Rillion AI Matches Complex Invoice Lines to Purchase Orders

Rillion, a provider of financial management software, has launched a new artificial intelligence feature within its Rillion Capture platform aimed at resolving difficulties in matching invoice lines to purchase orders (POs). The issue arises when suppliers consolidate multiple product details onto a single invoice line, deviating from a more detailed PO structure.

This common scenario, referred to by Rillion as the "T-shirt case," often necessitates manual intervention, slowing down the procure-to-pay cycle. The new AI capability allows users to provide simple, plain-language instructions to Rillion Capture, guiding how it should interpret and restructure non-standard invoice lines. For instance, an instruction like "If a line contains multiple variants, such as different sizes of the goods, break it into multiple lines" enables the system to generate separate invoice lines for each variant.

Traditional OCR (Optical Character Recognition) systems typically struggle with such inconsistencies due to their reliance on predefined templates and fields. Rillion Capture, however, leverages AI to comprehend invoice content and reorganize the data to align with the PO structure. This approach aims to reduce erroneous automated matches and the need for manual corrections.

The feature can be applied to various non-standard formats, including those where multiple colors or products are listed on a single line. Instructions can be configured globally or on a per-supplier basis without requiring template modifications or custom development. Andreas Jungmar, Product Manager at Rillion, likens the process to explaining interpretation rules to a new colleague.

The introduction of this AI functionality seeks to decrease manual workload, enhance data accuracy, and increase the rate of automated PO matching for finance departments.