“It’s one small step for man – one giant leap for mankind.” With these simple words, Neil Armstrong became the first man on the moon in July of 1969 launching a new era in the exploration of space. We are on a similar critical junction in the world of machine learning and artificial intelligence applications. Consider these metrics from a recent article in Forbes:
- Machine learning patents grew at a 34% Compound Annual Growth Rate (CAGR) between 2013 and 2017, the third-fastest growing category of all patents granted.
- International Data Corporation (IDC) forecasts that spending on AI and ML will grow from $12B in 2017 to $57.6B by 2021.
- Deloitte Global predicts the number of machine learning pilots and implementations will double in 2018 compared to 2017, and double again by 2020.
One area in the vanguard of this revolution is Financial Management, Accounting and Controlling. According to most CFOs, nothing in their financial accounting system is more important than CASH, so it just follows that the areas attracting some of the most intense scrutiny for machine learning are those that impact CASH. Not only to answer questions such as “How much cash do I have on hand?” or “Which of my bank accounts have the greatest cash balance?” but rather, “How can I accelerate the velocity of my cash flow?”. This latter question gets to the real question of cash management. In essence, the CFO typically wants to accelerate his cash receipts and to stretch out his cash payments in order to a) maximize float and b) optimize his invested cash. In a world of 24-hour financial markets where the movement of huge volumes of cash can result in real financial earnings, this is a key concept.
You have probably seen Sven Denecken’s blogs on “Driving Efficiency and Automation in Finance Through Intelligent ERP”, but I’d like to take you one layer deeper here. Consider Remittance Advices. These are essentially notices from a customer that “the check is in the mail”. If we could actually automate our receipt of these advices and then mechanically read the remittance advice, our speed of processing this information would not be limited to the pace at which our accountant or cash manager could:
- Open the advice (if the advice comes via snail mail, add time for waiting for the daily post)
- Determine who the advice was from
- Confirm that this is indeed one of our clients
- Scan a paper form and then run an OCR application against it to digitize the remittance advice
- Extract the information
- Load it – correctly and without errors – into our cash management system
- And then hand it over to the Accounts Receivable team
Then the A/R team will need to run their processes against the remittance to determine if:
- Which invoice the remittance matches up to
- If the remittance has already been paid by checking my bank account
- If the 3-way-match is accurate
- If it is not correct, how do we handle disputes?
Clearly, both of these processes – remittance advice extraction and the application of cash to our invoices – meet most company’s criteria of a Labor-Intensive Process that could be automated. Further, the automation of these processes also offers a tremendous opportunity to accelerate cash collections – a critical demand from the office of the CFO. If we look at Remittance Advice extraction, consider how #MachineLearning can be deployed:
Image of Remittance Advice Extraction
By deploying such a machine learning application (not coincidentally currently available from SAP today!), financial accounting, treasury and controlling workers can be freed up to be more than knowledge workers. They become decision workers.
So, you see…the deployment of a remittance advice extraction service using machine learning is not only one small step for machine learning, but one GIANT LEAP for cash management!!
- SAP Managed Tags:
