How Pixieset achieved 35% AI feature adoption by solving the right problem with Amazon Bedrock

This post, co-written by Ry Rainey and Graham Gibson from Pixieset, explores how the company achieved 35% AI feature adoption by effectively solving a critical problem using Amazon Bedrock. Photographers and artists tend to be skeptical of generative AI due to its potential to threaten their craft and flood the industry with synthetic content. In fact, a 2025 MIT study found that 95% of enterprise Generative AI pilots deliver zero measurable returns.

Pixieset is a comprehensive photography business service that provides photo galleries, websites, stores, studio management, and photo editing. With over 8 billion photos hosted and trusted by millions of photographers globally, Pixieset recognized that to gain acceptance for their new AI feature, they needed to address a genuine pain point faced by their users.

The team identified that most photographer websites contain little to no alt text, which is crucial for search engine visibility. Writing unique and accurate alt text for hundreds to thousands of images is a time-consuming task that photographers often defer. This became the focal point of their product development efforts.

Through an internal hackathon, the team discovered the potential of generative AI and returned to the customer problem with a renewed perspective. They moved quickly from concept to production, launching an AI image alt text generator for millions of users within four months. Graham Gibson, Staff Product Manager at Pixieset, highlighted that in the first week alone, Pixieset generated alt text for over 750,000 photos.

The key challenge was not just building the feature but driving its adoption. To ensure trust and avoid alienating creative professionals, the team focused on solving a real problem: photographers spending time on work that isn't photography. Instead of embracing flashy applications like AI-generated images that could encroach on photographers' creative work, they targeted metadata, the invisible scaffolding that makes a portfolio visible to technologies and search engines but is tedious for photographers to create manually.

The AI-generated alt text feature is designed to be incremental, with a photographer reviewing and approving each suggested alt text before expanding its use. This approach allows photographers to build confidence in the output quality at their own pace, retaining final say over the captions.

The feature utilizes Amazon Bedrock's event-driven architecture and seamlessly integrates with Pixieset's existing infrastructure, requiring only a single API call without additional infrastructure or model hosting. By leveraging Amazon Bedrock's fully managed solutions and secure access to leading foundation models, Pixieset achieved rapid scaling and zero downtime since launch.