Seven months after deploying an AI-enabled contact centre, Tower Insurance says the system has cut the total time customers spend on calls by roughly 796,000 minutes – a 15% reduction across all service interactions. The insurer attributed the outcome to changes in how frontline staff handle routine call tasks, rather than to a single feature of the new platform. Average call length fell by 2 minutes and 38 seconds per interaction.
The contact centre runs on Amazon Connect, a cloud-based platform that applies real-time transcription to every inbound and outbound call – whether the call concerns a sale, a policy query, or a claim. Under the previous setup, representatives were expected to take notes while simultaneously listening and responding to customers. With transcription running automatically, that dual task is removed from the interaction, and call records are generated without manual input from staff.
The change has a functional consequence for service quality as well as speed: details that might be missed or misrecorded when a staff member is typing during a call are now captured by the system as the conversation unfolds. Tower chief executive Paul Johnston (pictured) described the reduction in customer time as evidence that AI implementation does not have to come at the expense of service standards. “When customers contact us, it’s often at moments that matter. Reducing the time it takes to get help, without sacrificing care or judgement, is critical to maintaining trust. This work shows how AI can support our people to deliver better outcomes for customers,” Johnston said.
Johnston also indicated that the investment would continue into the next financial year. “Looking ahead, we’re poised to invest further in technology and AI-enablement throughout FY26. This is a critical part of our strategy that will drive greater efficiencies and continue to enhance our customer experience over the medium and long-term,” Johnston said. Beyond transcription, the Amazon Connect platform also provides automated quality monitoring and AI-assisted prompts for agents during calls. Tower has extended the use of these tools past the initial trial period, indicating the company intends the deployment to be permanent rather than experimental.
Read next: Aon unveils AI claims platform
Tower was one of seven organisations globally chosen to take part in an inaugural Amazon Web Services (AWS) Partner Led Customer Success pilot for Amazon Connect, and the only insurer from the Australia-New Zealand region to be included. The pathway into the pilot ran through Deloitte New Zealand. AWS selected Deloitte as its implementation partner based on the firm’s experience with cloud and contact centre projects. Deloitte then put Tower forward as a suitable candidate given the insurer’s scale and operational setup. Kylie Bryant, a partner at Deloitte New Zealand, said Tower’s willingness to commit to the project contributed to the outcome. “We see Tower as a genuine leader in digital transformation. Their commitment to innovation and customer centricity is setting a new benchmark in the insurance industry and beyond,” Bryant said.
Tower’s contact centre results are part of a pattern taking shape across the New Zealand insurance sector. Machine learning, data analytics, and intelligent document processing are now in active use at multiple local insurers, with AI applied across underwriting, claims handling, and customer-facing functions. IAG New Zealand – which sells insurance under the AMI, State, and NZI brands – formalized a multi-year arrangement with Google Cloud focused on data and AI tools, with claims and sales teams among the primary users.
Suncorp New Zealand, operating as AA Insurance in the consumer market, has taken a parallel path. The insurer built two internal AI tools: a claims management application that won the Financial Review AI Award 2025 for Ethics and Responsibility and a second tool that uses Azure OpenAI to deliver underwriting reference information to contact centre staff without requiring them to search multiple internal systems.
Among the major brokers operating in New Zealand, Aon launched a claims platform in late 2025 that draws on AI for claims analytics and resolution support, with deployment across its international network – including New Zealand – scheduled through 2026 and 2027. Marsh operates separate AI-driven tools for claims intake, document handling, and fraud analysis. WTW’s pricing and claims platforms apply machine learning models to risk assessment and reserve analytics.
