by Arik Hesseldahl

ServiceNow’s Intelligent Automation hints at a smarter, more productive workplace

May 30, 2017
Artificial IntelligenceEnterprise ApplicationsIT Leadership

The cloud company is deploying machine learning techniques to speed up common processes in IT, human resources and security and predicting problems before they happen.

There was a time in my life, not so long ago when I felt oppressed by tiny pieces of paper: receipts.

For much of my working life, I’ve spent about a third of each year traveling for work, and running up a lot of expenses.

About once a month, I spend most of a Saturday or Sunday with a pot of coffee, a pile of receipts, and a spreadsheet. I do this on weekends because it’s basically impossible to devote the time to concentrate during a busy work week. My day-to-day work commands a higher priority. But still the expense report has to be done. So I sacrifice valuable personal time on the weekend to do it. That’s time I’ll never get back.

The pang of those lost weekend afternoons comes back again and again as a personal reference point in the unfolding discussion around artificial intelligence and automation and how it’s likely to affect the workplace in the coming years. That messy pile of receipts, I’ve always thought, should have been eliminated years ago, replaced by a smart assistant that can prepare an all-but-complete expense report requiring only my minimal attention and approval to submit up the chain of command. My weekends — and likely those of countless others who have suffered similarly — would be saved.

If the above sounds in any way familiar you’re not alone, and there’s data to prove it. A recent survey commissioned by ServiceNow, the cloud software company that specializes in streamlining work flows, found that a broad portion of its customers around the world feel the same way.

Nearly half of the 1,874 senior executives the company surveyed estimated that they spend an average of 16 hours a week or the equivalent of two full days working on administrative tasks like expense reports and other time-sapping but necessary busywork that takes them away from their more valuable day-to-day work.

Additionally nearly all of them — 86 percent — say the velocity of work overall is increasing at such a rate that within two years or less their companies will reach a breaking point at which, they’ll have no choice but to turn to some sort of intelligent automation to keep up.

The pressure has been on corporate leaders to do something to spur productivity gains for a decade. Since 2007 the rate of productivity growth in businesses outside the agricultural sector as tracked by the U.S. Bureau of Labor Statistics has averaged 1.2 percent. For a historical comparison you’d have to look the economically stagnant years of the mid-1970s, remembered for high inflation, recession and an energy crisis, for a rate so low. It’s also less than half the rate seen during the eight years ending 2007, the first year of smart phones. So much for mobile productivity.

ServiceNow’s survey quantifies what many executives get intuitively but often struggle to articulate: One way or the other machines will be arriving on the scene to streamline and in some cases take over many time-consuming, value-sapping but necessary business processes.

And while the thought causes some anxiety in some economic and policy circles, about the potential for job losses, there’s probably more to be gained in the end. As it happens ServiceNow earlier this month made some significant moves that point the way toward how automation will be put to the most effective use.

The company is known for its System of Action, a sophisticated cloud software application that organizes and streamlines work flows, essentially the step-by-step procedures for accomplishing thousands of different kinds of tasks that occur in routine but immense volumes at large companies. Its specialities are in the areas of human resources, customer service, and IT processes.

The one thing these processes share in common is that they begin with a request by someone — they may be a customer, an employee or a vendor — that requires a response and some kind of action by someone else at the company. And more often than not they’re handled in an inefficient hodge-podge of ways: emails, phone calls, paper forms or a subjective combination of all three.

All of them are manual, meaning they’re subject to the benign whims of whichever person responds first. A task may not be treated with the priority it should. It may be mis-assigned to a person who responds to it incorrectly, or it may be forgotten altogether. ServiceNow has built a business worth $1.4 billion in annual sales in automating that process of managing and categorizing these requests. Raise your hand if you’ve ever had to chase down the right person to fulfill some inane administrative task at the office, only to lose hours or days of productivity while doing it, then multiply it by the extremely large number of your choice, and you see the opportunity that ServiceNow is attacking.

Now it wants to take the next logical step and inject what it calls intelligent automation to streamline many of these repeatable processes further. Having run literally billions of these service transactions through its application, ServiceNow has an excellent set historical operational data suitable for training an intelligent system.

Built largely on the work of DxContinuum, a machine learning startup in Fremont, Calif. it acquired earlier this year, ServiceNow’s intelligent automation engine specializes in quickly categorizing and routing requests that arise from the thousands of Internet of Things devices that are getting deployed to networks by the thousands.

But the Intelligent Automation Engine goes a step further in using that historical data in new ways: It can be used to watch for and then call attention to the early signs of the sort of IT system failures that can bring a business to its knees. It might sound an early alarm on a security breach seconds after it happens and when it can still realistically be contained and before it becomes fodder for a news headline.

Over time all this transactional and operational data can be further mined to create a benchmark against which you can compare your company’s performance against that of peers, and to create models predicting performance.

I chatted with John Donahoe, the former head of eBay who recently took over as ServiceNow’s CEO, and in his comments you can hear him speaking for his CEO peers who are constantly under pressure from shareholders and their boards to find new wells of productivity that lead to growth.

“What I think most CEOs want is to devote as much of their company’s resources as they can on innovation and on serving their customers,” he said. “And they want to devote as few resources as possible on managing the complexity of a global enterprise.”

Here’s another informative nugget from that survey: Companies who had automated 70 percent or more of their business processes saw revenue growth than those who hadn’t. In fact they were six times more likely to experience revenue growth in excess of 15 percent than less-automated companies, and were twice as likely to beat their internal financial goals.

What it translates to is this: The economic incentive for large companies to embrace the automation of their businesses processes is huge. And while the starting gun has already fired, the race is still in its early laps.

At first we’ll see automation making easier the day-to-day lives of employees in IT departments and HR offices. Later on it will relieve us all from the drudgery of value-sapping repetitive tasks we all loathe like expense reports.

In fact there are few jobs in the global economy that will completely untouched by automation. A McKinsey study earlier this year estimated that roughly half of the work that people are paid to do today could be automated by 2055 using technology that has already been demonstrated.

Intelligent automation will, when ideally deployed, complement, enhance and inform the work that people do, while people will supervise, improve and intervene on automated processes when necessary.

Yes, the changes that are coming included the potential for the loss of some jobs that will be made obsolete. This has always been true of economic progress for the last 200 years or so. But they’ll be the minority: The same McKinsey study found that less than 5 percent of existing jobs could be completely eliminated by automation.

And that’s why we should see the coming changes not with angst, but with hope. While automation will cause significant workplace changes, they logically imply a virtuous cycle: Companies seek growth, so an overall boost in value creation usually results in new rounds of reinvestment to continue that cycle of growth. That reinvestment has to include the creation of new jobs. 

Jobs, one hopes, that won’t require wasting weekends on expense reports.