Problem solved

HR departments frequently handle repetitive employee inquiries about company policies, benefits, and administrative procedures. Manually responding to these queries is time-consuming and prone to inconsistency. This tool automates the retrieval of accurate policy information using a vectorized knowledge base and can even take follow-up actions like sending confirmation emails. It streamlines HR operations, reduces response times, and frees HR staff to focus on more strategic work.

Use cases

• Employees asking about leave policies (sick, vacation, bereavement) and receiving immediate, sourced answers.
• Automatically sending an email to HR with a summary of the employee query and the bot's response for record-keeping.
• Handling multi-step requests, such as looking up a policy and then emailing a manager for approval.
• Reducing HR ticket volume by providing a self-service knowledge base accessible via chat.
• Assisting HR staff during busy enrollment periods by answering common questions about benefits and deadlines.

Audience

• Developers
• ML Engineers
• Data Scientists
• Enterprise

Pros

• Demonstrates effective use of LangChain agents and the ReAct pattern for complex, multi-step reasoning.
• Combines document retrieval (via Chroma) and external API actions (Gmail) in a single conversational interface.
• Chain-of-thought explanations improve transparency and reduce hallucinations.
• Human-in-the-loop confirmation for sensitive operations like email sending adds a safety layer.
• Modular design allows easy extension with new tools or policy documents.

Cons

• Not production-ready; lacks authentication, multi-user support, and comprehensive error handling.
• Depends entirely on the quality of the underlying LLM; may still produce incorrect answers if not carefully prompted.
• Requires manual setup of multiple APIs and local infrastructure, which may be a barrier for non-technical HR teams.