Fabrice: Machine learning is a form of predictive modelling. It uses a broad range of algorithms to automate the prediction of key business outcomes based on past experience. There’s quite a lot of hype about machine learning right now, but people have actually been using some of these techniques for years. The big leap forward is the amount of data that’s now available (e.g. claims files, broker submissions, complaints, weblogs, telematic sensor technologies and social media). We’re also seeing a huge rise in the computational capacity and tools that allow us to analyse all this information in real-time and keep improving models as experience deepens. In short, the technological potential has been hugely accelerated and democratised.
From product targeting to personalised services, digital providers have set the pace in machine learning. And now, we’re beginning to see what’s possible in insurance. Machine learning can help insurers to provide digital, personal and relevant solutions for their customers’ needs, and run their operations with an agility that’s never been possible before. The business benefits include more focused sales and marketing investments, more informed risk selection and pricing, reduced claims and operational costs, and more favourable outcomes for policyholders.
Fabrice: Development of machine learning within the insurance industry is still at an early stage overall. Over the next few years, machine learning will move from the data labs to mainstream business operations, and that’s when its full impact and benefits will be felt and business models will be fundamentally changed.
Fabrice: Our primary focus is on practical application rather than obtuse theory. We’re here to help clients use machine learning to tackle real business issues and give them a competitive edge.
We can help clients at all stages of their journey, from exploring the potential and running initial pilots through to augmenting and commercialising their capabilities. Crucially, we can get them quickly up and running even if they don’t have deep data and technological capabilities in place or don’t want to make significant investments up front. For the more advanced organisations, we focus on helping them to scale up and adopt machine learning in their day-to-day operations.
Fabrice: This is a field that satisfies my natural curiosity and love of innovation – it’s developing at such a pace. A good data scientist clearly needs excellent statistical and modelling skills. But I think the most important quality is the business acumen needed to apply these advanced techniques. The people in my team are experts in a range of areas including risk, pricing and marketing. Actuaries make great data scientists as they’re trained in applying models to business operations. But you don’t have to be an actuary – what really counts is curiosity, agility, open-mindedness and understanding of the business. Our key measure of success is to enable our clients to quickly address business issues and deploy business solutions.
Fabrice: PwC is pushing back the boundaries in machine learning and I wanted to be part of that. My colleagues and I are part of a bigger team within Actuarial Services that’s focusing on big data, blockchain, InsurTech and other potential game changers in the insurance marketplace. From partners through to the people we’re recruiting from universities, there’s a zeal for innovation that really sets PwC apart.