Boise State University builds campus-wide AI platform on AWS and cuts per-user costs by 80%

This article describes how Boise State University built a campus-wide AI platform on AWS, achieving 80% cost savings per user while maintaining FERPA compliance and providing equitable access to AI tools.

• Fragmented AI adoption across campus created equity issues and unsustainable $7.2M annual costs
• FERPA regulations prevented using commercial tools that train on student data
• Built boisestate.ai platform in six months using Amazon Bedrock with multiple models
• Platform provides access to Amazon Nova, Claude, and Llama through single interface
• Achieved $3 per user per month cost versus $20 commercial subscription rate
• Used AWS Lambda, RDS, DynamoDB, S3, and Control Tower for infrastructure
• Soft launch approach generated organic adoption without aggressive marketing
• Faculty can create custom assistants with uploaded documents and instructions
• Key lessons: start narrow, connect to ROI, leverage existing AWS agreements, small teams work

Boise State's approach demonstrates how universities can build compliant, cost-effective AI platforms using AWS consumption-based pricing while maintaining data privacy and user choice.