AI Transformation Across Alpino: HR to Inventory Management

Abstract

Alpino Health Foods is a rapidly growing D2C company in the Indian health food sector. But while scaling up, three major bottlenecks developed in the organization affecting productivity, decision-making and profitability. The problems identified are high discrepancy and delayed reconciliation in finance transactions, high early attrition and ineffective resume screening in HR and stock sites during high marketing campaign periods impacting ROAS. These issues needed an integrated system that can improve efficiency with the help of AI. The project was selected because there was immense pressure on internal operations of Alpino that were very heavily manually driven or semi automated. In the finance department, ~10000 transactions happened each month and relied on manual invoice checking, while HR was interviewing and hiring 3-4 employees every month and spending heavily on marketing campaign without a strong forecast model on demand which made Alpino a good candidate for studying how AI can improve operations of a mid sized FMCG company. The capstone project was implemented as "Alpino AI Operating System". The module in Finance, automatically extracts invoices from inputs and validates the Gst amount by using OCR and custom validation rules that helps in reducing reconciliation delays. The module in HR, predicts an employee at the risk of leaving the organization with the help of xgboost and predicts candidates to be interviewed by the usage of bert that compares semantics of the resumes and provides matching candidates for recruitment and reduces candidate screening time. While the module in marketing helps in improving campaign customer segmentation and reduce stock out during high demand by creating a better forecast of demand to impact and improve ROAS, was implemented using Streamlit with synthetic data similar to the business case of Alpino. The project has been successful to address three business problems. The finance module has better invoice validation, and quicker reconciliation of payments, whereas the HR module helps identify at risk employee, along with a quality selection of candidate, while the marketing module enables for better customer segmentation with improved demand forecasting that reduce stockouts and improved ROAS, thus it was evident that AI can improve the decision making of operations, and reducing manual work load, for a sustainable growth. The study concludes that the implementation of AI is feasible and can generate positive ROI at Alpino. The investment made would pay off in 12-18 months of operation that would estimate the investment as 25-40 Lakh and will prove that even low cost implementation can lead to a larger improvement at a growing D2C company. Hence, it is possible to create such a system that may help to create such a model for other FMCG companies.