Bill Kerr: The CEO of telecom giant Vodafone will never forget the day the company tested a robot against one of its best employees. The task? Making address changes. The robot was seven times faster than the human at a quarter of the annual cost. So the robot is 28 times more productive than a person. Vodafone says robotics and artificial intelligence will reduce costs so quickly that it must act fast; otherwise, its competitors will steal customers away with innovative products and lower prices. These changes have big implications for workers. Vodafone is already retraining people to keep pace with new technologies and racing to build new skill sets.

Welcome to Harvard Business School's podcast, Managing the Future of Work. I'm Professor Bill Kerr. I'm joined by the CEO of Vodafone, Vittorio Colao, who says workers in many industries could face a rough ride if they don't get retraining. Welcome Vittorio.

Vittorio Colao: Thank you.

Kerr: Vittorio, what was your reaction when you saw how much faster the robot was than the human in making the address changes?

Colao: Well, the immediate reaction I cannot say. It was, in English, it would be a four-letter word, and in Italian a five-letter word. But essentially it was I really need to act fast. It was really for me the moment of awareness that incredible productivity improvements were ahead of us, but also big challenges from an organization, and the people point of view. And also, the realization that we had to introduce as fast as possible that type of technology everywhere, but also think about the next steps in terms of international balance of our resources and type of skills that we needed to follow this development.

Kerr: We've seen huge technological advances in the past, for example, going from horse and buggies to cars to landing on the moon. How does this newest wave of technological change in which machines can think make this time different?

Colao: I believe this time is really different. I believe it's different because of the combination of super-fast broadband, machine learning, and the ability to connect essentially every point of the planet with data, which stays in the cloud, and is incredibly rich, we change the way we can do many things. And if you think people talk a lot about driverless cars, but you know it's not just driverless cars, think about microsurgery: I mean it's possible, it’s possible already today, and we do it with a hospital-based in London to give remote consultation on surgery in every part of the world, and the next step is to connect surgical robots to databases, which contain the memory of all the operations that have done in that particular case. So then the question comes, if you had to choose, would you prefer a doctor who has done the same surgery to your knee 25 or 50 times in his life, or would you prefer a machine who can have instantaneously the memory of a thousand procedures, and they would know exactly what to do to your knee? Well, I think we all know the answer.

Kerr: So how do these advanced technologies widen the scope of jobs that can be performed by machines?

Colao: Well of course, machines initially were capable of doing repetitive tasks, and with, I would say, a relatively accurate but not completely precise accuracy in preforming the task. Now two things are happening: One, you have a complete precision which goes beyond humans, clearly recognizing patterns, having the best and more precise surgery, the accurate interpretation of something that is happening in milliseconds is something that the machines already today do better than the humans, and that's one side of the story. But the other side of the story is that [they] can make decisions, and each decision improves the next decision.

Kerr: Let's take marketing. At Vodafone can computers compete with humans at marketing campaigns?

Colao: I don't think they compete. It's a bit of an exaggeration. But for sure, again, I remember when I started I actually hired a HBS graduate many, many years ago to come, and do what we called customer relationship management. And her job was to use the power of our IT infrastructure to give segmented offers to the customers. But again, the intelligence was her intelligence applied then by a machine to the right segments.

Today, what we can do is actually give the perfect offer for each customer, but not predefined by a human. So we have different engines that we use. But again, another eye opening moment has been when I was talking to a colleague in South Africa, and he told me, "Listen. The most important thing is that you don't allow the marketers to put their fingers on the machine because the logic of the machine is the right one. And if the marketers put their logic above the logic of the machine you'll end up with wrong offers.” So it's not competing. You're really integrating, but of course there is a higher level of input that comes from the machine.

Kerr: Vodafone is using artificial intelligence to screen job applicants. How do machines make decisions about candidates?

Colao: Machines don't make decisions about the applicants, but the machines help us with two problems. The first problem is, how can you really assess tens of thousands of resumes today coming from many universities that otherwise we would not follow, and rank them in terms of appeal? And therefore, that's a mass selection job that before was conducted by humans, by people. But people are inconsistent across them. They're inconsistent even sometimes within a group, and you're not guaranteed that you really get the same criteria applied to candidates coming from University A versus University B. In this moment, in this way you have a ranking, which is based on objective factors.

The second is to take away bias. The reality is that we all have biases, and each recruiter has a bias, and if I see somebody who fits my biases I'm more inclined to be positively influenced by these. A machine is not. And therefore this, in terms of diversity, in terms of inclusion, in terms of women, races, looks, facial expressions, all of this the machine can handle in a neutral way, which I think is good because it gives fair chances to everybody.

Kerr: Another innovation that will have a big impact for the workforce is chatbots. For starters, what is a chatbot?

Colao: A chatbot is a chat like the one we have on WhatsApp, or on iMessage, but it's not with a human. It's conducted between a human, a customer in our case, and a machine.

Kerr: Does that mean when I'm computing in an online chat and I think I'm communicating with a person it might in fact be a robot?

Colao: This is a very good question. In our case it's not. We had a long, long discussion and we came to the conclusion that in order to preserve trust with a customer you need to know that you are dealt with by machines. So TOBi, which is the name of our chatbot always appears like a little robot, little face, an icon of a robot, and it clearly says, "I'm not a person. I'm TOBi." And it's TOBi everywhere. And interesting enough when TOBi has difficulties, it's understanding what's going on, TOBi says, "Let me pass you to my colleague." And then, the colleague is real, human, and the human continues the conversation.

Kerr: The use of chatbots is expected to grow almost 40 percent over the next four years. Why?

Colao: There's two reasons why every company, every CEO should look at chatbots. The first one is the asynchronous nature of the relationship with a chatbot. We are all becoming more and more used to multitasking. We are all becoming more and more used of I ask something to the Vodafone call center, but in the meantime I read something, or I watch TV, or I do something. That's a very, very distractive and inefficient way of interacting with a company if the company still uses a call center because you leave people waiting. With a machine you have no problem. A machine can handle thousands of conversations in parallel and [if] you want to take half an hour to watch your Netflix, and then come back to us we're perfectly fine. So it's very convenient for the customer to deal with their own issue in an asynchronous way with all their time.

And the second quite frankly is they're accurate, so it's very hard for a machine to be in a bad mood. It's very hard for a machine to not remember exactly what is the roaming price in the U.S. If they give you an answer, it's the right answer. And so it's more accurate and it's more convenient.

Kerr: As people get used to robotic assistance like Alexa at home, to what extent do you think that will make them more comfortable with chatbots in a commercial setting?

Colao: I think immensely. We are all becoming more and more used to interactions through machines whether it's voice-based, or text-based. If I look at my children, if I look at the young generation, they really are surprised when we say, "I'm going to call the call centers of the water company." That they look at us like dinosaurs. They say, "Why do you need to call them? Just write to them. Do something." So there's a big generational shift. The beauty of voice, of Alexa, of voice-based technology is it is more natural and more immediate also for people a little bit older like us.

Kerr: Vodafone is using several chatbots. How do you teach them to answer the questions? And, what happens when customers get frustrated?

Colao: Well, first of all we teach them to answer the questions. So we look at how a certain flow has gone, and if there is something that can be improved. It's a little bit like in the old days. We used to have the debriefing sessions at the end of the day where the call center agents would say, "Hey, we need to change this. We need to change that. This is not good. This is bad." The same happens with cognitive engineers who look at the path, and what has happened in the conversations, and they realize what has to be changed, and that is fed back into the new system. So it's not very different, but it's clearly now digital.

What happens when people are frustrated? Well, at the end of the day people get frustrated with bots for the same reasons. So they are frustrated with humans is when you don't get the answer, or when the answer comes too late, we have certain keywords, or certain signals that indicate there is a stress level going up from the customer. And when that happens, there's a threshold beyond which the bot says, "Hey, let me give you a human now." And we hopefully resolve the problem.

Kerr: Will chatbots result in a dramatic reduction in call center staff?

Colao: Well for sure, it will result in a different composition. I think call center staff, yes, will definitely reduce. But on the other hand, you will have an increase in people who work on these developments, work on the kind of… if you want higher level of automation, not all of them have to be engineers. Some of them are simply, you know, people who know how to structure things.

In my days I remember 95 percent of the people were call center agents, and 5 percent of the people were just handling IVRs, or automatic responding units. It will change. It will go down the number of people who manage the live calls. But it will go up the number of people who actually have to measure the technology, and the answers that the bots give to the customers.

Kerr: Technology and digitization are vital for Vodafone. How is this changing the senior executives?

Colao: Well, we are becoming humbler. We understand that we don't understand everything. We are trying to be more experimental ourselves. We have introduced the obligation to try ourself much more than before. Again, in an old world it was relatively easy to test our products and pretend that we knew them. Today, it's much more complicated. I personally use my chatbots, my call centers, I try products, and I have to dedicate more time to that because it is more complicated. If you don't try it, you don't understand it.

We also have to be more exposed to the outside world. We were clearly correct I think, very focused on our own business model. Now, we spend more time looking at others’ business models. We travel more. We are more open. We are less obsessed with protecting our own innovation, but much more open in actually looking at others’ innovation, and try to incorporate. So banks, insurance companies, car companies, whoever, we are more open and more interested. Now it requires a different allocation of time, clearly.

Kerr: You have a program called Digital Ninjas that many workers my age and older would love. Tell us about how it benefits Vodafone executives?

Colao: Digital Ninjas is in the end a simple program. We just pick a number of younger colleagues, who are really digitally savvy, and we couple them with the senior executives, and these guys come, every now and then, either to explain to us how to use certain things or to judge how we use them. So for example, in my case my Digital Ninja told me, "Vittorio, I looked at your social feeds. You are really pitiful. You really are boring. You don't think that anybody would be really attracted by what you write?"

In another context, he looked at my use of digital note taking, and he said, "That's pretty good. It's like a three-year [old] boy walking. You need to learn to also run, and to also climb, and to also do everything else." Now, I'm not saying I'm adopting everything that he told me, but of course it's useful to see how they can do. They show us for example collaborative working, which I'm not particularly good because I still do my stuff and send it by email. And they say, "No, no, no, no. That's not the right way." So, this is the type of interaction we have.

Kerr: What is Vodafone doing to make sure its workers have the new skills to keep pace with technologies?

Colao: Well, the topic of how to keep the workers updated is immense. We are producing a huge amount, partially internally, partially externally of content on the digital Vodafone way of doing things. For example, our marketing, digital marketing courses are really very good. We are even willing to exchange them with other companies at this point because we really need to be sure that we are state of the art. We ask people to do these tests, and these courses, and then eventually from time to time we have classic training sessions. So, for example, the top leadership of Vodafone has to do a three-day's training program on digital. Among the things they have to do, they have to do a little bit of programming, and a little, very tiny stuff, very easy stuff, but just to understand how it works, and just understand how they can set up stuff. And then, most importantly to understand how to ask the right questions to the people who really know and are deep in these matters.

Kerr: For society as a whole, how painful will be the implementation of automation in artificial intelligence?

Colao: Well we need to distinguish the long-term with the short and the medium. In the long-term, I'm an optimist. I think society always have wonderful antibodies, and always reallocate resources in ways which are not thought [of] before and productive. But for sure in the medium-term, call it the next 15 years maybe 20, there will be a reallocation of resources that will be difficult, so I think the job that leaders, political leaders, academic leaders, and business leaders have is to find ways to mitigate the inevitable reduction of certain jobs with the slower, good but slower, creation of new job opportunities and new ways of working. So I think the challenge for the next 10, 15 years is really how to create models of experimentation and adoption of technology solutions that create jobs. Not just simply reduce costs and improve profits.

Kerr: What steps do governments need to take to address these enormous challenges?

Colao: I'm not sure it's governments or cities, or what is the administrative level. For sure, it's easier if you go bottom-up rather than top-down. What we see is that certain cities are much quicker, and much more effective at adopting innovation that actually creates higher value in public goods, and in services that citizens get.

For sure, governments, and cities, and everybody should do at least in my view three things. They should, first of all, equip themselves with much stronger technical skills. I think any relatively large administrative or political units must have a technology department that explains and identifies opportunities, not just treats these as problems to regulate.

Second, we need to have a better approach to experimentation, physical, and digital, online experimentation. Contained, we should try to see what can be done, and experiment more. And third, include more startups and smaller companies into the equation. Very often public tenders and public projects are very good for larger companies, but it's very difficult to access them if you're a small company. So I think the three things together, you know a bit more technology savviness, a bit more expertise, a good way of involving smaller companies and experimentation spaces is what probably is the recipe for these next 15 years.

Kerr: Advanced technologies like digitization and artificial intelligence pose significant challenges for society ahead and are also great opportunities for companies. We appreciate Vittorio Colao, CEO of Vodafone, for joining us today to talk about how his company is approaching these new technologies. Thank you, Vittorio.

Colao: Thank you very much, great to be here.