Automation brings with it the threat of obsolescence for the traditional offshore model as smart machine technologies erode offshore’s competitive advantage, according to a recent Gartner report titled “The Rise of the Machine Leads to Obsolescence of Offshoring for Competitive Advantage.”
The report states that the increase in the use of smart and cognitive technologies and algorithmic business models “will debunk the idea that talent and humans are synonymous.”It claims that “labor arbitrage will cede to hyperautomation arbitrage” and “sourcing leaders will need to consider new approaches or fall behind the competition.”
Nearshore Americas spoke to one of the report’s authors, Frances Karamouzis, vice president and distinguished analyst in Gartner’s research advisory group focusing on business and IT services, about the implications of automation for offshore and how such technologies might drive the evolution of offshore models.
Nearshore Americas: The global market is changing as new trends impact on geographic locations and on companies in the IT sector. How would you characterize the current state of the offshore market globally? Are countries like India and the Philippines still the big winners in terms of offshore deals?
Frances Karamouzis: It depends on how you define winners. Our global report has a breakdown of labor pool size currently employed by each of these markets, maturity level by country and also cost level by country. We mapped all three of those characteristics into one bubble chart. So if it is pure size — by total size of total revenue by country — then indeed India is still the largest. But it tends to focus more on applications deals and on deals that are larger scale and price-sensitive, but sometimes the smaller deals in the digital space or smart machine space do not require as much scale. It requires more intellectual property. So not all of those deals are necessarily going to India or the Philippines for that matter.
Winner is in the eye of the beholder. If it is pure total revenue then they are still winners. If it is by virtue of impact, those are not necessarily going to India or Philippines. They are not necessarily organized by country — they tend to be organized by what the initiative is, who the client is, and sometimes the client may not use a third-party provider. They may source skills from Colombia or South America or Mexico or Czech Republic, and it might be their own employees that they leverage in those countries. It is usually a multi-disciplinary team working from a number of countries that come up with the solution.
The traditional offshore model has changed. It is not necessarily discrete, not that it has ever really been discrete. Clients never said I am gong to go offshore for a project, and 100% of it that was done in India or Brazil. There was maybe a percentage done onsite, a percentage done offshore, and maybe some nearshore. So the market for using offshore services is still growing and relatively robust. It is diffuse in that, 10-15 years ago, a disproportionately large percentage of that was going to one or two countries. Now that is potentially for any given client may include three countries or five countries.
People have evolved to more strategically sourcing to the right shore. And sometimes the right shore might be onshore, and sometimes it might be nearshore or offshore. So that mix of what shore you select, combined with services, is really changing.
Is the overall offshore market, on a relative basis, still growing? Yes, the pie is still growing. And the reason for that is that locations like the U.S. and the U.K. have not necessarily overcome the skills gap. The problem is that a large number of companies in the U.S. say I need these many developers or these skills sets, and even though we have unemployment, we don’t have the qualified people in those areas to meet those needs. We don’t have enough STEM skills still. It is the same in many developed countries, including the U.K. and Germany.
NSAM: So how is automation and similar trends impacting on the offshore model?
Karamouzis: Until we close the skills gap, there will still be this level of growth of offshore. In the past, people looked to offshore or nearshore for competitive advantage, to somehow differentiate themselves. That is no longer the case, that value proposition is diminishing and eventually going away.
They are actually now looking at the use of offshore for competitive parity, meaning that everyone else is using it therefore there is peer pressure to use offshore simply because it gives you a certain price point for a certain skill set that you can’t find in other locations. And so you want to use it to stay competitive and have parity with what industry peers are doing.
But if you want to look for competitive advantage then the use of smart-machine technologies, the adoption of certain digital strategies, and certain Internet of Things strategies, are how the new means of competitive advantage are going to be constituted.
NSAM: According to a recent report by Gartner, 40% of outsourced services will leverage smart machine technologies by 2018, rendering the offshore model obsolete for competitive advantage. What kinds of cost savings can this virtual talent over the traditional offshoring model?
Karamouzis: Well, it is not a compare and contrast to the offshoring model, because some things that you offshore are not necessarily the same things that you would use in automation. In some cases they are and in some cases they aren’t, but there isn’t a direct comparison. The comparison that people look at is what is the business outcome: How much of this can I automate, how long will it take, and what will my payback be?
There are lots of things that can be automated. Some, what we call low-hanging fruit, are easy to automate and are relatively low cost. We have published a number of case studies where we have shown the result of how much savings have been yielded by different types of automation and it varies from 10% to 70%, depending on what you are measuring and what the options are. It is highly attractive option because in some cases it causes higher levels of predictability, lower error rates than humans, and it is very repeatable. You can run a machine 24 hours, and it is working on Christmas or Easter, and it has nothing to do with being tethered to a time zone or to the working hours of human being. So there are different benefits that you are getting.
The savings are quite huge, and the overall costs are better because the machine part is easier to scale. If you want to hire an additional 1,000 people, it takes time. It takes time to hire them, time to train them, and so on. One you get a machine process working, if you want to add a 1,000 servers on the cloud, that takes no time at all. It’s a much shorter time frame.
It is a much different approach to the model. In the past it was not possible because the technology was not yet commercially available. But now it is.
We think first of all that smart machines and the whole market are going to evolve in the next decade. So you will get more and more technology and more and more options. There are lots of commercially available options already, but we are going to see a geometric increase in the number of options and what those commercially available elements can do. We think this market is something that we are going to be tracking for the next 10 years or more. It will take a long time for newer and newer breakthroughs, but in the meantime the market sizing is expected to in the trillions of dollars. It is already in the billions of dollars.
NSAM: Are vendors and clients ready for the impact of automation? How are they changing their approach?
Karamouzis: Right now it is fairly chaotic in terms of the reaction that people have had. You are getting all kinds of reactions that are all over the map. Part of the issue is that there isn’t a common term for “smart machines.” Some people are calling it “artificial intelligence.” Some are saying artificial intelligence is a term that has gone by the wayside and describes technology from 10 years ago. Others are calling it “cognitive technologies.” Others are calling its “machine learning.” So there is no one definition of what smart is or what artificial intelligence is.
So because there isn’t a common definition, lots of people have different interpretations of what it is. “Automation” as a term has been around for 30 years or more. And so some people are calling this market “hyper-automation” because it is using higher-order technologies. Some people look at it and welcome it and say this is the going save us a lot of time. Others look at it and say this is the rise of the machine — it is the beginning of the end for certain jobs and this is going to change the economics.
In a recent debate I attended in New York, four world-renowned experts were debating just this: What will this mean for the economics and what will it mean for how people value work?
One of the examples was of higher-order translation software. You can give it a recording of two people speaking in German, and it can translate it in English. The way that the technology works is that they took a machine and developed an algorithm and then they fed into it literally hundreds of thousands — or millions — of examples of German being translated into English. And what the machine did was figure out patterns within it to learn how to translate into German. It even recognized slang terms, or sayings, and was able to translate those. But what it did was take the labor of millions of transactions of humans who translated German to English and understood and developed all these patterns, and now it can translate it.
His argument was: Now what are you going to do, are you going to fire all the German translators? Is that ethical? You are actually now charging for the use of this machine, but you are not paying back a royalty to the translators whose work you needed to make this work. You can’t fire all the translators, because you need some sub-set of them to translate new words or new slang that comes into use. You can translate this to any industry or body of use. None of it has been determined as yet.
My view is that automation could be used for good or for evil. It depends on how people and governments accept these things. It is a little like global warming. If you allow every country to do what they want, it will eventually impact all of us. All of these technologies — no longer have a single impact on a given country. They have global implications, and they require global governance, and that’s the difficulty of newer technologies. It creates more of a global economy.
Economics and free-flowing trade between countries has been debated for a long time, and technologies like this exacerbates this issue. This makes everything much more possible, so it is harder to govern. It is harder to say that I would lose a job in India or in the U.S. or in the U.K. It is more difficult to understand when you implement some of these things where the job will be lost or gained. Some countries are saying it will create a rebound effect because we need more knowledge workers to run some of this automation, so we will take jobs back from the developing countries to the developed countries. But it is difficult to know the exact impact.