RPA meets cognitive capture

Robotic process automation (RPA) has been getting a lot of attention from CEOs and investors recently. For example, Blue Prism, a leader in that arena, is listed on the U.K. AIM market with a valuation of 35x revenues.

Capture and RPA are synergistic. RPA automates keystrokes and tasks. Cognitive capture extracts data and interprets it to facilitate automating business processes. Both areas are designed to reduce or eliminate the use of humans to enter data. Both reduce the cost and time delays associated with manual key entry.

One big advantage of RPA is that users do not need to spend a lot of time configuring it, because the “robot” records how a human operator acts, which it can then play back. It can work without having to reinvent processes and is at its most powerful when automating repetitive tasks.

Robotic workflow

Derived from a mixture of screen scraping and macros, RPA eliminates key entry through building a robotic process that records the movements and key entry operations of an office worker. The intent is to minimize coding in creating a robot workflow with rules that can be automatically invoked. One major attraction is that users should not have to change their backend processes—many of which were created around mainframes in the 1960s and 70s. User companies then expanded those systems with distributed PC front-ends that provided interfaces to external services that invoked calculations often based on Excel. The result is that many office workers spend their time cutting and pasting from one area to another, retyping cells from one spreadsheet to another or typing selected data from a form into a spreadsheet.

Prompt services

The time spent on key entry by white-collar labor in the United States continues to increase with an average document costing around 63 cents to key data from, according to our research. It adds up. In manufacturing alone, we estimate that more than $1 billion a year is spent on key entry. Much of that is keying from one spreadsheet to another, and the key entry costs continue to spiral up. When you add in the time to access and verify results from a supplier or other portal, the costs multiply.

In the case of customer relationship management (CRM), customer engagement management (CEM) and digital experience management (DEM), it is increasingly critical to process incoming account openings, loan applications, claims and queries quickly and promptly. Customers do not have patience and will use and recommend the supplier who handles their needs quickly and efficiently. Manual data entry just cannot keep up—particularly as we consider multimedia inputs with the need to check social media and other input sources.

Many companies have outsourced standardized key entry work to the Indian subcontinent, the Philippines or other low-cost areas, but that slows approvals and decision-making. In the case of highly regulated industries such as finance, it may not be allowed or safe.

RPA is a small segment of the software market, currently worth less than $200 million in worldwide revenues. It can only handle repetitive tasks with known data. When data formats are new and unknown, it does not know what to do. Cognitive capture is the enabler that will cause the RPA industry to grow rapidly as it broadens its capability to better understand the data and images accessed. However, we are at the very early stages of cognitive capture-enabled RPA.