Mistakes happen, but sometimes they can be quite expensive for a company. Intelligent technologies make it possible to relieve employees of error-prone tasks and save organizations from unwanted costs.
After the exceptional year 2020, companies have already started early to calculate their losses for this year. In addition to quarantine and remote work, which will continue to be with us, mistakes made by employees are also a regular cause. Due to lack of experience, lack of attention, or fatigue, for example, an employee may fail to complete a task on time, send a document to the wrong colleague, or make a typo.
Human errors and their consequences
At first glance, these moments of negligence don’t seem dramatic, but sometimes such undetected oversights add up to millions in losses for a company. Here are three vivid examples that underscore this:
- In August 2020, due to a small clerical error, employees of Citigroup’s credit department accidentally transferred nearly a billion dollars to Revlon Inc.’s lenders, who had been involved in a legal dispute with Revlon for many years and had almost lost hope of receiving that money. And, of course, they were unwilling to give it back. Citigroup then filed suit.
- In October 2020, nearly 16,000 coronavirus cases went unreported due to an Excel error by Public Health England (PHE). The problem: PHE’s own developers chose an old file format – known as XLS – to do the job. Each template could only handle about 65,000 rows of data, instead of the million-plus rows Excel can actually handle. And since each test result generated multiple rows of data, in practice this meant that each template was limited to about 1,400 cases. Once this number was reached, further cases were simply not recorded.
- And even in Germany, errors occur in companies that can be costly. For example, an architect’s office made a typing error when calculating the cost of structural steel. The result: additional costs of almost 800,000 euros.
What do these and many other errors have in common? First, they are related to reading, processing, analyzing and capturing documents. Second, before the error occurs, employees repeat the same operations over and over again. Third, in many situations, data is entered manually. For example, an employee copies numbers, enters customers’ last names into a table, performs calculations of financial results, etc.
At first glance, it looks like we live in a digital age today and such things should be an exception in business processes. But when you have to double-click the same area on the screen dozens or even hundreds or thousands of times, use Ctrl + C and Ctrl + V, or check the box in a database for the repeated time, it is not surprising that on the 1001st time an unintentional mistake happens. If other factors for distractions are then added, for example, because a person is working from home, the risk of error increases even more.
Work routine is not only bad for business, but also for the employees themselves. It is a cause of frustration, and poor operational processes can even lead employees to consider changing jobs, according to a recent study by ABBYY.
The call for robots
Entrepreneurs who are interested in technology and understand its value to the business know that humans are not robots and repetitive tasks can be done better by machines. In manufacturing, this was realized some time ago, but in offices, the first software robots moved in only recently. Today, we see a growing number of RPA, process and intelligent document processing projects around the world. According to Gartner, the RPA market will reach nearly $2 billion next year and continue to grow at double-digit rates through 2024.
So where do we go from here? For example, if the chief accountant wants to restructure his department, the purchasing manager may suggest that tender documents should be digitized first to select suppliers more quickly. For the compliance manager, however, regulatory requirements for automation take priority for the company. These are subjective opinions. Can they be backed up by quantitative data? What can be automated quickly? What is more relevant to automation, i.e. whose manual work is more expensive and how can we measure that?
Process understanding is essential for planning
An employee survey is a classic method for obtaining information about business processes. This approach is useful for becoming familiar with processes and preparing new automation projects. Sometimes such surveys are conducted by consulting agencies because the expert judgment of an external consultant seems more objective. In large companies, intelligent automation is such a large project that they set up centers of competence – so-called “Centers of Excellence” – whose main goal is to identify business requirements, evaluate the projects and set priorities for automation projects. However, employee surveys are often influenced by bias and skewed responses.
Process Discovery and Process Mining
There is a smarter way to identify, prioritize and accelerate automation initiatives. In recent years, companies have begun using process discovery and process mining solutions to make smarter decisions. This includes the use of intelligent platforms that collect data from the company’s information systems and create a process map, break it down by phase, include interaction data from desktop users, identify common errors, redundancies or deviations and prevent them. They also analyze the reasons why certain processes are running slower than they should.
This process insight allows organizations to take a bird’s eye view of their entire operations: they can connect systems, people and data to understand exactly where automation is best suited, enables effective digital transformation and minimizes human error. The need to “customize” the process so that it can be understood not only by a human, but also by a machine, is one of the reasons for the growing use of process mining.
For example, a global telecommunications company integrated a process mining platform into its key information systems. The company identified nine processes to automate, using not only RPA but also chat bots and machine learning. The analysis and data processing took just six weeks and saved the company an estimated $8 million per year.
Humans will always be an important part of the workflow. However, managers can ensure that errors are minimized by equipping the company with more digital intelligence solutions.