HR teams at growing businesses face a recurring problem: new hires and employees ask the same policy questions repeatedly — leave entitlements, onboarding steps, expense procedures, IT setup — pulling HR professionals away from work that actually develops people.
Designed and developed a conversational AI assistant that answers HR policy questions, guides new hires through onboarding steps, and surfaces the right document at the right moment — trained on the company's own handbooks, policies, and SOPs using Retrieval-Augmented Generation.
The assistant integrates with Slack and Microsoft Teams so employees ask questions in the tools they already use. When a question falls outside the knowledge base, it escalates to HR with full conversation context — so no question goes unanswered, and HR can identify gaps in existing documentation.
HR documents, policies, onboarding checklists, and SOPs are ingested, chunked, and indexed into a vector database. At query time, the assistant retrieves the most relevant passages and uses an LLM to generate a grounded, plain-language response — with citations pointing back to the source document so employees can read the full policy if needed.
The platform includes an admin interface for HR teams to upload and update documents, review escalation history, monitor usage, and identify frequently asked questions that may indicate gaps in existing policies. Role-based access controls ensure sensitive HR content is only surfaced to authorised users.
The assistant significantly reduced the volume of routine HR enquiries handled by the HR team, freeing them to focus on employee development, hiring, and strategic initiatives. New hire time-to-productivity improved as employees could get instant, accurate answers to onboarding questions at any hour. HR teams gained visibility into what employees were asking, informing ongoing improvements to policy documentation.
Project Summary
Download the one-page PDF overview of this project.