Tower: thinking ahead with AI and machine learning

Born and bred in New Zealand, Tower Insurance has been supporting Kiwis when they need it most for more than 150 years. It now serves more than 300,000 customers – across New Zealand and eight Pacific Island nations – providing cover for their houses, cars, contents, businesses and more.

Tower has become a household name in insurance thanks to its focus on innovation and investment in large-scale digital transformation, which has enabled the business to evolve rapidly and keep pace with the latest in technology and customer expectations. Award-winning customer service is another point of difference that the Tower team pride themselves on.

The insurance industry is highly competitive, and customers tend to move around frequently, so finding ways to keep customers happy is a win-win proposition for Tower. ‘Thinking ahead’ isn’t just a marketing tagline; Tower is at the forefront of the industry with its use of artificial intelligence and machine learning.

Collectively, the business has millions of data points about claims and customers – it’s an almost overwhelming amount of information. The challenge is to be able to sift through that gigantic mountain of data to reveal useful insights that can help Tower improve its customer experience.

“Yes, we’re an insurance company, but we are a digital and data-led business. We are passionate about data because it’s through analytics and digital insights, that we can improve the experience for our customers,” says Gurbani Singh, Head of Data, Analytics and AI.

“We can learn a huge amount about customers – knowing why they choose us, why they leave and what happens in between.”

In late 2021, Tower came up with an idea to see if it could get more from its existing data. The team decided to start with a small-scale data analysis, focusing on motor insurance policies, since customers switch providers frequently when they buy or sell a car.

Tower used IBM’s Cloud Pak for Data to analyse data over a six-week period, with Datacom providing the supporting infrastructure. The immediate aim was to improve Tower’s understanding of the factors, events and insights that drive improvement in customer experience, says Ramy Amer, Customer Success Manager at IBM.

The longer-term focus, says Ramy, was empowering the Tower team to apply what they had learned in proactive ways: “to make proactive attempts to retain and improve the experience for customers, then to ultimately put practices in place to intervene and keep customers. And to measure that improvement in the future so Tower can optimise it and keep raising its game.”

The proof of value project lasted for just six weeks but captured four years’ worth of claims and customer data from its motor insurance policies. After the six weeks were up, the model had undoubtedly proved its worth.

“We picked some core data points we thought would have a higher likelihood of predicting churn,” explains Kieran Simmons, Tower Customer Experience Owner.

“Within that six-week turnaround, we not only had a better understanding of how the model worked, but we were also able to get some valuable new insights – and this is just a drop in the bucket of what we could feed into this model.”

Tower, IBM and Datacom worked together toward a single goal, which helped deliver excellent results in a short timeframe.

“This was a joint venture and we couldn’t have done this without IBM and Datacom,” says Liz Cawson, Head of I&T Operations.

“This was brand new technology that we were not familiar with inside Tower, and when we reached out to Datacom for help, they were pretty damn swift in coming to our assistance . Datacom really enabled us to explore what this technology was capable of.”

"We're really proud of the work our team did on this project with Tower and IBM," says Tim Moor, Product Portfolio Manager for the Datacom Cloud team. "We were able to deliver at pace and help drive innovation with an Open Shift platform underpinned by Cloud X, our hybrid-ready cloud.”

"The vision for Tower is that a better understanding of the causes of churn and customer dissatisfaction will enable them to put better support processes to help customers, as they are interacting with Tower's frontline staff and making decisions about their insurance cover. We really enjoy being part of projects like this where you can see a tangible benefit for the company and for the customers they serve."

And the scope for further analysis seems almost unlimited. With so much data available, Tower is now considering how it can use the model to analyse its other products and focus on other data points to uncover further analysis. The team is also considering ways to turn its discoveries into actions for the customer service team.

“We’re so enthusiastic and excited,” says Gurbani. “If we can do all this in six weeks, imagine what we can do in a year.”