Trinity: AI-Driven Transition Planning for Students with Disabilities on AWS

University Startups has developed Trinity, a conversational AI platform designed to assist students with disabilities in creating individualized transition plans aligned with IDEA standards. Built on Amazon Bedrock and AWS serverless infrastructure, Trinity employs a hierarchical six-agent system to handle specialized tasks such as recommending colleges, careers, and training programs, while ensuring compliance with privacy regulations like HIPAA and FERPA.

The architecture features a multi-agent setup where each agent has a specific responsibility, collaborating under an orchestrator to generate comprehensive, IDEA-aligned transition plans. This design improves domain focus, retrieval accuracy, and path-aware sequencing, resulting in faster, more reliable outputs. The system separates recommendation and plan generation into two phases, enabling real-time, structured outputs suitable for export as PDFs.

To support accessibility, Trinity integrates Amazon Polly and Transcribe for audio features, and uses Amazon Cognito with Canvas LTI for seamless school login. Responsible AI measures, including content filtering and hallucination reduction via Amazon Bedrock Guardrails, are embedded throughout.

The platform has successfully reached over a dozen U.S. states within its first year, with international expansion into Saudi Arabia and Kuwait underway. Educators report increased efficiency and improved student engagement, with the system automating much of the IEP process and providing tailored, actionable pathways for students.

This scalable solution exemplifies how combining domain expertise, advanced AI architecture, and cloud infrastructure can transform special education planning, ensuring every student has a clear, personalized path forward.