Policy work was disconnected from action.
Organizations need to turn changing EHS documents into training assignments, status visibility, and auditable reporting without manually rebuilding the same chain every time.
Cal Poly DxHub × AWS · Applied AI
AI Compliance Platform
A policy-to-training workflow designed to reduce repetitive administration while keeping people responsible for review, assignment, and compliance decisions.
Map the workflow before adding AI.
I defined user roles, document ingestion, question-answer support, assignment logic, reporting states, and administrative views, then connected them in a working Python prototype using Amazon Bedrock.
A usable system, not a chatbot demo.
The prototype showed an end-to-end operating model. Time and visibility figures below are project targets reported in the portfolio, not audited production outcomes.
Up to 90%
targeted reduction in manual administration
Up to 2×
targeted faster reporting workflow
+60%
projected compliance visibility
Top 100
selected from 1,300+ applicants
Evidence
Source material behind the case.
The project page distinguishes prototype targets from measured production results.