The starting point for the development of our General Ledger Recommendations began with the artificial intelligence model we connected to Accenture’s BuyNow procurement solution and our SAP S/4HANA® accounts payable solution. Accenture Procurement Plus worked with CIO Applied Intelligence to develop the solution for both purchase order-based and non-purchase order-based spend. For purchase order-based buys, BuyNow needed to improve the assignment of general ledger accounts at the point of purchase requisition. For non-purchase order-based buys, Accenture needed to improve the assignment of general ledger accounts within the accounts payable process. Both situations were manual, time consuming and prone to error.

To increase accuracy and efficiency, Accenture Procurement Plus and CIO Applied Intelligence designed an artificial intelligence model to recommend the best account from more than 5,000 available general ledger accounts. Drawing from the accounting on historical purchases and invoices, the model recommends the general ledger accounts for new buys at the point of entry. The solution was directly embedded into the Accenture BuyNow platform experience.

From a technology perspective, CIO Applied Intelligence approached the problem as one of classification. First gathering historical purchase and accounts payable activity, the team ingested the data into Accenture’s data lake. The team then used the data to build and train a General Ledger Recommendation model with an AI machine learning algorithm. Using the patterns created on the historical data, the model was also able to make a predictive recommendation of a general ledger for a purchase.

Once the model was trained, the team embedded the solution into both the Accenture SAP S/4HANA accounts payable process for non-purchase order-based buys and the BuyNow procurement process for purchase order-based buys. The model continues to learn and retrain itself to increase its accuracy in making classifications for new goods and services buys.