Transformation—Artificial Intelligence

Transcript

Narrator

The future of project management is changing fast. On Projectified™ with PMI, we’ll help you stay on top of the trends and see what’s really ahead for the profession—and your career.

For an easy way to stay up to date on Projectified™ with PMI, go to iTunes, Stitcher, Google Play music or PMI.org/podcast.

Stephen W. Maye

Hello, I'm Stephen Maye, and this is Projectified™ with PMI. I'm here with my co-host, Tegan Jones, and in this episode we're talking about trends in artificial intelligence.

Over the past few years, we’ve seen terms like machine learning, predictive analytics and natural language processing become central to the conversation about next-gen business technology. 

But while there’s been a lot of talk about artificial intelligence, the actual impact has been, well, rather limited. Looking forward to the next few years, however, it seems like we may actually see AI start to move the needle. 

Tegan Jones

Well, the money’s definitely there to make something happen. According to IDC, worldwide spending on AI and cognitive systems is set to triple in the near future. They predict investment will grow from 24 billion U.S. dollars in 2018 to 77.6 billion U.S. dollars by 2022.

And if you wanna look even farther out, PwC projects that artificial intelligence will contribute 15.7 trillion U.S. dollars to the global economy by 2030.

Stephen W. Maye

Well, that is a lot of money going in, but when will businesses start to see a return on that investment? 

Tegan Jones

That’s a tough question. But a report by Accenture recently predicted that between 2018 and 2022—so that same time frame we were just talking about—investments in AI and human-machine collaboration could boost revenues by an average of 38 percent across sectors. 

But to see those types of results, companies will have to go beyond just automating processes and invest in AI that can really communicate and learn.

Stephen W. Maye

And, you know, this is particularly relevant when you’re talking about using AI within your customer service platform. I mean, few things are more frustrating than trying to communicate with an automated system that absolutely refuses to understand you.

But AI has the potential to bring so much more to the table. And that’s something I recently discussed with Oliver Broom, who’s the program director for artificial intelligence and data and analytics for Virgin Media in London. We’ll hear his take on how AI is transforming the customer experience a little later in the show.

Tegan Jones

We’ve also got Sindhu Joseph, who is the founder and CEO of CogniCor Technologies, a global company that builds AI-powered interactive assistants. She talked about what it takes to define and then deliver a clear return on investment as an organization integrates AI technology. 

But first let’s hear from Veronika Sokolova, who’s an intelligent automation program manager for British Telecom in London. She’s gonna talk a bit about where BT has been deploying its robot army—and how it’s helping to streamline global operations.

[musical transition]
Tegan Jones

The race to automate is on. But being the first to cross the finish line isn’t always enough to beat out the competition. While the excitement around artificial intelligence can be contagious, Veronika Sokolova says the time has come to start turning the hype into hard numbers.

Veronika Sokolova

I think today people are stepping back, and rather than looking at it from that point of view of having a solution—what’s the problem—people are asking themselves: What’s my biggest problem? What’s my biggest challenge I’m trying to solve? And which one out of AI solutions out there going to help me to solve this problem?

Tegan Jones

Veronika says the first phase of BT’s robot rollout was focused on automating simple tasks across departments. But there’s only so much low-hanging fruit.

Veronika Sokolova

And then suddenly, you slow down, because it starts becoming very difficult to find those small pockets of processes which feed all the criteria in your matrix—you know, the ones which are repetitive, the ones which don’t have many exceptions, the one which don’t have human interaction to them. And this is where the challenge begins, because on one hand, you are running out of those easy to pick up processes. On the other hand, your C-suite is expecting to see your army of robots grow because that was promised three years ago.

But in this three years, we did get to the point where we have dozens of robots which have freed hundreds of people from doing very, sort of, routine and repetitive tasks. We’ve given those tasks to robots, and now these, you know, hundreds of people they can focus on something more value-add, more interesting, more intellectually stimulating. 

Tegan Jones

But she says getting to the next level will take a more holistic, end-to-end approach to automation. 

Veronika Sokolova

Rather than looking at a process and saying, "Oh, I can automate this 20 percent, or this 10 percent,” we are saying: “Before we do anything, before bring my robots, let’s just look at the process. Does it make sense the way we are doing it? Is it being implemented the same in the U.K., in Hungary, in India, in Indonesia? Could it be more consistent?”

And my job is to make it happen and ensure that we take the AI tool and we use it to solve the problem. Because it’s all good and well having a machine learning tool here, or a couple of robots there, but what’s impact on the bottom line? And that’s part of my job as a program manager, A) to ensure it happens and it works well once it happened, and B) to ensure that there is a quantifiable benefit which I can handshake with our finance team and agree that look, you know, it’s been a success. 

And you need to be very clear with your senior management when and how they will see those benefits coming in, so that everybody’s expectations are aligned. 

[musical transition]
Stephen W. Maye

That seal of approval from the finance team is one of the clearest measures of success that a project team can shoot for. And I think Veronika made some great points about how you get to that final handshake. Especially when you’re working with cutting edge technology, it’s all about setting and managing expectations with stakeholders up front.

Tegan Jones

That kind of clear communication also really helps project teams move quickly in the face of change. You know, right now there’s a big demand to be adaptable, so that the organization doesn’t fall behind. 

Stephen W. Maye

You have to stay competitive, but you also have to be really aware that these types of innovative projects come with a lot of uncertainty. If you don’t take that extra time up front to focus on your plan and assess exactly what you’re organization’s capable of, you run the risk of seriously disappointing your sponsors.

Tegan Jones

But our next guest has a few thoughts on how to manage this uncertainty while still making big moves. Sindhu Joseph is the founder and CEO of CogniCor Technologies in San Francisco, and we’re about to learn more about the process she uses to determine the potential value of AI for her clients. Let’s go to her now.

[musical transition]
Sindhu Joseph

Hearing and reading about artificial intelligence and its benefits are, it’s interesting, it’s exciting, it’s innovative. But exploiting that technology and incorporating that into your business is a completely different story. So, what we have been seeing in the past three, four years is that since the world of business is not very educated about AI, it was quite hard for us to deploy this kind of technologies and derive the ROI that the business is expecting. Because the expectation from this technologies are so much hyped that the business leaders have no idea what AI provides. 

The very fact that AI is an intelligent system, that means that you cannot get 100 percent of the time the response that you are looking at for the same kind of input. So this kind of uncertainty is very unsettling for project managers that are not used to this kind of project, and that means that they have to work on people’s mindset, the processes within the company. I’m talking about the processes because all of these AI-driven technologies involve a huge amount of data. And the quality of data kind of decides the quality of implementation. So if we don’t have a great process to get the best quality data inside these algorithms, then it becomes a junk in, junk out kind of system. So, those are the things that we need to focus while deploying this kind of solutions.

So when I say, you know, the projects can deliver ROIs for the businesses, it may not be always immediate delivering of ROI. So if you start measuring the ROI the moment the system is implemented and in place, then you might be up for disappointment. I’m not saying that this is always the case, but majority of the AI projects needs to have certain time frame for them to make sure the AI to learn based on the customer interaction, learn about the characteristics of the user base that is interacting with the system and then based on that, on a longer period of time, derive the ROI.

One of the examples that I want to give is in one of our implementations. After six months of implementing the system—this is a lead generation system for a mortgage selling business—and we were able to generate the leads and were able to increment revenue for that business by 100 million. So that is a huge benefit, but it takes time to get there. 

So if you look at past couple of, maybe a decade or so, Google has been one of our, you know, the go-to interface for everything. So we search for information, and we get the information. But if you look at human nature we are always, you know, by nature, conversationalists. We would like to make conversations. And somehow Google has made us step away from our natural instinct to converse, and I think the AI technology is going to bring back that conversation into our business world. 

[musical transition]
Stephen W. Maye

It is nice to hear Sindhu say she thinks AI’s going to bring a bit more conversation back into the customer service experience. One of the big struggles with this type of automation so far has been that even the most advanced AI—take Alexa or Siri—it still lacks that human touch, that ability to process conversational requests the way we really expect it to.

Tegan Jones

Yeah, there’s definitely still quite a lot of pain points with this technology. But I think AI is growing up. As Veronika mentioned earlier, a lot of simple or repetitive tasks are already starting to be taken on by machines. And that’s having a big impact on the workforce.

According to that same Accenture report that I mentioned earlier, nearly half of executives say traditional job descriptions are becoming obsolete as machines take on more routine work. And nearly two-thirds say that the share of jobs requiring collaboration with AI are going to increase in the next three years.

Stephen W. Maye

That definitely translates to some big changes for project managers and team members, as well as the people in charge of hiring and training that talent. 

And I recently got some insights into what that looks like from Oliver Broom, who’s the program director for artificial intelligence and data and analytics for Virgin Media in London. 

Tegan Jones

Let’s hear what he has to say.

[musical transition]
Stephen W. Maye

Oliver, obviously artificial intelligence is showing up as a theme and as a trend and set of trends in essentially every sector of business and personal life right now. What do you see in your sector as some of the most significant trends that are emerging?

Oliver Broom

I think within telecoms, the biggest trends, they’re all around customer experience, right? Transforming the customer journey. And that is where a lot of the focus of AI is right now. It’s a very competitive marketplace, like most in technology. There’s a lot of players, particularly in Europe. Margins are slimmer than they used to be. And there’s a lot more competition for end users.

So, I think where AI is kind of playing in that space is around this idea of conversational artificial intelligence. And this is about transforming the customer into action. So we see it in the use of messaging bots, with virtual assistants. We see it in things like speech analytics, where artificial intelligence will look at the voice pattern of someone making a call into a contact center, and then make some very accurate predictions about the state of mind of that person. Perhaps they’re calling to make a complaint, and artificial intelligence can pick up on that very early in the conversation and flag it to an agent, and help to steer that conversation to a better resolution for the customer.

Stephen W. Maye

You know, I think we’re kind of very comfortable with this idea of yes, please model that engine and figure out where a component is going to fail. That’s great. Predict the impact on your network so you can get out ahead of that, from a development perspective.

But we may get less comfortable when there’s the idea that the AI is going to, is going to assess my mood and my intent. Where do you think we are, as sort of a society of people, in how we’re responding to or are prepared to respond to AI getting into our headspace?

Oliver Broom

I love how quickly we got into this topic, because it is so important. AI is largely tech-driven. But the softer side of things is absolutely critical.

I think, where we are as a society, you really have to break it down into demographics. And it’s very, very difficult to do. There are lots of people who will give away their data for a free coffee. But then there are a lot of people who are deeply uncomfortable about that. And it’s a very personal perspective, because it really depends on how you feel about your data, and what that data says about you, and the context in which you exist within your society.

My personal view—and it may not be shared by everybody—but I think you have to decide where you are, particularly as a PM if you’re involved in these kind of topics, where you sit on that spectrum. If you’re doing projects that implement these kinds of tech, you will come across that broad spectrum of individuals who believe deeply in what you’re doing, but are also deeply against what you’re doing. And I think you have to decide where you are on that spectrum. And whether you want to be part of that conversation, or whether you just want to be black and white about it, and say, “I’m just delivering the box.”

Stephen W. Maye

Can you share an example, Oliver, of where AI is showing up in a current or recent Virgin Media project?

Oliver Broom

Yeah. I think about the AI road map—the vision for what AI could bring in the future—there’s this idea that’s coming through now called dynamic case management. And the case management piece is where Virgin Media is making some investment now over the next couple of years. It’s a very significant pillar of what we’re doing over the next two to three years. And this is really about treating every interaction with a customer as a case, and being able to raise it and surface it, anywhere in the business, and across omni-channels, so over messenger, over the phone, online being a huge one for our customers. And it’s kind of the first step to aggregating a lot of this information around customer interaction, which in the future we’ll then be able to do data analytics on. And we’ll also be able to service our customer interactions in the way that’s best for our end users.

So if they like to use online, then they’ve got to switch to the phone, and they get in the car and go to work and it becomes email. That case, that journey or that conversation, goes with the customer in the best way possible for them. That will require a lot of data analytics. There’s a machine learning part to that around how we do price analysis. And that’s kind of the main thread of our AI road map I think, over the next few years.

Stephen W. Maye

What else do you see in that development and continued evolution? If you go out three to five years, whether you can speak directly about Virgin Media—and that’d be great if you can—but at least in your industry or in your sector, what are we going to be experiencing five years from now that would be perhaps very different from the experience we have today?

Oliver Broom

Let me talk about companies like Virgin Media. That’s probably fair to say. Because there’s so much uncertainty about exactly where these things are going to go, but there are some key topics that are absolutely going to be pursued.

I think the first thing to say is the most successful implementations over the next few years are not going to be laser-focused on artificial intelligence as a platform or a box. Although there are companies out there that will try to sell AI as a platform or a box, I think they’re going to be more a feature of the network, an output of the architectural design decisions. So kinda one tool of many. But I think it will become very focused on automation and prediction. What is the point of creating more data and storing it in a more intelligent way if you’re then not gonna do anything with it? What can you do with that data once you’ve defined it and stored it in a more intelligent way? We can automate a whole bunch of stuff that in the past was very labor-intensive.

So a lot of the tasks that are being done across the business, in every area, that take up your best people’s time to do, very manual work, will for sure be automated. And that will probably happen very quickly. But then on the prediction side, what will happen is with these data pools, these data lakes, we’ll be able to draw insights with quite high degrees of accuracy, where you can then start to predict things that will happen. So, for example, we might predict during certain events that there is an increased likelihood of generating network faults. 

Stephen W. Maye

How are you staffing up to deal with this anticipated future? And of course acknowledging that you’ve even said, we’re not clear exactly what that future looks like. We’re talking about trends and directions, and some of that’s going to have to take shape over time. So how are you staffing up for that future?

Oliver Broom

If you’ve worked in IT project management for the last five years, you’ve probably got the fundamental skills to project manage projects that will become very AI-focused in the future. So the kind of recruitment decisions we’ve been making over the last few weeks is very much focusing on the fundamentals, the core project management skills still.

But there are a couple of things I kind of look for in project managers. One is to have that enterprise-wide experience is really good. Because for some reason—I mean I think I know why—but for some reason, projects that have kind of a sniff of AI or data analytics, they tend to go up that strategic ladder quite quickly and get a lot of focus from senior management and become very high visibility. And they tend to reach out quite far across the enterprise. Because the potential is that they could have a wide-reaching impact, and when you start to aggregate data.

So a skillset which is, has an experience of working across enterprises is really good. Senior stakeholder management skills: really good. Because there’s probably going to be a lot of that in these initiatives.

Stephen W. Maye

So if you think about someone who’s new in the space—so just beginning or wanting to move into the artificial intelligence space, projects and programs that have a significant artificial intelligence component—what’s your single most important piece of advice for that project or program manager that wants to move that direction?

Oliver Broom

One piece of advice. Well first of all, I would say “good luck,” definitely. But one piece of advice, I think, is just… God, I’ve got so many of those, Steve. 

Stephen W. Maye

I tell you what: If you wanna list two or three, then we may just force you to prioritize. But sure, go ahead.

Oliver Broom

All right, let’s do three then. I think be very clear about why the project exists. I know you don’t necessarily have any control over that, but is it part of a hype cycle? Is it delivering on a strategic outcome? Can you stand by the reason that it’s being done, and can you believe in it? I think identify whether it’s linked to a strategic plan from the business, and what is the expected impact.And I think, also, be prepared to reach out across the business. Because AI projects, they tend to have quite a broad reach. Because if they’re dealing with data analytics and these kind of topics, they’ll be probably hitting on systems that maybe haven’t inter-worked before. They might be pulling down data from a platform and using it somewhere else, in a very different way. You might be very used to working with a good team. But be prepared to widen your stakeholder engagement, because it’s very likely that you won’t be working in a black hole on AI. Everybody’s gonna wanna be involved.

But yeah, watch the scope creep. And make sure that where you’re interacting with new platforms that perhaps you haven’t interacted with before, and you’ve got new teams there to work with, that you acknowledge that, because it could be a slightly different environment to what you’re used to.

Number four, which I wanna sneak in, is everyone will be interested in what you’re doing, if it’s artificial intelligence. So you’ve gotta handle that, too, from a comms perspective. Not everyone needs to be involved, but you might want to think about flying the flag for your project a bit, because people will want to know what it’s about, because it’s kind of like the hot new thing. So get internal comms involved if you’ve got them or do something yourself, because that will probably only be to your benefit. And that was four, Steve, sorry.

Stephen W. Maye

That’s okay. Three, three and a bonus. And with that, Oliver Broom has the last word.

We’ve been talking with Oliver Broom, program director for artificial intelligence and data and analytics at Virgin Media, and he is located in London, England. Oliver, it has been a pleasure. I have thoroughly enjoyed talking with you.

Oliver Broom

Yeah, me too. Thanks a lot, Steve.

Narrator

Thank you for listening to Projectified™ with PMI. If you liked this episode, you can subscribe on Apple Podcasts or Google Play Music. We'd love your feedback, so please leave a rating or review.