(Previously published in Outsourcing Magazine, on August 16, 2016)

For the last few decades outsourcing has been on a journey… literally! Outsourcing once meant having another company do your work, perhaps on your property. Later, outsourcing meant moving work to the suburbs. When work went digital, it moved to another town, in another state and soon to another country. China and India once looked like the ends of the Earth, but we kept moving… to Bangladesh, Vietnam and beyond. Where will outsourcing go when we have truly reached the ends of the Earth?

20 years ago I managed a New York investment bank’s support services. We needed to add staff, but our headquarters was full. We had space a few blocks away, but breaking up the staff created major inefficiencies, because workers were trained to work side-by-side, in one big room. When we learned to work “remotely”, efficiency improved everywhere, including our headquarters. If moving blocks away worked, nothing was stopping us from moving work halfway around the world!

In a few months, we had a successful pilot program in India. In a couple of years, every Fortune 500 company had pilots in India or China. The “ends of the Earth” were getting crowded, and it was obvious that costs would soon rise. Purchasing departments were also evolving, becoming procurement departments, and demanding ongoing year-over-year cost reductions.

An experienced outsourcer knows that lower-cost wages offshore initially deliver big financial benefits, but in a few years those benefits are lost to inflation when the program does not incorporate continuous improvement. Offshore locations often have twice the rate of inflation of onshore programs. If you launched a pilot years ago, locked in low rent and other costs, developed performance standards and incorporated continuous improvement, your program will deliver financial benefits for years to come. But the lure of lower wages at a new location has kept outsourcing on the move.

The tension between wages and productivity is highly visible in China. In the past decade, China has experienced thousands of labor protests every year, usually in remote rural areas with little press coverage. Workers have complained about mandatory overtime and unpaid hours. Yet, outsourcing contracts with Western corporations require fair treatment of workers. When Foxconn, China’s largest employer and the firm that assembles most of Apple’s “iProducts”, had very public labor disruptions, change came rapidly.

In 2013, Western media started to report on Foxconn workers. Workers challenged their managers’ orders and demanded better pay and improved work conditions. The strange thing, the historic thing, was that the workers won. Western outsourcing clients, led by Apple, supported workers and demanded that Foxconn follows their contracts and provide agreed to wages and work conditions. The result? Higher wages, reduced overtime, and government support for (some) public protests. Workers were so successful that wages in China outpaced neighboring countries. In Vietnam, wages are half those of China. Cambodia’s wages are even lower. These countries are now the newest “ends of the Earth”.

The government of China has been focused on full employment in manufacturing, not high productivity. If China’s factories offered the best automation and highest productivity, outsourcing revenues would be spent on foreign-built equipment rather than domestic wages. Work in China often takes more workers to perform than it did pre-outsourcing. Low wages offset these productivity issues, or they used to. Winning all of those labor disputes raised Chinese wages, but didn’t impact productivity. This growing gap in wages makes China’s neighbors increasingly compelling locations for outsourcing.

Time to move to the next “ends of the Earth”? Maybe not! China and India have a combined population of 2.6 billion. We once thought that they would provide all the staff and resource we would need for the next century of outsourcing. The search for the next low-cost location could end. The only problem is that these neighboring countries are relatively tiny. Cambodia’s population is only 15 million. Vietnam and the Philippines are each under 100 million. Before a fraction of the work could be moved, competition will heat up and wages will skyrocket.

China and India were game-changers. For years outsourcers talked about “the India price” or the cost of China, shorthand for a dramatically lower price to perform work. But now, even Chinese factories outsource work to other countries. It’s still early days, but a growing number of outsourcing firms in China must move work to other countries to keep contracts profitable.

Outsourcing can no longer depend on moving to the next country with the lowest wages. The next stage for outsourcing requires a different model, a model that values productivity over wages. That model started in the 1970’s when Japan began outsourcing to South Korean firms Goldstar (now LG), Samsung and Hyundai. Just like China, South Korea began outsourcing with simple work and grew in complexity over time. They survived their period of labor unrest and developed a standard of living comparable with Europe. While South Korea had a slight head start, both South Korea and China have been centers of outsourcing for decades. Yet, today, it is wealthy South Korea that outsources billions of dollars of work to inexpensive China.

In the 1980s, China, Thailand, India, the Philippines, and Vietnam paid similar wages. Each country has dramatically raised pay over these decades, largely due to outsourcing. China has the highest monthly wages ($656), followed by Thailand ($489), India ($302), the Philippines ($279), and Vietnam ($209). China’s single-minded dedication to outsourcing is one reason why China has outpaced the wages of their neighbors. Monthly wages in South Korea ($2,903), though, are in another tier altogether, just behind the UK ($3,065). This is the result of South Korea’s focus on productivity. One of the most visible signs of productivity? Robots! Lots and lots of robots!

South Korea has the highest robot-to-worker ratio in the world. According to MIT, South Korea has 478 robots for every 10,000 human workers, versus a mere 36 in China. Depending on the type, a single robot can replace just a few workers or hundreds. Conservatively, one replaces at least 10-20 workers. Let’s just say 15. Let’s see how this works.

If a factory with 10,000 workers bought 478 robots, it would need just 2,800 workers. That’s a massive boost in productivity. China’s government has set a goal to match South Korea’s robot ratio by 2020. If achieved, China would regain price competitiveness with Vietnam and Cambodia.
With over 100 million industrial workers in outsourcing and domestic manufacturing, China would need 4.8 million robots to meet their automation goals. Unfortunately, global robot production is just 250,000, and China already buys 25% of global production. Even if China bought every robot in the world… it’s not enough. The only option for China is to build its own robots.

Expert estimates vary, but a low estimate is that China’s 100 robot companies (plus another 100 robot support firms) currently build 50,000 robots annually. Last year, Foxconn alone produced another 50,000 robots for internal use, and Foxconn is not a listed robot manufacturer. That gives us 100,000 Chinese robots that are excluded from global sales numbers. Why are these robots excluded? Partially because they are only sold in China, and partially because major manufacturers see China’s robots as inferior and/or obsolete. That may be true, today, but China is pouring money into fixing their robot gap.

The regional government of Guangdong has earmarked $140 billion to replace human labor with robots by 2018. Zhejiang has guaranteed another $120 billion to automate 36,000 factories, within five years. That’s a much bigger investment in robotics than the entire global revenue of the robotic industry ($32 billion in 2014).

The International Federation of Robotics (IFR) reported sales of 250,000 robots in 2014, 30% more than 2013. If we add robots from China, global production rises to 350,000. To reach 4.8 million robots by 2020, China’s $260 billion stimulus must increase annual production to at least 60% to 70% through 2020. Production must rise even higher if global robot demand increases, which it will. South Korea’s robot ratio will not stand still. As China’s massive outsourcing industry improves its productivity, other major outsourcing locations must respond in kind, or leave the market.

By 2020 annual robot production will rise from 250,000 to over 2 million, making China the world’s largest robot buyer and manufacturer. Moore’s law dictates that computers (and robots) double in capability every 1-2 years. That means that two million robots in 2020 will do the work of 10 million of today’s robots. If one “2016 robot” = 15 humans, then 2 million 2020 robots could replace 100 to 150 million human workers. That’s nearly the size of the entire US or European workforce. Lower cost, more capable robots, produced in vast numbers. This will utterly transform global manufacturing, and outsourcing.

The shift from wages to productivity means that “the last place on Earth” will no longer be the country with the lowest wages. Instead, robot ratios will determine outsourcing locations. Manufacturing will move to factories with the best robot ratios. As labor costs decline in importance, shipping costs and global risks will make offshore manufacturing less appealing. Outsourcers need to take a hard look at the locations they use. In just three to five years, many contracts will be up for renewal. If offshore centers cannot rapidly improve productivity, many programs will be “re-shored” to domestic locations.

Meanwhile, artificial intelligence (AI), Natural Programming Languages (NPL) and deep learning systems will arrive in corporate America. In the early 21st century, outsourcing firms learned to replicate and outsource rules-based work: secretaries, administrators, presentation centers, corporate libraries. Now, AI allows outsourcing to move up the value chain and take on knowledge work, where the worker makes decisions rather than just following rules.

Automation drove 20th-century outsourcing, but accurately documenting work rules took a lot of time, was costly, and the final product was often fragile. New technologies are dramatically more efficient and more robust. Expensive traditional programming is replaced with inexpensive machine learning. AIs now learn from doing work – producing reports, writing memos, approving requests, sorting documents and making business decisions – while receiving feedback from subject matter experts.

AIs will easily replace junior positions, and then continue to learn how to perform ever more complex functions, eventually replacing senior knowledge workers. This opens up a vast new market for outsourcing, but it also creates opportunities for new types of outsourcing firms, especially if they can sell “trained” AIs as a service.

Outsourcing will reach into new markets, including transportation. Trucking firms cannot hire the drivers they need today, and as young drivers continue to reject long-haul trucking careers, the driver gap will grow. Transportation firms may buy their own fleets of self-driving vehicles, or like Uber, they may focus on innovating logistics, and outsource everything else.

Ivy League business schools and Fortune 500 firms have been partners for decades, but business school graduates have already begun to compete with AI-based financial analysts in new digital banks. As AIs arrive, will corporations “teach” proprietary knowledge to AI systems, turning them into productive knowledge workers? Alternatively, will Moody’s or Standard & Poors sell fully-trained but customizable AIs to corporate America? Or will financial firms simply hand over analytic functions to outsourcers, and let them deal with the mix of people and robots.

In just a few years, humans will work side by side with millions of AIs and robots. Outsourcing will create new services and transform the way that corporations train and retain staff (and their intellectual property). More work functions will be outsourced, more jobs will be outsourced, and outsourcing will move forward faster than ever. Outsourcing itself will be transformed as offshored work is automated and moved back onshore. Outsourcing, the great driver of change, is about to make everything change…. One more time!