The global insurance sector is investing heavily in artificial intelligence but seeing minimal returns, a phenomenon dubbed “Pilot Purgatory.” This key insight comes from a new report, Ultimate AI Strategy for Insurance, by Simplifai, a leading provider of Agentic AI for P&C claims. Using data from McKinsey, EY, Deloitte, Swiss Re, and others, the study highlights a major gap between AI spending and outcomes, while revealing a consistent, repeatable approach among the few carriers actually achieving meaningful results.
With US and European insurers generating approximately $3.3 trillion in premiums in 2025 while combined ratios hover near 99%, the pressure to reduce costs and compress cycle times has never been greater. While AI was broadly seen as the solution, the data tells a more complicated situation.
Record Investment with Minimal Impact
According to the Ultimate AI Strategy for Insurance report, 99% of insurers now have generative AI initiatives underway, and 14% are spending more than $50 million per year. Yet only about 42% have deployed AI in even a single function, and production deployment remains the exception. Most activity clusters around customer service chatbots and document summarization, useful, but not transformative. The least common deployment by far is end-to-end workflow automation in underwriting or claims, ironically where the real financial leverage lies.
The report points to three consistent barriers: finance teams unable to connect AI activity to P&L outcomes; risk and compliance teams navigating the EU AI Act, EIOPA guidelines, and NAIC Model Bulletin without clear frameworks; and the deep integration complexity of legacy policy admin and claims platforms. Together, these forces have produced what the report calls “pilot purgatory”, a cycle of promising experiments that consume budget and talent without producing durable business value.
“Insurance doesn’t have an AI problem. It has a strategy execution problem. Technology works. The business model doesn’t,” noted Artem Gonchakov, CEO of Simplifai
Attaining Results
A smaller group of carriers are seeing dramatically different results. Their common thread is not the use of more sophisticated technology, which every carrier has access to the same models and platforms. The difference is their approach. Leadership at these companies start with workflow and business outcomes rather than technology curiosity. This entails deploying AI end-to-end, rather than as point solutions bolted onto existing processes. In addition, design governance is implemented from day one rather than retrofitting compliance after pilots succeed.
This architecture utilizing “Agentic AI” employs systems that plan, reason, and execute complete multi-step workflows under governance, rather than generating outputs that still require humans to interpret and act on. In practice, it means handling a claim from First Notice of Loss through payment without manual handoffs or taking an underwriting submission from intake through binding with human oversight built in at defined thresholds.
Carriers deploying this way report:
- 30–40% productivity gains in claims and underwriting operations
- 25–35% reduction in cycle time from FNOL to payment on automated-eligible claims
- 10–15-point improvement in straight-through processing rates
- 40–50% of eligible claim volume under governed automation within 18–24 months
- Audit-ready governance that satisfies EU and US regulatory expectations without adding manual workload
A Narrowing Window for Competitive Advantage
The Ultimate AI Strategy for Insurance report makes clear that the gap between leaders and laggards is compounding rapidly. Natural catastrophe losses are projected to approach $145 billion in 2025. Premium growth is slowing toward 2% while loss severity rises. Insurtech entrants built on modern architecture are automating their way to lower expense ratios. Carriers that move to production-grade, workflow-integrated AI now will build advantages in cycle time, loss ratio, and customer satisfaction that, the data suggests, late movers will not be able to recover.
The report provides a six-phase execution framework covering strategic readiness assessment, AI partner selection, change leadership, first-win deployment, value measurement, and systematic scaling, with specific entry and exit criteria at each stage. It also addresses the regulatory landscape in detail, making the case that governed AI deployment is not a compliance burden but a speed advantage: carriers that build explainability and oversight in from the start reach production faster than those that treat compliance as a final validation step.
The complete Ultimate AI Strategy for Insurance report is available here.
Related News:
