Using science to make business decisions

How leading insurance company AIG is taking an exciting new science and technology-led approach to the way it operates
Other finance

AIG has put science at the heart of its business. For the past few years, the global insurance company has been building up its "Science Team" - expert computer scientists, researchers, analysts and businesspeople charged with driving AIG's strategy by ensuring the business is making rational science-driven decisions - and asking the right questions.

Here David Pardoe, EMEA Science Business Partner, explains more about the Science Team's work, how it fits into AIG's business strategy, and the opportunities there are here for graduates.

How would you introduce AIG's Science Team and how what they do fits into an insurance business?

What the Science Team does is all about using advanced data science and machine learning techniques to change the way decisions in the business are made - historically AIG and other insurance companies haven't done this.

The Science Team was set up from scratch in 2012, but now consists of over 150 people globally, with a large number of these employees having come from outside AIG and outside the insurance industry.

We've recruited employees, and hence techniques and ways of thinking, from industries that are significantly further forward than insurance in terms of using data to influence how they do things - for example, retail banking, mobile technology, and of course the big technology companies like Amazon, Ebay and Google that have been at the forefront of this area.

So the challenge for the Science Team is work out the best way to apply these tools and techniques to the vast amount of knowledge and experience that AIG has built up over the years around insurance and how to sell it to customers.

Why was the Science Team set up?

The creation of the Science Team was driven by Peter Hancock, who was the chief executive for the property and casualty part of AIG's business at the time, and who was forward-thinking enough to recognise that utilising data and understanding what it can tell you about your business should be an integral part of AIG's strategy in the future.

When he created the Science Team, he asked them to report directly to him, which we still do - very few large companies have our kind of function reporting directly to the chief executive and it shows how influential the Science Team is on the business as a whole.

What are the main types of work that people in the Science Team do?

Broadly speaking, we do three types of work:

"Quick win projects", which are about identifying a business objective, building analytical models that address that objective, and getting the results of this analysis into the hands of people who make business decisions, all very quickly. These projects comprise about 20-30 per cent of the work that we do.

"Transformative big hit projects", which are about taking ideas and thoughts from quick win projects and turning them into large scale programmes of work. For example, a transformative big hit project might take the findings from a quick hit project focused on a specific market in a specific region or country and implement them globally. These projects comprise about 40-50 per cent of the work that we do.

Finally, there are research and development projects, where we aim to take cutting-edge, innovative approaches to problems that have never been tackled before.

Working on these projects we have people with different specialisms - for example, the team includes data experts, quantitative analysis experts, data communicators who translate complex data into accessible formats, business partners who work with stakeholders across the business, and even psychologists.

Can you give an example of a piece of work the Science Team has done recently?

A good example is a predictive model we've developed that uses internal and external data to forecast whether or not certain of our commercial clients are likely to renew insurance policies they hold with us.

It provides an early warning to those who manage these client relationships and indicates which clients it's worth making extra effort with.

How do you see the work of the Science Team developing in the future?

Generally we're looking to turn our quick win projects into transformative big hit projects, so taking techniques we initially use in one market and applying them more broadly. It's an exciting process - starting with small-scale projects and turning them into massive ones with global reach.

In terms of the research and development part of our work we're focusing on researching the latest advances in technology and working out how they can be applied in insurance.

If insurance companies want to compete in the future, they've got to embrace the latest technology, processes and tools. We think we at AIG are in some ways ahead of the game, but there is still plenty that we could develop in this area, so there are huge opportunities in front of us.

What opportunities are there for graduates in the Science Team?

We look for graduates with experience of areas like statistics, machine learning, computer science, mathematics, econometrics - generally anything data and numbers-orientated.

Even graduates with advanced degrees will spend a good proportion of their initial time learning more about these techniques and methods and how to apply them, not just to our data, but also to our business problems - we're most interested in how data can be used to inform decision-making.

Graduates will then progress through a relatively structured framework and can choose to become an expert in a particular area or branch out, as they grow and develop within the team.