On the final day of the Business of Data Festival, Australian insurance firm Suncorp’s Group Head of Data Science and AI Centre of Excellence, Craig Price, shared insights on the key ingredients that go into data and analytics capabilities
Enlisting highly technical people and harnessing cutting-edge technologies important when it comes to building awesome data and analytics capabilities within enterprises. But they aren’t the whole story, according to Craig Price, Head of Data Science and AI Centre of Excellence at Suncorp.
The Australian finance, insurance and banking corporation’s executive spoke about AI opportunities and practical approaches to building data and analytics teams at the first Business of Data Festival in July 2021.
Asked what some of the great opportunities around rolling out AI were for organizations, Price said a simple but powerful example concerned improving customer experience in the digital age.
“Organizations are increasingly trying to get customers to use their digital channels,” he said. “That takes load off the contact centers and also off stores. That’s a good thing in many cases and corresponds to customer demand post-COVID-19.
“A lot of customers across age groups, including the older age groups, are more and more comfortable with interacting digitally with financial services organizations and other organizations.”
Part of the shift to digital channels includes the rollout of chatbots or intelligent virtual assistants (IVA), which Price said were great innovations toward improving the customer experience of navigating storefronts and supplying information to businesses.
“The underlying engine and intelligence-driven by data and analytics is, first of all, going to recognize the customer when they turn up, and it will remember previous interactions,” he said. “It’s a great customer experience because you don’t have to retell your story or key in what you’ve already done.”
One example Price said was simple but effective was a change-of-address function, in which a customer can inform an AI on arrival to the site that they would like to update their address details and provide that information to the bot without needing to navigate to a particular subsection of a page.
“A more sophisticated example might be if a customer is at high-risk of lapsing a particular product, then for us to be aware of that and for an experience to be appropriately incentivizing them, from providing a message, a nudge or a real offer,” Price said. “The next-best-action or next-best conversation engine is very much going to have its time in the sun over the next few years.”
Cornerstones of Building Better Data and Analytics Capabilities
People, process, technology and culture are the foundational cornerstones of a great data and analytics capability within any organization. Price said he shares this view in his data leadership position at Suncorp.
Starting with people, he said, while touting the advice of ‘getting the best people you can’ may sound obvious, organizations are still regularly constrained in their ability to do this.
“I think that sometimes budgets or even just time pressure around trying to get people on board can constrain,” he said.
To get around this, Price said he has bought into the 10X developer theory, the notion that individuals at the top levels of performance in a particular technical discipline can produce up to 10 times the result of an average performer in terms of outcomes.
“It really is worth your while to try and get those types of people for key roles,” he said. You obviously can’t do that for every position in the organization. And if you have a fixed budget, sometimes less is more. A smaller team of higher capability individuals can get you better outcomes.”
Price said a good way to go about recruiting, in this case, was practical testing; having a candidate demonstrate their ability to approach a real business problem and solve it.
“That’s where we are focusing, because a lot of people that we see, especially in our principal data scientists, they’ve got PhDs, they’ve got fantastic skills but, really, it’s the application of those skills that separates the really great from the good,” he said.
Going Beyond Technical Proficiency
Many data and analytics executives would agree that it takes more than technical capability to initiate data-driven change within an organization, and Price acknowledges that other skillsets are definitely critical.
“You can have the best technical teams but if they can’t communicate and can’t influence business stakeholders, then it’s going to fail,” he said.
Of course, the more high-level skills and skill diversity sought with new hires, the fewer and further between qualified individuals are going to be. Finding ‘unicorns’, as they are known, remains a challenge for all kinds of businesses.
“The simple response that we’ve found effective is, by all means, if you stumble across a unicorn grab them,” Price said. “But if not, then you’ve got a team and you can actually get the mix of skills across the team.
“We’ve had senior leaders in my team that span from very technical PhDs right though to really good business analytics professionals with fantastic consulting skills, and we’ve seen those skills come together very well. It lifts the game of the entire team as we do that.”
Installing the Right Process and Technology
Price believes a data science platform should form the backbone of the process and technology aspect of building great data and analytics capabilities in business.
“I do see really great value in having a data science platform,” Price said. “One of the key things you need from a great data and analytics team is the ability to prototype rapidly. It’s all about trying to find value in what you’re creating, rather than just a cool solution.
“[It’s beneficial] having a platform and a set of processes that allow for rapid prototyping, and then, if it’s valuable, being able to take that to a production asset without having to throw it all away and reengineer everything from scratch.”
Price said that two-phase capability had been particularly valuable to his team, adding that they had initially tried to build a platform themselves, before shelving the idea.
“We realized that we didn’t have an army of engineers and we weren’t going to get there,” he said. “So, we’re utilizing one of the third-party vendors, using their workbench to allow our data scientists to get the high-capacity compute they need and then to collaborate.”
Building a Data-Driven Business Culture at Suncorp
What does culture mean from a data science perspective? Price agreed with many of our other festival speakers when discussing this topic, citing two factors as being particularly important. The first, he said, was around business focus.
“For data science and AI to be most effective, I don’t believe in coming to the problem with no context,” he said. “I encourage all of my teams to have a very good understanding of the business process or the operational process that they are trying to fix or optimize. If they have that, they are able to devise a solution.
“When people haven’t got that context, they can go off on a tangent or not bring the right data to the problem, or somehow not be aligned with what the business stakeholder is trying to do.”
Price added that the second requisite cultural element for data and analytics teams is being value-focused and seeing beyond the aspects of the job that some data and analytics professionals find more immediately exciting.
“All of us from technical backgrounds love new technology, new methods and cool shiny new toys to play with,” he said. “But what it comes down to is the right technical solution that is going to demonstrate the value best.”