Suncorp halves disclosure list with images of 9 million homes
Suncorp has claimed an industry first by using artificial intelligence (AI) to crunch images of 9 million Australian homes, doing away with half the questions asked of prospective policyholders, such as whether their roof is tile or Colorbond or if they have a swimming pool.
Suncorp EGM Digital Distribution Katherine Carmody tells insuranceNEWS.com.au the images, captured using aeroplanes, were bought from Aerometrex and GIC and will be updated every 4-6 months.
US predictive analytics firm Arturo then helped Suncorp deploy AI to crunch and sort the image data, categorising the images and feeding them through to both its online and call centre customer interactions, slashing the questions about property characteristics put to new home insurance business and renewals.
“The questions don't exist anymore. This isn't pre population, this is complete removal,” Ms Carmody said.
"It's only going to get bigger and bigger. We're not stopping here because we know how important this is for our customers, to make insurance simpler and easier.”
Arturo says it trained its sophisticated deep learning models specifically for Australia to deliver country insights, and delivered more than 30 unique property characteristics across almost 9 million records, or 85% of Australia, in under 48 hours.
The initiative has just been shortlisted for a “best use of technology to revolutionise customer experience” Ashton Media award.
“This is an industry first for Australia. It's super exciting in terms of the combination of technology, AI and the customer experience aspect to it,” Ms Carmody said.
“We’ve got a massive amount of coverage. What we can see in those images is in very high resolution.”
The AI models convert the images into data which has been fully integrated into Suncorp's core systems across all its brands.
“If you can imagine flying over Sydney, we have every single photograph of every property and those images are converted into a massive data file, and then we've completely integrated that into the sales process for buying insurance and renewing. So it's very, very detailed and we know all that information without having to ask customers,” Ms Carmody said.
This both improves risk assessment accuracy and alleviates customer anxiety and uncertainty over their answers, with precedents like Google Earth helping to quash any privacy concerns.
“Maybe 10 years ago there would be more of a Big Brother kind of feel to it but without doubt, the simplification for customers – if anything, they're expecting more of this, not less,” Ms Carmody said.
“It's become expected because we're an expert and customers are expecting us to do as much work as possible to make it simpler for them. That expectation from customers is increasing more and more.”