Problem

Manual GL coding drained AP resources and delayed processing

Manually processing over 35,000 non-PO-backed invoices each year was a significant challenge for the AP team, even with 13 FTEs, many of whom were contractors. Every scanned document image emailed from all 260 locations needed to be manually reviewed, creating a bottleneck within the approval workflow. Eliminating the manual capture of these documents was critical. The company began looking for ways artificial intelligence (AI) could remove the need to scan invoices and physically review every image.

Beyond automated invoice reading, the AP team also needed a system to learn and apply their complicated general ledger (GL) codes and cost center structure. Manually assigning the correct entity for every invoice from every location was difficult. Reclassifying invoices took a lot of time, as they corrected errors and mistakes. TruGreen also wanted to consolidate the many existing tools and software systems. The company was concerned about the time needed to train employees and the risk of a loss of institutional knowledge if turnover took place.

The AI can learn when a transaction was changed and why it was, and then it can make that change for us and remove that activity from the process.”

 Director  of Corporate Systems at TruGreen

Solution

A frictionless integration for automated GL coding and non-PO invoices

TruGreen saw the transformative power of Autonomous AP’s AI and its autonomous, or completely hands-free, processing as the key to reducing manual work. The AI’s finance domain knowledge and ability to learn continuously without leaving the workflow gave the company confidence it could increase the speed and accuracy of its invoice processing. It was already saving time and effort with AppZen’s Expense Audit, which had saved the company 550 hours of work, caught $356K in out-of-policy expenses, and reduced spending by $22K in the last year alone. The company was counting on AppZen as a trusted partner to do the same for its accounts payable process.

As soon as TruGreen integrated Autonomous AP with its JD Edwards ERP, the solution immediately began reading invoices and images to find location and supplier information. Using transaction histories and the company’s chart of accounts, it accurately assigned GL codes and cost centers. Soon, the company was automating the end-to-end processing of over 40% of non-PO invoices, saving hundreds of work hours.

With AppZen, we can read documents and images to reference location data, invoice items, and supplier information, then look at our transaction history to propose the GL coding. This eliminates a significant amount of manual work.


Director of Corporate Systems at TruGreen

 

Result

70% GL automation with $1.6M in risk and duplicate detected

Nine months after launch, AppZen's technology is fully automating 60% of TruGreen’s invoice volume with high accuracy. The remaining invoices take just 2-3 minutes to process. Autonomous levels are still rising, thanks to our fast-learning AI. Approximately 70% of all GL code assignments are now autonomous, as well. Consolidating platforms eliminated even more manual effort. With no need to touch nearly two-thirds of the transactions, the company has realized significant savings in reducing FTEs. It now takes only 3 dedicated employees to handle all invoice processing. There’s plenty of room to scale while focusing on key business tasks.

In addition to delivering high levels of autonomous processing, AppZen’s sophisticated AI models provide robust auditing across aggregated invoice and T&E spend channels, uncovering more savings. Of nearly 20K invoices, Autonomous AP flagged 2K high-risk transactions, found $785K in overall high-risk spend, and removed $870K worth of duplicate spend. Its dashboard analytics can pinpoint which suppliers send the most duplicates or high-risk invoices, improving supplier compliance.