How to build a HR Policy Chatbot

Workflow Overview

1. User Question (Input Node)

• Purpose:
  Collects the HR policy question from the user.
  Example: “How much compassionate leave do I get annually?”

• Node:
  User Question
  (in-0)

2. SharePoint Knowledge Base Search (Knowledge Base Node)

• Purpose:
  Searches your SharePoint-based HR policy documents for relevant information to answer the user’s question.

• How it works:
  The user’s question is sent to this node, which performs a semantic search across your HR policy files in SharePoint.

• Node:
  SharePoint 1
  (knowledgebase-1)

3. AI-Powered Answer Generation (LLM Node)

• Purpose:
  Uses an advanced AI model to generate a clear, conversational answer for the user.

• How it works:

  • The AI receives both the user’s question and the search results from the SharePoint knowledge base.

  • It crafts a brief, polite, and accurate response, citing the relevant HR policy document if possible.

  • If the AI cannot confidently answer:
    It will inform the user and automatically send an email to the HR team for follow-up.

• Node:
  OpenAI
  (llm-0)

4. Display the Answer (Output Node)

• Purpose:
  Shows the AI-generated answer to the user in a user-friendly format.

• Node:
  User Output
  (out-0)

How the Data Flows

1. User submits a question →

2. SharePoint Knowledge Base is searched →

3. AI generates an answer using the search results →

4. Answer is displayed to the user

Special Features

• Citations:
  The AI includes citations from the HR policy documents, so users know exactly where the answer comes from.

• Fallback to HR Team:
  If the AI can’t answer, it notifies the user and emails HR for human follow-up.

In summary:
This workflow automates HR policy Q&A by combining user input, SharePoint document search, and AI-powered answer generation, ensuring users get fast, accurate, and well-sourced responses—with a human fallback if needed.