This blog is the second of a series of four blogs about disruption in the digital age. Read Part 1.
Software robots that imitate the way humans interact with software are increasingly being used to automate routine business processes. And, as a recent Deloitte University Press report points out, with good reason: Automation can significantly reduce the effort involved in many processes.
Consider the clear account accrual/reversal process that needs to be done at each period close in finance. A listing of all general ledger clearing account balances must be obtained. Then someone needs to identify any accounts with a balance greater than zero. Each balance must be corrected one-by-one and reversals posted for the next period. It’s repetitive, time-consuming and mundane – and a perfect target for automation. Not only can the process be sped up, but accuracy can be improved because robots are less prone to human error.
As these examples demonstrate, repetitive, rules-based tasks that are usually performed by people sitting in front of computers are ripe for automation. Robots can open e-mail attachments, complete e-forms, record and re-key data, and mimic many other human activities. As a result, they can lower costs, reduce error rates, enhance service levels, shorten turnaround times, increase the scalability of operations, and improve compliance.
Cognitive process automation
However, by its very nature, robotic process automation can be used only to a limited extent in complex scenarios that require human judgment. But with advances in cognitive technologies such as natural language processing, speech recognition, and machine learning, that is changing.
Cognitive process automation has the ability to go beyond robotics to deliver better business outcomes such as greater customer satisfaction, lower churn, and increased revenues.
For example, a U.S. bank has used a number of cognitive technologies to automate its billing system. By replacing the manual, paper-intensive, error-prone multilingual process with an automated, executable business process workflow, the company uncovered revenue leakage of about 10% and recovered about 4%.
At a global information services firm, a cognitive platform is being used to handle the seven million annual faxes it receives from clients. The platform converts document images into machine-readable text and then uses user-defined rules and machine learning capabilities to extract the data and categorize the faxes accordingly.
Increased computing power has made cognitive computing possible, as has Big Data. The enormous data sets provided by various sources – for example, text, images, and geospatial data – are the basis for training machines and allowing them to learn.
In addition, basic research in machine learning has led to more sophisticated learning algorithms and a better understanding of the basic principles of learning itself. In turn, this is allowing machines to acquire capabilities such as seeing, reading, writing, listening, and talking.
Machine learning has huge potential. As such, it is a strategic topic for SAP. And the Machine Learning Incubation team of the SAP Innovation Center Network is working on different machine-learning use cases to make SAP applications intelligent. SAP is also collaborating closely with its large partner ecosystem to bring about the cognitive enterprise.
Find out more
- Read the full Deloitte University Press report, Robotic process automation. A path to the cognitive enterprise.
- Read the first blog of the series about disruption in the digital age: Digital Disruption: A Four-Step Approach.
- View the SAP news story, Machine Learning: Go for the Intelligent Enterprise.