Philip Armstrong joined Great-West Life over two years ago and hasn’t looked back. Actually, Philip never looks back, choosing instead to live by the motto "respect the past but embrace the future." In his interview with Emerge, he explains how success today depends on your ability to think ahead.
Philip, what's the next big thing IT leaders will embrace, and why?
All forms of automation. It’s really as simple as that.
Some of the simplest forms, like repetitive task-type automation, are showing up in environments in the form of bots. Today, bots are pretty dumb—they just do what they're told. But for low-level tasks, they automate quickly, they have repeatable results, and they're very cost-effective.
We’re starting to see robotic process automation platforms coming into all aspects of large, diverse organizations, from corporate functions to business, to operations, and into IT. And they're great, but I think we're on a path with bots. In the next few years, the largest two or three global suppliers of this type of technology are going to introduce AI into the bots themselves.
From there, you’re going to start to see an environment where these rather dumb process bots start to think and act, learn and adapt. If you can imagine bots on a process line, like an automobile or a factory line, there might be bots behind the bots that are monitoring processes and adjusting based on workflow. There'll be other bots that will be stripping off metrics and creating dashboards and management reports automatically. And there'll be other bots that are counting widgets and spotting trends and feeding other systems. Finally, there'll be supervisor bots that manage the deployment and the workforce of the bots.
Where are the humans in this equation?
Well, this is really going to eat into a lot of the manual labor that's deployed today and reduce that cost. And for a lot of large organizations, labor and salaries with fully loaded costs are their largest costs. So, most large, people-intensive companies are moving into automation in a big way.
As you move up the value chain, you start to see investments in another form of automation that impacts things like big data analytics and algorithms. In this case, automation is sifting through huge amounts of data (volumes that can be overwhelming for a human analyst) to derive trends and inferences from that data.
These types of automated technologies can quickly identify gaps in the marketplace via algorithms and big data analytics—underserved areas where you have opportunities to introduce products, services that are working, services that are not working, and future trends before they even happen. This allows you to anticipate your customers' needs.
Automation can even enhance cyber-security by sifting through very, very large log files and determining false events versus actual real events. We're very good at collecting and storing this information, but without this kind of augmentation, it's becoming overwhelming for human analysts to use it.
As you move closer to the customer, automation pops up again in the front office. Now we’re talking about things like conversational AI (artificial intelligence). In past years, we had a dumb chatbot, but today chatbots are becoming more intelligent. They can actually understand the context of words in the placement of sentences. They can hold a conversation with you and are closer to passing the Turing test. They can sense mood and adjust their responses. They can operate across different channels, whether web, text, mobile, or even voice. They can talk to you in over 40 different languages. And they're available 24 hours a day, 7 days a week. They don't ask for raises, and they don't take bathroom breaks or personal holidays.
These technologies are getting very powerful and more sophisticated. Instead of that robotic voice that you can easily trip up because you know you're not talking to a human, they're becoming more and more sophisticated. It’s possible that in a year or two, you won't actually know whether you're talking to an automated attendant or a real live human.
It’s possible that in a year or two, you won't actually know whether you're talking to an automated attendant or a real live human.Philip Armstrong CIO, Great-West Life
I feel as if I’m listening to a 1950’s science fiction novel, one in which the sentient robots take over the world. It’s really not that far off.
No. Because of course, when you string these three types of automation together, you’ve got automation in the front, middle, and back office.
What if the bots actually talk to each other? What if, for example, when you phone into the help desk, you're talking to a conversational AI system that is understanding from the conversation what you need and what service you require? What if, as you’re talking, they are profiling you using big data analytics and kicking off an automation bot in the back office that actually performs the operation you're calling in about? And then, once your problem has been solved automatically, the bot then positions another product for you in an upsell or a cross-sale capability? All of that is done without the aid of a human.
Given automation’s potential to solve a problem from start to finish, what kinds of skills are you going to need in your IT department in the future? And how do you go about developing those skills now?
That’s a $64,000 question, isn't it?
Some studies around AI skills were just done across Canada recently, and they projected out the demand versus the supply for AI skills coming through university programs in the next two or three years. There is a massive, massive gap between supply and demand, and not in a good way.
»Related content: Building AI that can build AI, The New York Times
Anyone that's got AI skills can write algorithms. But to succeed, they also need to know how to apply it to business models. To sit with one foot in the AI world and another foot in the business world, with the ability to communicate between the two—that is the premium skill that everyone's looking for right now. Business translation skills. People are hiring a few of those people and using them to mentor and grow their own workforce.
It’s also important to partner intelligently in order to leverage your R&D spend. You have to pick the strategic IT partners who are right for your organization. We have four primary technology partners, very large global tech players, and we leverage their skill sets, their knowledge, and their R&D investment. You can't do it on your own these days.
How do you decide what to outsource and what to keep in-house?
That depends upon the risk level within your organization. I work in financial services for what started as an insurance company, so risk management is completely ingrained and embedded in our organization. We have a fairly low tolerance for risk. But even so, we've selectively chosen applications that are best suited to the cloud. We've selectively chosen applications from software-as-a-service providers. And we've deliberately chosen other things that we want to run in-house in order to retain a tighter control.
I think having an aspiration to be 100% cloud or 100% outsourced is not realistic for some business models. It depends upon your appetite for risk and how much you’re using those applications as a strategic advantage versus commodity. Some decisions to outsource are purely based on, "Well, it's not a sustainable model to run that type of capability in-house so we're going to leverage our partners." And then of course cost is a big driver too.
How are automation and AI going to impact your industry and what you do?
For the industry, it's really going to shift the skill set of our global workforce. Not just at Great-West Life but at all companies in the financial services business. As a result of automation, we're probably going to see a decline in lower-level jobs and a demand for higher-level skills. That may cause a bit of a talent war.
For me personally, I have CIOs in each country and region and about 3,400 IT staff globally. My job is to survey the technology landscape, understand changing business models and customers' needs, and partner with my internal executive peers in the business, to listen to them and understand their challenges and opportunities. The real work is marrying up emerging technologies like AI and machine learning, and providing the explanation and education needed to understand and leverage what these technologies can do and how they potentially redefine the business strategy.
In the past, we used to do a business strategy maybe three or four years out, and then work to underpin the business strategy with technology. That model is a little outdated. Today, when we create a business strategy, we try to bring powerful emerging technologies into the conversation because those technologies can reshape the way that we look at our business strategy. They have to be evaluated at the same time, in partnership with business peers.
And that's my goal: to educate my business customers around the art of the possible and the ways in which we might be able to integrate these technologies into their vision for the business.
You highlight an interesting shift that's happening between IT roles and business roles today. How do you distinguish IT leaders from their business partners? Or, do you?
That's a great question. At Great-West Life, we're going through what's termed as a "digital transformation." At least, that's how we started. But it really isn't about digital.
We've reorganized our business department to be more customer-centric. We've aligned our IT resources and federated them, we've distributed them, and embedded them, so that they're congruent and lined up with business functions. We're creating full stack teams of individuals from the business, individuals from corporate functions, like legal, and compliance, and cyber-security, together with developers and business resources.
And we're creating these cross-discipline teams to work on a specific outcome. So, we're making them outcome-based, and really focused on a new agile way of working. We've seen agile at scale fail many, many times with large organizations that are in different buildings, spread across the country, in different time zones, and so, we've modified agile at scale to be more congruent and aligned with our business partners, and made it more outcome-based.
What that's done, to your point, is completely blurred the lines between what is a business resource and what is an IT resource. We just have resources now. We just have project resources. And we’re bringing their different disciplines, skills, and knowledge to the table to define an outcome.
Tell me more about seeing agile fail in large, global environments. Are you saying agile is most effective on a small scale?
Here's the secret sauce. If you're a startup company, and you have about 30 people in your organization, and you all sit on the same floor of the same building, and you have one product and your goal is to take that product to market, agile is fantastic.
If you're an organization that's comprised of 26,000 people around the world, in different countries, in different time zones, in different buildings, in different provinces within the same country, or different states within the same country, and you have distance between you and language difficulties and cultural difficulties, agile does not work. It's just too big.
There’s an emerging philosophy called "agile at scale," which attempts to overlay the agile model onto large organizations. But here’s where it fails: If you take a team of 12 people and shove them in a room, and say, "You're not coming out of this room for 8 weeks," and here's the outcome that you want, you have to empower the functional representative in that room to speak on behalf of that entire function for the organization.
It doesn't work because when you come out of that room, you've got the boss, or the boss' boss, or someone in another country, or someone that's thinking of something different that will always have a different point of view or a challenge to the decisions that were made in the agile process in the scrum.
To set up the team for success, there has to be a way to check in with what I call "the corporate heartbeat" to ensure the decisions they're making, the representational statements they're making, and the design that they're coming up with incorporates everybody's opinion. It has to be socialized very, very quickly, yet that socialization process in some organizations can take months. Well, that's not agile.
I know an “agile socialization process” is a little bit of an oxymoron, but that's the way to actually get agile working in large, distributed organizations.
Philip, what’s keeping you up at night?
Nothing? You’re not worrying about what’s coming around the bend, in terms of the next wave of innovation?
No. I don't worry. I would go nuts if I worried. I've been in this business long enough that I sleep really well.
I tend to live in the future, thinking about things two or three years ahead and imagining what the world will look like. And then I try to come into work every day and connect that back to today so that we can create capabilities and skill sets and an environment where we're positively positioned to take advantage of the future. I'd rather be proactive and lead the way, creating the future as opposed to reacting to it and always being in catch-up mode.
»Read next: How to navigate digital transformation, by Colin Rowland