AI-Powered HR Assistant for Grievance Redressal, Policy Guidance, Onboarding, and Attrition Risk Management

Organisations require HR solutions that are not only reactive but also transparent and proactive throughout the employee lifecycle. This paper introduces an integrated HR assistant merging sentiment-aware grievance analysis, authoritative policy advice, a practical onboarding suite, and explainable attrition risk scoring into a unified modular platform. Grievance management makes use of NLP to identify sentiment, categorize types of complaints, and expose priority suggestions. Policy direction uses a Retrieval Augmented strategy to provide grounded, auditable snippets from the HR policy handbook. Onboarding module creates customized 30/60/90-day plans, conducts mentor/buddy matching through embedding similarity, suggests microlearning material, and assists with regular new-hire pulse checks to track early sentiment. Attrition risk is quantified with supervised models and survival analysis; model predictions are combined with SHAP-based explanations to render feature-level drivers understandable to HR stakeholders. The design prioritizes a human-in the-loop workflow for escalation, privacy-conscious persistence for pulse data, and modular components that are easily integrated with company documents and LMS systems.This article illustrates how RAG, explainable ML, and conversational UX can all together promote HR from reactive casework to active, evidence-based employee care.