Millions Invested, Minimal Impact
Insurers have committed substantial capital to AI and automation, from underwriting models to claims triage. Industry studies from Simplifai, McKinsey, EY, Deloitte and Swiss Re paint a consistent picture: heavy spending on pilots and point solutions, but a small share of projects ever reach scaled, revenue-generating deployment. Simplifai’s survey, for example, found only a minority of programmes produced measurable returns, and broader consulting research reports that a large portion of AI pilots do not progress beyond testing.
The Real Problem: Strategy, Not Technology
Leaders often treat AI as a technology problem. That is the wrong frame. Insurtech executives cite a fundamental disconnect between AI investment and business workflows. As the Simplifai CEO put it, insurers face a “strategy execution problem” where models sit in lab environments rather than driving daily decisions. The result is a gap between proof of concept and production impact, with data, governance, process and incentives misaligned.
Beyond “Pilot Purgatory”: Keys to Real AI Impact
Moving from experiments to measurable value requires three strategic shifts. First, link AI outcomes to core business metrics and make product owners accountable for adoption. Second, integrate models into end to end workflows so automation replaces manual steps rather than adding more tools. Third, invest in operational tooling: monitoring, model governance, and retraining pipelines that keep performance consistent in production.
Successful insurers treat AI as a capability, not a project. They align executive sponsorship, change management and clear return targets before procurement. They standardise data and measurement and prioritise a few high-value use cases that ripple through the organisation.
For insurance leaders and investors, the takeaway is plain: the competitive advantage will go to organisations that solve execution. Capital is available. The scarce resource is the ability to turn models into business outcomes.
