AI HR Policy Search Agent

How Cassidy automates--header-using-ai

Step 1: Sync HR Knowledge to the Knowledge Base

Cassidy connects to SharePoint, Google Drive, Confluence, and uploaded PDFs to continuously ingest handbooks, benefits guides, and policy documents—preserving structure, metadata, and effective dates so answers stay current.

Step 2: Trigger on Employee Question

The Workflow activates when an employee asks a natural-language question in Slack or Microsoft Teams, capturing their query along with profile attributes like location, employment type, and business unit.

Step 3: Retrieve and Filter Relevant Policies

Cassidy searches the Knowledge Base using hybrid retrieval, then applies role and location filters to surface only the policies that apply to that specific employee—whether they're a California FTE, a German contractor, or a Toronto-based union member.

Step 4: Generate Cited, Grounded Response

The Agent drafts an answer constrained to retrieved sources, embedding inline citations with section numbers, effective dates, and deep links back to the source document so employees can verify and explore further.

Step 5: Offer Next-Best Actions

Alongside the answer, Cassidy presents actionable buttons—start a PTO request, initiate an LOA workflow, enroll in benefits, or submit an expense—connecting directly to your HRIS or ITSM system.

Step 6: Escalate with Full Context

When confidence is low or the query involves sensitive matters, Cassidy opens a case in ServiceNow, Jira, or your HR system—attaching the conversation transcript, retrieved citations, and employee attributes so HR can resolve it with full context and audit trails.