Insurance companies are investing heavily in artificial intelligence to improve risk assessment, pricing, underwriting, claims and customer engagement. Yet having increasingly sophisticated models does not necessarily mean insurers can act on their insights quickly, consistently or at scale.
Earnix argues the industry is facing an AI execution gap, with insurers struggling to connect models and data to the business rules, workflows and human decisions needed to turn analysis into action. An actuary may refine a model, a data team may generate a score and an underwriter may receive a recommendation, but siloed systems, manual approvals, reliance on IT and fragmented business rules can still stand between intelligence and the final decision.
The challenge is becoming more pressing as insurers contend with changing risk conditions, rising claims costs and greater pressure on margins and customer experience. In markets such as France, insurers are also navigating increasingly interconnected risks while operating across legacy technology environments, multiple distribution channels and complex approval processes.
Insurers already have a growing number of AI use cases across the business. Models can identify emerging claims trends, support risk selection, improve pricing and flag customers at risk of leaving. Generative AI can help summarise information and produce recommendations, while agentic AI can coordinate multi-step workflows.
The problem is that these capabilities often remain isolated. Pricing, underwriting, claims and customer engagement systems may operate independently despite the decisions they support being closely linked. A change in risk can affect underwriting appetite and pricing, while a pricing or underwriting decision can influence what is ultimately offered to a customer.
When these systems are disconnected, valuable insights can become trapped between development and production. Employees may then have to manually interpret the output of one system before feeding it into another, slowing down the decision-making process and increasing the potential for inconsistency.
Earnix’s AIOS is designed to address this gap by connecting existing systems, data and models with business rules, workflows, human approvals and operational actions. The AI orchestration system can use predictive, generative or agentic AI depending on the decision being made, while keeping governance and human oversight within the process.
The company identifies several requirements for making AI an operational capability, including reducing the time between a signal being identified and action being taken, adapting decisions without rebuilding core systems, applying automation at an appropriate level and embedding governance into the decision-making process.
This is particularly important for insurers operating in highly regulated markets, where AI-generated decisions need to be explainable, auditable and subject to appropriate oversight. AIOS is designed to work across existing technology environments rather than requiring insurers to replace their core infrastructure.
The wider shift is moving the conversation around AI in insurance away from the sophistication of individual models and towards what happens after a model produces an insight. For insurers, the competitive advantage may increasingly come from how quickly they can connect that intelligence to pricing, underwriting, claims and customer decisions.
With Earnix AIOS positioned as an orchestration layer between intelligence and execution, the company is betting that the next stage of insurance AI will be defined less by better models and more by the ability to turn those models into governed, measurable action.
Read the full Earnix analysis here
Read the daily FinTech news for the intelligence and early insight that keeps industry leaders ahead of what’s next.
Copyright © 2026 FinTech Global
