How Data Errors Thwart Firms’ Automation Goals

Automation of business processes can quickly get derailed if the data that a system is working with is incorrect. Take B2B trade, for instance: A company’s automated invoice and purchase order management technology can’t do much if the error that was placed within the system is erroneous, as personnel must then step into the business process and solve that data error manually.

So, while corporations have taken steps to automate their systems, Brent Halverson, president and CEO of cloud-based business solutions firm ecmarket, says that’s no longer good enough.

“We see a trend of businesses moving away from semi-automated workflows to a fully automated workflow,” he told PYMNTS in a recent interview. “You need to understand the data elements of what’s causing orders or invoices to have someone have to look at it.”

His company operates Conexiom, which automates processes for B2B organizations, like invoicing. Conexiom has recently focused on data management with its services. Last year, for instance, the firm rolled out Invoice Automation to help manufacturers and distributors capture invoice data for their accounts payable departments.

And earlier this month, ecmarket and Conexiom announced a partnership with DATAgility.

“Conexiom can identify where the problems of data are, while DATAgility can help cleanse that data, so people can get more automation out of invoicing or sales order automation,” Halverson explained of the collaboration.

As more corporations, large and small, want to automate business processes in the accounts receivable and accounts payable departments, they’re beginning to understand that full automation can be a challenge. Halverson identified some of the ways data errors can occur.

For example, he said, product descriptors and pricing can be wrong when reconciling between a purchase order and invoice.

“One of the biggest challenges is pricing discrepancies,” he explained, “where someone sells based on a contract price in their own ERP system, but the price from their customer is off.”

Similar discrepancies can occur on inbound purchase orders on the seller’s side. “Companies now are really looking to automate these inbound purchase orders,” Halverson said. “They want to get full automation, and one of the things they don’t want to do is have people touch the purchase order.”

But that level of automation can be difficult to achieve. The CEO said he was recently speaking with one client that experiences 50 percent of inbound purchase orders that need to be handled by a customer service or inside sales representative for review. Part of the problem is that automated solutions may not be able to go through data on a purchase order on a line-by-line basis, said Halverson. So, when pricing or a product number is wrong on a single line item of a document, the issue can’t be handled automatically — putting processing of the entire invoice or PO on hold.

Often, too, this situation requires a lot of back-and-forth communication between buyers and suppliers via email or phone. Today, explained Halverson, many businesses are looking for solutions that automatically send a document with erroneous data back to the sender — instead of having the recipient initiate the correction process.

“I think what we’ll start to see is that people will start to push back the problem to people that send the information to begin with,” he predicted.

While the executive noted that “world-class” companies are often educated and aware of not only issues of data errors but a need to correct the problem, there are other businesses that need to take a new approach to their automation efforts. He spoke of another prospective client that was looking to integrate an optical character recognition (OCR) solution to automate data capture, something Halverson said works best on images — not for PDFs and emails. And while the prospect believed that about 80 percent of data on documents came into the company via fax (a good candidate for an OCR solution), Halverson suggested that the business implement some type of tool to know for sure how much data comes in from the company across channels.

“They did, and they were surprised to find that 80 percent of the orders they got came by email, and only 20 percent were from fax,” the CEO recalled. “It’s like companies still live in the past a bit, and they don’t think through the trends that are happening. The pace of change is so fast, and I’m not surprised that some companies have to spend the time to think about it.”

Still, he said, businesses have their fingers on the pulse of automation. While that’s a start, organizations need to keep an eye on data quality within areas like procure-to-pay, accounts payable and accounts receivable. The service providers that offer solutions in these spaces, along with their corporate clients, should be more progressive, added Halverson.

“From the people I talk to, there has got to be a lot more innovation,” he stated. “And I don’t think the leadership of a lot of these companies understand the challenges that are out there.”

Even for the companies that do understand, the hurdles to achieve top data accuracy and quality are tall.

“They still have big challenges to analyze that data,” Halverson noted. “There is a lot more work to be done in this space of analyzing data.”