Over the next 10 years, the work of 110 million to 140 million knowledge workers around the globe may be handled...
by cognitive robotic process automation systems. That's according to KPMG's Cliff Justice, citing research from McKinsey Global Institute last week at the World BPO/ITO Forum's Global Sourcing & Cloud Summit in New York.
This shift to robotic process automation -- which digitizes labor through the use of advanced machine intelligence, engagement, analytics, big data, social media, mobile technologies and cloud computing -- will change the knowledge worker labor market as we know it, according to Justice, who characterized robotic process automation (RPA) as a double-edged sword. While RPA is expected to ultimately expand the job market according to some economists and researchers, the latest wave of automation in its early stages will disenfranchise many workers, exacerbating the income inequality seen today.
Justice cautioned that the RPA systems won’t necessarily cause 110 million to 140 million knowledge workers to lose their jobs, assuming that “the economy and demand for knowledge workers continues to grow at projected rates,” he said in an email exchange. New RPA technology will fill some of the future demand for knowledge workers, he said.
A portion of the displaced workforce will be freed up from doing repetitive clerical and administrative tasks to focus on innovating and generating revenue, but others will be automated out of a job.
There are three classes of RPA technology, said Justice, who is partner and U.S. leader for shared services and outsourcing advisory at KPMG. The first class is basic process automation, which includes sophisticated macros, screen scraping and business workflow technologies that sit at the OSI presentation layer and are not integrated into the IT system. Class 2 -- enhanced process automation -- are technologies that use natural language processing and can, for instance, understand unstructured data and apply that understanding to the process automation. The third and most transformative class is autonomic/cognitive platforms.
The Class 3, cognitive platforms "have the ability to parse context and understand meaning like [IBM's] Watson [supercomputer] did in Jeopardy," Justice said. "As that technology merges with robotic task automation, you have a whole different class of digital labor. You have technology that can understand your customers [and] run queries against rules engines. [If the response] falls within parameters, [the technology] can inform the robot to carry out a transaction and actually do things that in the past required decisions."
For queries that produce results outside a given set of parameters, the system can route the request as an exception to a human employee, who can review it and answer it.
Cognitive robotics advancements will have a major impact on the labor market, Justice said, shifting the knowledge worker framework from one of labor arbitrage, which reduces costs for relevant functions by anywhere from 15% to 30%, to one of labor automation, which reduces costs by 40% to 75%, he said. "In the very near future labor arbitrage is not really part of the conversation. … The concept of outsourcing, which has been sort of perversed into a vehicle to move things to low-cost areas, [is] a dead concept today," Justice said.
Justice likened the shift to robotic process automation to the Industrial Revolution. It will transform "the whole idea of how you operate your business," he said, suggesting that rather than looking to untapped geographic regions -- "the next India" -- to drive greater profitability, the new untapped potential will be found through digitization and automation. "If you think it's … going to be about low-cost labor, you're dead wrong."
One key element of RPA, Justice said, is that labor automation is scalable. "[With labor arbitrage], if you're going to scale the business, your cost moves up as you do it," Justice said. Labor automation, on the other hand, "follows the price/performance curve, not a labor curve. Technology is faster and cheaper, whereas labor doesn't scale that way."
What about the impact of the shift to cognitive robotic process automation on the economy?
Justice pointed to work by MIT researchers Andrew McAfee and Erik Brynjolfsson, which he said concluded that robots will eventually create "more jobs than we could ever have imagined." But he also cautioned that before this job growth, there's likely to be a lot of social unrest as income inequality increases.
Justice also cited Thomas Piketty's 2014 book, "Capital in the Twenty-First Century." Piketty "talked about how when return on capital is the predominant measure of domestic income and it exceeds labor, income inequality accelerates and you have social unrest," Justice said. "That's what we're seeing now. Capital is exceeding labor in the western world and in the developing countries."
The good news in this massive disruption, he said, is that cognitive robotics applied to the knowledge worker space opens doors for innovators. "And you open the door for massive amounts of opportunity. This is an entrepreneur's dream."