However, capitalising on the promises of cloud and AI is proving elusive for many insurers, resulting in a “two-speed” market with an ever-widening gulf between those succeeding at digital transformation and those just keeping the lights on. What makes the difference is whether the foundations are in place. Successful cloud and AI transformation at scale requires a single, trusted data foundation, modern flexible technology architecture, digital-first governance and new workforce skills.
Part of the problem with AI deployment is that many insurers are starting with a bottom-up, efficiency-focused, use case-based approach – which initially looks promising. But for those without all the foundations in place, achieving scale is close to impossible. Some insurers find themselves automating one part of the process only to create a bottleneck in another, eroding the business case. Even when data quality is good, insurers often overlook or underestimate the difficulty of automating the risk and control component of AI governance.
One of these challenges is highlighted by the experience of HAAST, an Australian Insurtech using AI to innovate in Marketing Compliance. As Chief of Staff, Jack Tizzard, explains, “A systems thinking approach is required. We map out the workflow end to end with an understanding of all the various different teams, who interfaces with whom, and where the dependencies lie. We then deploy our platform into that workflow across the system in a way that solves the bottlenecks.”
In contrast, leading insurers are getting transformative results from taking a top-down, domain-focused approach to realising benefits from AI. For example:
- Taking an agentic AI approach to the large pools of human resources still used by insurers. For example, contact centres for personal insurance that greatly rely on human-to-human conversation.
- Prioritising high-value applications of AI in underwriting and claims, including ingesting, augmenting, summarising and prioritising submission data. The focus here is three-fold: enhanced service by cutting time to quote from days to hours; productivity improvements of up to 200%; and the ability to prioritise the right risks and claims to get the best outcomes.
In the US, Ascot Group Chief Information Officer, Owen Williams, sees huge potential in the use of generative and predictive AI in underwriting and claims: “Broader adoption of explainable and auditable AI frameworks is required, as models increasingly influence underwriting and claims, regulators and boards expect transparency. Vendors who combine performance with strong governance will lead the next phase of market development.”