| Abstract |
This capstone project will present the development of the AI-Powered HR Analytics
and Automation System created in collaboration with the AHRD. The system will
integrate modern data engineering, machine intelligence, and automation of HR
workflows that will replace most of the time-consuming manual processes with a
digital, scalable solution. It uses Supabase as a secure cloud database, Python as a
programming language for extracting and preprocessing data, GPT-4.1-mini for AI
cleaning, storytelling, recommendations, and email generation, Streamlit for
interactive dashboards, and knowledge-driven Botpress to drive an HR chatbot. This
solution will deliver an end-to-end pipeline that can transform raw HR data into
actionable insights.
This project addressed several challenges related to inconsistent Arabic text that
scattered non-standardized job information, and poor visibility into HR patterns, by
combining a hybrid cleaning strategy using both the rule-based methods and the AI
normalization, data accuracy significantly increased from 81% to 97%. The next
section will highlight the dashboard that covers a comprehensive workforce,
demographic, structural and the geographic analyses while the AI Storytelling
generates narrative explanations for easier interpretation. The automated alert
subsystem detects HR policy conditions for sending official AI-generated
notifications. Supported by the anonymized workforce data and the official HR
documents through a Retrieval-Augmented Generation the chatbot could support
policy-aligned, context-aware responses. Overall, the PoC demonstrated how the AIdriven
automation can improve decision-making to enhance HR operations and
underpin the Ajman Government's digital transformation vision.
|
