What Is It Like to Work With Androids? Merck KGaA, Darmstadt, Germany, Finds Out

Last summer, the €16 billion science and technology company Merck KGaA, Darmstadt, Germany, brought in an HR specialist named Elenoide to handle employee development questions at HQ.

Brought, not hired. Elenoide is an android, a carefully crafted work of robotics capable of some of the most advanced AI known to HR tech.

Elenoide isn't meant to replace or even supplement the company's workforce. She's an experiment. Every interaction she has with an employee, says CHRO Dietmar Eidens, helps Merck KGaA, Darmstadt, Germany, "find the best differentiated approach to addressing a megatrend driven by big data, and steer it such that we actively create the best fit for us."

Employee reactions to Elenoide, which have been positive, help Merck KGaA, Darmstadt, Germany, predict human/robot relationships in the workplace, but answering development questions is a low-stakes interaction -- phase two of the experiment will test an android project member and an android project leader.

Elenoide is an android, a carefully crafted work of robotics capable of some of the most advanced AI known to HR tech.

"To manage that effectively from an employee relations perspective, from a preparedness perspective, from a leadership competency perspective -- that is our primary interest," Eidens says. He wants to establish the principle that tech is positive, that it's an opportunity -- hence, Elenoide's Darmstadt debut.

But advanced as Elenoide is, as Eidens explains in this CHRO Conversation, it's only the first step toward getting the answers HR needs.

Emond: What led you to build an android in the first place?

Eidens: The genesis of Elenoide evolved around a few points. The first one has to do with our ambition and purpose. Science, technology, innovation, curiosity are all part of our DNA. They always have been there -- it's what's made us successful as a company for more than 350 years.

The second driver has to do with that megatrend around big data that affects any kind of business function these days, in particular people data that we deliver to our businesses through technology-driven solutions in our HR portfolios.

The third driver is that we want to adequately prepare our workforce of 56,000 people for the introduction of this kind of technology -- which is quite advanced -- and related technologies. We don't want to be affected by one of these many megatrends; we want to actively steer them such that they become meaningful for us.

Emond: How did you introduce this project to the workforce? Starting with why Elenoide looks like she does.

Eidens: Elenoide's design -- her look, for lack of a better term -- was a conscious decision. We wanted to make her look as humanlike as possible for two reasons. For one, Peppers are small and cute, you know? But they look like machines. They don't present a threat in any form or shape. But Elenoide is a replica of a real person, a member of the academic team at the Technical University of Darmstadt, where she was prepared for the on-site field studies. One of the theories we're working with is that people are likelier to accept human-looking technology, at least initially.

So for the first wave of the project, we chose a typical situation within the company, a relatively simple and straightforward inquiry around employee development opportunities. We programmed the android with, in essence, the contents of our learning and development portfolio that we have on our intranet. That was the testing ground because it's neutral.

It's a situation that would not be perceived as affecting any single person's job directly. That allowed us to see how employees (across all age groups, blue-collar/white-collar roles, nationalities, genders and so on) respond to the introduction of technologies and technology-driven solutions where they might not expect it -- HR, in this particular case.

Emond: How did you set up the test?

Eidens: First, we asked for volunteers, and 350 of our employees signed up in less than 48 hours, so interest was very positive. They chose between random allocation of a Pepper, a human resources employee or the android, or they could choose for themselves between a human, a Pepper or the android. Afterward, we interviewed all of the volunteers, and across age groups, genders, citizenship and educational background, their response to the availability of technology for development was extremely positive.

When we asked if they consider this technology promising for us, we got very high ratings. When we asked if this technology can create new jobs at our company, their rating was over 5 on a 1-to-7 scale, 7 being high. Would this technology be useful for your current job? Again, ratings just below 5. So, a unanimously positive reception to technology in areas where it wasn't present so far. And no major or immediate negative reactions, either, particularly among those who were interacting with the android only.

Emond: Elenoide, Pepper, your intranet and your HR team -- they can all provide the same answers to development questions. So why would your employees choose Elenoide over other sources?

Eidens: We asked the group who chose Elenoide exactly that question. Their answer probably has to do with what I said at the beginning of our interview -- this is a science and technology company. Many, many of our employees, even those in blue-collar roles, especially if they've been around for a while, have a genuine interest in scientific developments.

Of course, that is an important prerequisite because it indicates a general openness toward new technologies.

Emond: The thing about machine learning, of course, is that it learns. What is Elenoide learning? What are you learning from Elenoide?

Eidens: That was quite interesting to observe. One early indication was, which is not new, that it takes a while, particularly in these repetitive situations, for the software to become intelligent. So the first 50 or 60 volunteers got basic yes-or-no answers. The more the field studies progressed, the more Elenoide learned to make connections.

So, for example, when the employee said, "I would like to learn more about our learning programs in the area of management," Elenoide would say, "Your management courses in the learning curriculum are 'one-two-three-four-five.'" She would play back what she had been programmed with.

The further the test series went, the more she would ask validating questions -- "What else can I help you with?" Things like that. If the answer from the employee was "Nothing, thank you," Elenoide would suggest finding out more about the development opportunities for a project or recommend an expert.

She was starting to use the data that she had received from the answers of other participants; plus, the software started combining certain parts of the learning curriculum. That was nice to see. And nice to have. It's great that the ramp-up time to get a machine ready and act as if she or he were human is so fast. But it's not our key focus; it's not what we're trying to figure out.

Emond: Do your people see androids doing basic automated tasks? Or as higher-level partners?

Eidens: Right, another important finding was that the vast majority of the participants declared that they could see technology like an android robot as perfect for complex tasks in their own work environment more than as support for simple tasks. That information was important for us because it's the first indicator of the type of work for which we would bring in technology-driven solutions, whether Pepper-like or android or maybe something else.

It's the first indication of the selection criteria that we would use across the board because we're not going to digitalize 56,000 people's work here. That's not the intention. There are areas that lend themselves to the introduction of technology, like with automation in the 1980s and '90s, you know? Some roles were affected to a very high degree and others almost not at all.

What we're trying to do is find the best differentiated approach to addressing a megatrend driven by big data, and steer it such that we actively create the best fit for Merck KGaA, Darmstadt, Germany. Which already tells you that there was, from the beginning, an idea about a potential phase two, which is going to be more about real-life work situations that do affect current employees' job responsibilities and ways of working.

That is what we're discussing with our academic partner at the moment, which we'll launch next summer.

Emond: With android technology?

Eidens: Yes, in two real-life job roles. It's time to bring the project to operationally relevant scenarios at the workplace, one as a project member with a specific set of capabilities, the other as a project leader. Quite classic jobs. Of which we have thousands, all around the globe and across all businesses and functions. That was our requirement because those project environments allow us to examine where tasks and types of work may be done more effectively and/or more efficiently by a machine.

Secondly, those environments allow us to see the implication of human/robot teams. We want to understand the impact on communication and collaboration between project team members when one is a robot and the others are human. To understand the leadership capabilities, competencies and skills that you need when you work with a human/robot team.

The first wave of the experiment, asking Elenoide for information, is a neutral workplace situation. It doesn't present a threat to the current job. In the next wave, when we introduce technology to augment contributions to two real-life roles, we expect a different reaction. That will drive a very different dynamic, but that is what we need to focus on.

The experiment we ran with Elenoide this summer -- I mean, that type of technology has been around for a while. We've been using chatbots, for example, in recruiting and in our online job postings for three or four years now. There are Peppers running around grocery stores. That's really not anything new.

But what is new, and more personal, is supplementing, improving and in certain ways also replacing existing roles, even when they consist of very repetitive tasks. To manage that effectively from an employee relations perspective, from a preparedness perspective, from a leadership competency perspective -- that is our primary interest. We believe that the biggest lever for introducing technology in the workplace is the perception that it's not a threat, but an opportunity.

Emond: You mentioned leadership competencies that Elenoide can provide. Leadership is usually considered a very human quality -- what role could an android play?

Eidens: Well, we don't know. If you are leading a project or a team and part of the deliverable of that work package is done by machines, what does that mean for you as a leader? Do you need to communicate more or less? Do you need to apply different types of leadership skills, or can you manage a machine like you manage human beings? Do you even need to manage a machine? Or do you need to be even more involved to make sure that humans and machines work together effectively? Is the leader's role reducing the anxiety, explaining what the machine does and doesn't do?

Of course, it might simplify leadership, because AI can easily identify requirements ahead of time to drive sourcing and recruiting activities, for example, or identify internal development activities. Or it could also go the other way, you know?

It could make project lead roles much more complex, such as ensuring the most effective organization of work and collaboration between humans and machines. That's a very different profile of a project leader from today.

Maybe the introduction of technology at the workplace will cause the revival of the good old soft skills that so many thought aren't important anymore.

Emond: Those are big questions for HR.

Eidens: And we need to answer them. I mean, the technology is here; it's ready today. I had a key experience six or seven years ago, which has fascinated me ever since, at Narita Airport in Tokyo that told me about the potential of machine technology and artificial intelligence. I was looking for some very basic thing -- flight information or something -- and my iPhone was out of batteries. So I headed to the information counter, and two meticulously dressed women from the Narita Airport Authority approached me to offer help.

They looked almost exactly the same, and only at second glance did I realize one was a human and the other was a robot. Then the human asked me, "Sir, would you like to talk to me or a machine?" And I'm like what the heck is this thing, you know? And for some reason, I still can't figure it out, I decided to talk to the machine.

That experience completely turned me around. And maybe in a few years, there will be a different form -- because at the moment, the technology comes in the form of Pepper and Elenoide -- that none of us knows yet. I want to establish the principle here that technologies provide an opportunity. It's positive. For that, I need employees and leaders to understand what the implications are. And theoretical academic experiments in the lab don't deliver that. Because people don't experience work that way. That is why it was so important for us that we bring Elenoide here on the campus, at Merck KGaA, Darmstadt, Germany, and we're going to do the same again next year in the phase two field studies. Because only that is real life, real time.

So the question is not "Is it coming?" The question is "When and how?" And I want to decide how I can prepare our people for opportunities and challenges, yes, but also what it means to be an android's colleague.

Watch this video to learn more about their experience with Elenoide.

Jennifer Robison contributed to this article, which was based on an interview conducted by Larry Emond.