Love’s Travel Stops and ICG have worked together for many years implementing back-office solutions to streamline operations. This successful partnership has yielded significant benefits for Love’s. In a strategic move to further efficiency, Love’s started an OCR project with ICG that went live in April this year. The collaboration has already produced impressive results and continues to do so.
The Solution: OCR/AI/ML
ICG specializes in creating customizable and configurable back-office technology, and the process of implementing OCR is no different. ICG’s OCR process has a 98+% extraction rate and uses automated reinforcement learning and vendor rules to save companies time and money. However, ICG’s OCR process is more than just the 98+% extraction rate.
Using Large Language Models (LLM), the ICG data capture offering can discern different documents within a submission, identify invoices versus non-invoices/supporting documentation, and automatically separate different invoices within a file. Pairing OCR with AI and machine learning (ML) allows for continuous improvement over time.
In addition to capturing the header information at exceptional accuracy rates, ICG has a built-in recursive learning model that compares data from the paid transaction within the ERP to data initially captured on an invoice. Where differences are noted, the system automatically creates vendor-based rules to further enhance accuracy – all without user involvement. These vendor-based rules are not manually created templates, but are automated rules systematically created and applied through ongoing learning.
Love’s concluded that ICG’s OCR/AI/ML would be the best fit to improve the operations and efficiency of their AP department.
The Results
Instead of only measuring the success of the extraction, ICG and Love’s needed to understand the true accuracy of the automated process. ICG determined success by comparing the data automatically extracted by the system to the data used to pay the transaction in the ERP. ICG and Love’s specifically looked at 4 key fields: vendor name, vendor number, invoice date, and invoice amount. When all 4 of these fields were extracted with 100% accuracy, the invoice was considered to be automatically indexed. The percentage of invoices that correctly extracted all 4 fields is referred to as the rate of automatic indexing.
Even though the extraction rates were above the allowable threshold of 95%, the rate of invoice keyword automatic indexing was 55.45% on 5/1. In just two weeks, that score jumped to 78.64%. It continued to climb to 80.95% just a week later on 5/29. Following this, the rate hit 83.24% and has been hovering around the 80-83% mark ever since.
Through automated reinforcement learning and vendor hints/rules, ICG’s configuration of OCR, AI, and ML can help you attain similar results – or even better. The graph below shows the percentage of invoices automatically indexed according to the 4 key fields identified by Love’s.
The significance of this increase from 55.45% to 80-83% has not gone unnoticed by the Love’s staff. “Since implementing OCR, our AP team has gained back over two hours per person, every day. It’s been a game-changer for efficiency and productivity,” says AP Manager Stephanie Garinger.
Not only is this project saving Love’s time and money, but with the extra two hours per day per person, Garinger’s team can also take on higher-level tasks and create and contribute to new department and company goals.
Conclusion
ICG’s automated capture solution is easy, quick, and economical to implement, providing the payback executives expect from their technology spend.
OCR can help your company achieve high rates of automatic indexing leading to faster and more accurate invoice data entry. To learn more about how an OCR solution can help your organization, contact ICG today. We’ll coordinate an OCR test using thousands of your invoices, showing you just how much time and money you can save by using OCR/AI/ML. We also offer many other customizable solutions and work with your company to meet your needs and help you achieve significant results. To see our other solutions, request a demo or check them out here.