Tech Talk
Robotic Process Automation

One of the things I try to do in this column is provide readers with useful information and barring that, I strive to be entertaining. In that vein, I’d like to mention a weekly tech-related newsletter that I find useful and informative. It’s from Benedict Evans, a venture capitalist at Andreessen Horowitz, who shares his notes from the week of looking at trends and companies. (If seeing how a smart, thoughtful guy thinks about technology sounds interesting, you can sign up at www.ben-evans.com.)

An ongoing concern of our society is the replacement of employees by robots. It’s already happening in retail with self-checkout lanes at grocery stores, kiosks at McDonald’s, and robot baristas like the one at CafeX, just down the block from my office in San Francisco. And we’re on the verge of replacing thousands of people who drive for a living with autonomous vehicles, in the form of taxis and long-haul trucks.

That trend is now extending to the office. But rather than seeking to replace employees wholesale, the idea of robotic process automation (RPA) is to assign the repetitive, relatively mindless elements of office work from humans to software-driven processes. McKinsey & Company has this optimistic description: “[RPA] can do repetitive stuff more quickly, accurately, and tirelessly than humans, freeing them to do other tasks requiring human strengths such as emotional intelligence, reasoning, judgment, and interaction with the customer.” Of course, if the entirety of your work is mindless repetitive tasks, you could be out of a job.

This isn’t new. It’s just a new name and vastly more capable technology. The same process has happened in business before with QuickBooks and spreadsheets, to cite just a couple of accounting-related examples.

RPA software comes in several forms. The simplest replaces activities, which involve data collection: go here, get this data, perhaps process it in some (simple) way, and put the result over here. Clearly, much of that can be automated in the form of a software “robot.” The next step up is a tool, which provides templates for different kinds of office processes that an organization can use to develop its own robots. The do-it-yourself approach is followed by enterprise-level offerings such as Blue Prism, which are aimed at corporate IT departments who are concerned about compliance and security. Finally, there are custom solutions that are specific processes such as accounting and finance.

Basically, if you can make a flowchart of a process, where “yes” and “no” questions and well-defined actions make up the steps, it’s a candidate for RPA. The difference is that “before” RPA, you needed a programmer. RPA replaces that with a tool, much the way spreadsheets were a generalized solution to manipulating rows and columns of data with a custom program.

RPA is particularly useful in highly regulated industries, as it can enforce the rote aspects of regulatory compliance. It’s also useful in customer service where common requests can be dealt with by a software robot, allowing employees to focus on requests that involve judgment. McKinsey claims not only a first-year ROI of 30 to 200 percent, but also that employees are grateful to give up the mindless aspects of their jobs to a robot.

The real advantage of RPA for a small business is that it allows existing employees to be more productive. Rather than hire a new employee, a business invests in automating the repetitive work that takes up part of an employee’s day. One vendor calls its small-business service offering “RPA Robots for Hire,” akin to the way that small businesses hire temps to deal with backlogs. This list (tinyurl.com/yct6mpuf) identifies 30 processes which a business might automate using RPA, and is worth a look to educate yourself about the practical application of RPA.

Note that RPA only scratches the surface of so-called “white-collar” automation, as it deals with flowchart-type processes, where rules determine outcomes. Beyond RPA is the world of machine learning, which excels at recognizing patterns. These robots are making inroads into the world of stock selection and cancer diagnosis (where they display better-than-human accuracy in reading X-rays), as well as the decidedly-creative field of fashion. Pattern recognition is no longer a uniquely human skill.

As I said before, despite all the hype from McKinsey, Gartner, et al., RPA is nothing new. Software has always been about enabling computers to do what humans do, only with greater speed, accuracy and reliability. Programmers have moved from programming with wires to writing programs in high-level languages. Accountants no longer tote up ledger pages. Cars tell mechanics what is wrong with them. This is just another step along the way.

Business will continue to drive the trend toward RPA because it lowers costs and raises profits. The real question that faces our society is what those profits will buy: mass prosperity or increasing inequality. And ultimately, that will be a human decision with enormous consequences for all.