Gartner Fires Back at HfS as Debate Rages About Scale of Automation’s Job Destruction

Getting the real picture on automation can be confusing. The prevailing wisdom seems to be that automation will result in huge job losses, but this techno-pessimism may not paint a full picture. Nearshore Americas examines how you can make sense of the information around automation.

robotics automation

You’d be forgiven for thinking that your job will soon be lost to a robot, if you believe the techno-pessimistic hype about automation. World-renowned research companies like Gartner seem quick to declare that your next boss could be a robot — although, a closer look paints a different picture.

Huge increases in unemployment, lay offs, and the obsolescence of entire work categories, seem to encompass the prevailing wisdom about the impact of automation, from banking to contact centers and IT. Those looking into the possibilities of automation may find these reports confusing, without sufficient concrete detail to help them make a decision around this investment.

Calling Out the Scaremongering

Phil Fersht, CEO and Chief Analyst for global services analyst company, HfS Research, is outspoken about what he sees as the irresponsible techno-pessimism in reporting on automation. He recently responded to Gartner’s piece about robots replacing your boss with a blog post on the Horses for Sources website, challenging the analyst firm to provide data for the claims. His views seem to be echoed by many, with the post attracting a great deal of attention on social media – and Gartner refusing to respond.

Then you have analysts and companies like Gartner who have nothing better to do than to make stuff up — Phil Fersht, HfS Research

Fersht told Nearshore Americas that there is definitely a great deal of scaremongering around automation. “Well-respected academics, analysts, and consultants are coming up with far-reaching predictions on the impact of robotics on industry, on jobs, and everything, without any data to back it up,” he said. “It’s just people pontificating to get headlines as opposed to having measured insight and analysis based on hard facts, hard data, and a measured methodology.”

philfersht
HfS Research’s Phil Fersht calls Gartner out on their “irresponsible” reporting on automation

HfS Research’s own predictions have indicated that over the next five years there will be an impact on the services industry – people working in services jobs in BPO, ITO – that will result in a reduction in jobs of about 9%. The focus for those job reductions is really some low level transactional roles eventually being eliminated because of improvements in automation.

“But it is also going to create more jobs higher up the value chain in data science skills, digital skills, and things like that. Companies are going to have better data to act on and they will need more people who have more collaborative, creative skills to act on that, so while we will see a decrease in transactional roles we will also see an increase in more strategic type jobs for people,” Fersht said.

However many jobs we are talking about, we shouldn’t be looking at it as a one to one relationship to a machine — Frances Karamouzis, Gartner

In Gartner’s Defense

Gartner has, until now, chosen not to respond to Fersht’s posts, but analyst Frances Karamouzis told Nearshore Americas that Fersht had “obviously not even read the study that he was referencing” and that the study itself was released almost a year ago. She stands by Gartner’s predictions, but paints a different picture of what it all means, highlighting the fact that “however many jobs we are talking about, we shouldn’t be looking at it as a one-to-one relationship to a machine.” In other words, automation is not about job loss in the sense of a machine taking the specific job of a specific person.

Karamouzis: If a service company is not providing value beyond labor arbitrage, there may be no benefit to continue using them.
Karamouzis said that robo-bosses are not about a complete replacement of management roles.

Instead, Karamouzis said, it is about automating tasks that have traditionally formed part of a job role. The Gartner post that sparked the response from Fersht highlighted the idea of robo-bosses, a term Gartner coined, but Karamouzis said that this is not about a complete replacement of those management roles by machines.

“Our position is not what he is saying. Our position is that just like any disruptive change that has happened since the printing press, what people do, how they do it and how value is added into the equation is changing. What we were trying to do with regard to robo-bosses was not look at it as a boss in the strictest sense,” Karamouzis said.

Why Robo-Bosses?

Gartner took the approach of identifying a list of tasks usually associated with the manager or supervisor role. These included routing workloads, measuring tasks, and analyzing productivity statistics, as well as the more human aspects of providing feedback, creative leadership, worker relations, and dealing with bereavement.

From this list they then identified three specific tasks: routing workloads, measuring tasks, and analyzing statistics that they felt artificial intelligence — robo-bosses — could potentially handle.

“Because our predictions must be between 12 and 20 words and can’t be longer than that, we came up with the robo-boss as a term to succinctly capture the essence of this automation of tasks traditionally associated with a boss,” she said. She is clear, though, that this does not mean getting rid of human bosses in their entirety.

Perhaps the need to capture attention is part of the problem in the sector as a whole. As competition for audiences increases, media outlets, academic institutions, and even analyst firms are under pressure to offer insight in 140 character Twitter style sound bites in the hope that users will be tempted to read the full story. Often, though, the sound bite becomes the story and the nuance of the issue is lost in an increasingly cluttered environment. In this case it seems Gartner and HfS are not as diametrically opposed in their views of automation as they first seemed.

Higher Quality Jobs, but Less Jobs Overall

Hochstein: there won't be as many jobs
Hochstein: there won’t be as many jobs

Eric Hochstein, an Economic and Business Development Consultant, offers another view. “The automation industry is doing its best to convince people that jobs won’t be lost and jobs will get better because it’s in their interest. I would like to think that’s the case, but outsourcing, offshoring and automation has been all about getting labor costs to the absolute minimum and automation is the end of that progression.”

Sign up for our Nearshore Americas newsletter:

Hochstein does agree to some extent with Fersht’s notion that there will be some better jobs and he does not believe that all supervision will become robotic. “But I am convinced in both contact centers and business process environments we are going to see fewer people and those people may not be the ones who are best suited to the higher level jobs that will be remaining.” One thing is clear, though, Hochstein believes there won’t be as many jobs, explaining that we have seen the same thing in manufacturing.

“Jobs that went offshore, whether to nearshore or China, when they come back there are one third as many jobs because automation and technology have progressed so much. The jobs that come back require much different skills that those that left,” he said.

What Hochstein has seen is that as companies become more comfortable with automation in the business process area, they slowly expand their automation to more and more lines of business and increase displacement of current workers.

Automation is also going to create more jobs higher up the value chain in data science skills, digital skills, and things like that — Phil Fersht, HfS Research

A Balanced Argument

There is pessimism around but there is also a recognition that there are benefits to automation and there is the potential to create jobs through the use of these technologies. “Then you have analysts and companies like Gartner who have nothing better to do than make stuff up, which is literally what is happening. We think it is irresponsible for a respected research firm to come out with stuff that is complete rubbish and then not even have the decency to defend themselves; they just ignored it, even though we had 40,000 people accessing that blog,” Fersht said.

Fersht does think it will change, though – this is a short period of hype and it will become less pronounced in the future. “Companies are going to want to hear real business conversations around what enterprises are doing in industries to combat digital transformation, to embrace it, to get ahead of it, what they are doing in terms of robotics automation, and what processes can benefit from some automation. It’s about driving value rather than just cutting jobs,” Fersht said.

Regardless of how many machines you have, you still need a human in the loop — Karamouzis

Applications for Services Automation

Hochstein emphasized the need for a complete picture and better information when examining the use of automation in BPO and ITO, differentiating between RPA and cognitive automation. In the contact center and service delivery environment, structured data often has to be moved from system A to system B, from a customer record system to a sales system, for example. In that scenario you are just putting standard information, scraping screens and moving data around. Hochstein believes that that type of work is ripe for automation and is easily done by robots.

The more knowledge work where decisions need to be made often requiring unstructured data – different formats from emails, texts, handwriting, or different companies using different invoices – requires machines to be trained to interpret in a cognitive way. “That is the next generation of automation and in many cases the immediate holy grail and we are not necessarily there today,” he said. “But the companies that can separate their structured and unstructured and decision making with data can improve efficiency significantly.”

Karamouzis said that those looking to introduce automation, need to look at the absolute process, whether it is processing an insurance claim or a mortgage or developing an IT application. There may be 100 or 150 steps in the process.

“You need to look at the decision elements in each step to determine how much judgement is required. You are looking at the process and determining how much is rules-based, routine, and predictable versus how much of it is judgment based and what type of judgement is required,” she said. It is crucial to look at the data and determine whether it is structured, ad hoc, stable or coming in at high velocity. You also need to consider the human in the loop. “Regardless of how many machines you have, you still need a human in the loop,” she said.

Taking all of this into account companies can determine where the best applicability for automation is. It is clear that what is needed now is balanced, measured reporting on automation that does not flinch from the negatives but that also embraces the positive possibilities of an increasing use of these technologies.

Tags

JOIN THE CONVERSATION

JOIN THE CONVERSATION