Straight-through processing (STP) has been a vision and a goal for finance operations for decades, particularly for high-volume transactional processes like accounts payable invoice processing. STP allows invoices to be received, reviewed and approved without manual effort to key in data, route it for approval and create ERP transactions. In the past, where it was even applicable, STP was only achievable and practical for large companies with a low volume of exceptions to their business rules.
For businesses that historically experience a high volume of exceptions due to errors or non-standard processes, the burden falls on the AP staff to manually process everything—which ironically, produces more errors and exceptions due to process gaps and the inherent unreliability of manually transferring information from a document to a database by relying on a keyboard and mouse.
Process bottlenecks related to these tedious, manual processes account for the largest pain points for AP teams, according to a recent IDT survey, which found that 76 percent of professionals cited manual data entry and document handling, plus invoice process exceptions as major obstacles.
Today, however, the challenges of the past have been addressed by two significant innovations: machine learning and integration as a service platforms.
Machine learning technology has allowed companies to address the problem of process exceptions in a new way that doesn’t feel uncomfortable or disruptive to process owners. Considering that it is usually a cultural shift or a required behavioral change that undermines process changes, this is huge. Here’s the difference. In the past, if a company was able to address exceptions discovered in a given process workflow, they had two options. One was to change/modify the business rules to accommodate the exception. When that was possible, it required coding or intervention from IT. This often proved to be too costly or too complex over time. The other option was simply to identify and flag exceptions, which were then routed to human operators for handling (back to the same old problem). Machine learning systems applied to processes like AP invoice handling, flags exceptions for humans to handle. But it only asks them to do so ONCE. Systems like Artsyl’s docAlpha intelligent process automation platform, records and learns from that human intervention, automatically updating its knowledgebase and modifying its algorithm so it can handle the same exception on its own next time. So, no more need to change code, modify business rules or involve IT.
Integration Platforms as a Service (IPAAS) is a new approach to systems integration that results in far more flexible and adaptable communications between platforms like intelligent process automation systems and ERP or CRM systems. The big different is that upgrading or modifying one system doesn’t break the integration. And changing/modifying it (wait or it) doesn’t require complex coding or even significant IT involvement.
As a result, intelligent process automation systems can automatically cross-check records in these systems to ensure accuracy. Find an exception? You may have to bug a human. But far less often than the old way of doing things. And over time, that intervention diminishes. More flexible, intelligent integration means that today’s systems can more reliably ingest an invoice or sales order, extract the right information for routing or matching, validate that information against other systems, secure coding/approvals from human stakeholders and then automatically create the appropriate approved transaction records in the ERP/ECM/other system.
With these new tools and technologies available, affordable and easy to implement, STP can become the rule rather than the exception, and the era of hands free ERP may not finally be upon us.
In our opinion, it’s about time. You don’t hire accounting team members to do manual labor. Let the robots do the heavy lifting. And let your team free their minds for more valuable work.