Machine Learning: A Clear Case For Automating GR/IR Clearing

Paul Kurchina

The more companies become digital and global, the more finance organizations need to step up their game. It’s just not enough to report numbers, pay the bills, and secure steady cash flow. They must also deliver self-explaining, reliable financial information that provides the full story of how the business is performing in real time. But all too often, many financial teams struggle with isolated business systems full of inconsistent – and sometimes conflicting – information.

From the perspective of Smitha Chowdavarapu, senior manager at Deloitte Consulting LLP, CFOs should consider process automation to serve the business more efficiently. Finance teams have access to such a massive volume of data that they are well-positioned to use digital technology – such as machine learning – to analyze and interpret information more quickly and accurately.

One critical area that can benefit from machine learning is goods-receipt (GR) and invoice-receipt (IR) clearing. During the Americas’ SAP Users’ Group (ASUG) Webcast “Making the Future of Finance Real with Machine Intelligence: A GR/IR Use Case,” Chowdavarapu shared:

“Finance teams are under tremendous pressure to clear GR/IR by month- or quarter-end to represent the business’s liability correctly. If goods receipts are not consumed by invoices, liabilities are overstated – resulting in inaccurate financials for the organization. A mechanism to automate the clearing process could accelerate the entire close process – whether it’s at month’s end, quarter’s end, or year’s end.”

Simplifying the triple match: POs, goods receipts, and invoices

Even though GR/IR cleaning is one of the most manual, time-intensive processes in finance, many organizations rarely have the time to determine a better way to accelerate it. Finance organizations often spend an extraordinary amount of time sifting through data and compiling reports manually – all while pulling together purchase orders (POs), goods receipts, and invoices scattered across a wide variety of business applications. And the process is further complicated with stored data that is potentially erroneous, out of date, and incomplete.

Applying machine learning to this process allows finance teams to intelligently automate the matching of more than 75% of manual clearing items. “With machine learning, finance will still have to undergo semi-automated clearing when there are exceptions that need to be reviewed and considered,” said Chowdavarapu. “However, the volume of that manual work is greatly reduced because POs, GRs, and IRs stay within the processing rules predefined by the business and from which machine learning operates. And it is this benefit that allows CFOs to take finance to the next level of strategic thinking.”

Machine learning introduces several potential advantages that help not only accelerate, but also improve, GR/IR clearing such as:

1. Real-time analysis

Analyze the GR/IR information quickly and, depending on the confidence level of the machine learning model, derive insights automatically. Since data is not constant for most business processes, the algorithm can learn the data much quicker than humans.

In the GR/IR process, this is even more helpful, since the machine learning model learns the data provided to better identify patterns and any anomalies that may exist in order to continually improve the prediction rate.

2. Visibility into trends across periodic cycles

Pinpoint trends – large and small – that are hidden deep in the data or go unrealized due to talent or human limitations. Machine learning helps eliminate human error, improve the quality of output, and bolster cybersecurity.

The GR/IR scenario clears any missed POs, GRs, and IRs. Plus, it improves implementation and execution practices by more effectively leveraging insights into the company’s procurement processes and behaviors driven by data.

3. Reduced burden of manual analysis

Free finance teams from low-value, repetitive tasks so they can focus more on the strategic needs of the business. Machine learning can automate and prioritize routine decision-making to help achieve outcomes sooner.

For the GR/IR process, this level of automation clears any GR and IR with high accuracy by relying on common scenarios defined in the program. More importantly, the process takes much less than the average time it took to clear them manually.

4. Fast and efficient soft closings

Speed up periodic financial closings to give business leaders and other decision-makers the insight they need to make investments that are optimal and complementary, as well as leverage the right partner, vendor, and supplier relationships. Period-end closings are more efficient since clearing can be done in real time and manual interference is minimal.

Developing a sophisticated method for GR/IR clearing with machine learning

Machine learning presents a data-driven opportunity for higher efficiency, error elimination, faster action, and better decision-making. And this case can be made for finance, as well as the rest of the business ecosystem. But what makes the implementation of the technology unique to finance is knowing where and how to apply it.

Given the range of exception handling, regulatory requirements, and intense scrutiny that happens within the finance organization on a daily basis, some form of human intervention will always be required. However, even if the smallest task in a more-extensive process is automated, CFOs can free their talented staff from energy-consuming activities that yield little strategic benefit to deliver significant outcomes that can help shape the future of the business.

Machine learning is no longer just for smartphones or game shows. See Why Machine Learning and Why Now? to develop a strategy that will change the basis of competition in your industry.

Paul Kurchina

About Paul Kurchina

Paul Kurchina is a community builder and evangelist with the Americas’ SAP Users Group (ASUG), responsible for developing a change management program for ASUG members.