Case Study: Posti achieves safer invoice automation with Aito.ai

Posti flags invoice anomalies with Aito.ai using 97% product-match confidence

Posti, a Finnish postal service, sought to enhance the safety and reliability of its accounts payable automation processes. The challenge was to automatically identify potentially fraudulent or anomalous invoices within its purchase-to-pay system to prevent incorrect payments without enforcing overly strict rules. The company partnered with vendor Aito.ai to leverage its machine learning platform for implementing intelligent controls.

Aito.ai's solution used Posti's historical invoice data to create three AI-driven controls that predict anomalies. These controls assessed whether the products on an invoice were typical for a specific vendor, if the invoice amount was consistent with historical data, and if the invoice frequency was normal. The implementation, which required no data science expertise, allowed Posti to automatically flag high-risk invoices for manual review. This resulted in safer automation, with the ability to identify improbable vendor-product combinations with less than 1% certainty and to confirm legitimate ones with over 97% probability.

Posti