Fueling Efficiency: How Pilot Uses Vector Stores, Data Quality, and GenAI to Deliver Business Value

In the complex world of logistics, efficiency and accuracy are paramount. At Pilot, the largest travel center network in North America, managing fuel delivery operations was a time-intensive and error-prone process. Tasks like processing delivery records and validating fuel transaction data posed significant challenges due to the diverse formats and handwritten elements involved. After several attempts to use robotic process automation failed, the team turned to Generative AI to automate and streamline this critical business process. In this session, discover how Pilot leverages GenAI, powered by advanced text and vision models, to revolutionize BOL processing. By implementing few-shot learning and vectorized examples, the data team at Pilot was able to increase document parsing accuracy from 70% to 95%, enabling real-time validation against truck driver inputs, which has resulted in millions of savings from accelerating credit reconciliation and improved financial operations.