AI Quality Improvement Knowledge Agent

How Cassidy automates this using AI

Step 1: Trigger on quality signal or scheduled review

The Workflow activates when a safety event is reported, a dashboard threshold is breached, a regulatory gap surfaces, or a scheduled audit cycle begins—pulling in relevant context automatically.

Step 2: Retrieve and synthesize evidence

Cassidy searches the Knowledge Base for applicable clinical guidelines, eCQM specifications, HEDIS codes, local SOPs, and prior PDSA logs—returning grounded answers with citations to specific document versions and sections.

Step 3: Draft charter and measurement plan

The agent assembles a project charter with SMART aim statements, baseline metrics, and recommended outcome, process, and balancing measures aligned to CMS benchmarks and NQF stewardship.

Step 4: Compute measures and generate SPC charts

Cassidy connects to your EHR and EDW, applies measure logic with proper exclusions and risk adjustment, and produces run and control charts with equity stratification and benchmark comparisons.

Step 5: Guide PDSA cycles and capture learnings

The Workflow creates linked tests of change, collects cycle data, tracks predictions against results, and synthesizes lessons into KCS articles for reuse across the organization.

Step 6: Route for review with Human-in-the-Loop approval

All artifacts—RCA reports, driver diagrams, audit findings, and regulator-ready submissions—flow through configurable approval workflows with e-signatures and immutable audit trails.

Step 7: Publish and sustain

Approved content updates knowledge articles, refreshes standard work and order sets, and schedules re-audit reminders—maintaining an auditable record of data lineage, decisions, and version history for MIPS, HEDIS, and compliance reporting.