial judgment. We used prompt engineering to ensure the output reflected the tone, clarity, and cultural sensitivity that museum content requires. Every description is reviewed and approved by the museum's team before anything goes live.

Fits into existing workflows

The AI processing costs less than $10 a year to operate and requires no new platforms or disruption to how teams already work. We fit directly into existing process.

Outcomes

<$10
Per year to operate, down from multiple person-days of labor
1
Review session needed instead of multiple team days