PepsiCo, Inc. is one of the world’s largest Multinational Food and Beverage companies, with its products sold in 200+ countries, 280+ production facilities, 290,000 employees, and a net revenue of over $86 Billion.
PepsiCo launched a new app loyalty rewards program in select markets designed to drive loyalty among consumers to purchase more PepsiCo products. Developed for select markets around the world, the Loyalty App allowed consumers to incur ‘loyalty points’ on scanning their receipt whenever they purchased a Pepsi Product from a grocery store.
PepsiCo wanted to solve multiple challenges, the biggest being potential fraud. It is usually hard to determine if consumers manipulated the receipts in order to garner more loyalty points. For instance, customers can go to a grocery store and purchase a case of Pepsi Max. Adding just another zero on their receipt would allow them to incur 10X more points. This was a challenging use case, since this involved different languages and matching/reconciliation at the back end.
To tackle this problem, PepsiCo chose Kognitos. Kognitos leveraged Optical Character Recognition (OCR) and Large Language Models (LLMs) to create a single process to extract information from scanned receipts and detect any signs of fraud. This was enabled using Kognitos’ ability to do document understanding, handwriting detection and ensuring that the line items in the receipts add up to the expected totals. The Kognitos platform also enabled the business analysts to change the fraud detection logic rapidly and dynamically as the logic was expressed completely in natural language. And whenever fraud is detected, the Kognitos platform sends the data to a human user to validate and make a decision.
PepsiCo was able to reduce their costs by 70% on this initiative.
Kognitos also enabled PepsiCo to reduce their BPO spend by 6 figures and deliver loyalty rewards programs in underserved global markets.
Furthermore, the platform gave PepsiCo the ability to scale to 2 million receipts a year.
The Numbers: 2 Million (Receipts Analyzed in a Year)
70% (Reduction in BPO Costs)