Pixieset, a platform for photographers, utilized Amazon Bedrock to introduce an AI-generated alt text feature. This initiative aimed to address the tedious task of image SEO, which photographers often avoid. The feature was rolled out to millions of users and achieved a 35% adoption rate within four months, according to the AWS Machine Learning Blog.
Key Points
- Pixieset launched an AI-generated alt text feature using Amazon Bedrock.
- The feature was developed and deployed within four months.
- It achieved a 35% adoption rate among users.
- The tool automates image SEO work for photographers.
- The implementation avoided impacting the creative aspects of photography.
Context
According to the AWS Machine Learning Blog, photographers are considered a skeptical audience for generative AI. Pixieset's approach focused on solving a specific, non-creative problem for this audience. This contrasts with other domain-specific model developments, such as ONESTRUCTION's Ishigaki-IDS foundation model for construction and BIM workflows, which was built with technical advisory from the AWS Generative AI Innovation Center and utilized synthetic data and a three-stage training pipeline on Amazon EC2.
Why It Matters
This case demonstrates how targeting specific pain points, rather than core creative processes, can drive significant adoption of AI features, even among initially skeptical user bases. It highlights the potential for AI to automate non-creative, time-consuming tasks.
What To Do
- Note how Pixieset identified a specific, non-creative pain point for AI application.
- Consider how Amazon Bedrock can be used to deploy AI features rapidly.
- Evaluate user adoption rates for new AI features to gauge their effectiveness.
- Explore the AWS Generative AI Innovation Center for technical advisory on domain-specific model development.