For insurers, having access to generative AI is increasingly becoming the norm rather than a competitive advantage. As the same general-purpose models become available across the market, the focus is shifting towards how effectively insurers can apply AI to their own data, processes and regulatory requirements.
Research from Stanford has highlighted the productivity gains possible with generative AI, including for tasks such as drafting correspondence, summarising reports and creating internal tools. But Earnix argues that these efficiency improvements are unlikely to remain a differentiator as adoption becomes more widespread.
Instead, the InsurTech points to vertical AI designed specifically around individual industries. Such systems can be built to understand particular workflows, data structures and compliance obligations, with the aim of connecting decisions across an organisation rather than simply accelerating individual tasks.
Insurance already makes extensive use of AI. Models can help underwriters assess assets, location and regulatory factors and climate exposure, while claims teams use anomaly detection to identify potential fraud. Generative AI is also being used by service teams to interpret policyholder circumstances and produce personalised responses.
The problem is that these applications can operate independently of one another. Insurers often use a combination of internal and third-party AI systems, leaving information distributed between departments rather than shared across the business.
That fragmentation can create problems when functions including underwriting, pricing, claims and customer service are working from different systems and sources of information. The result can be inconsistent outputs despite the individual AI tools themselves working effectively.
Earnix believes this creates a case for insurance-native AI, with technology designed specifically around the industry’s operational, data and regulatory environment. The approach puts governance and explainability into the underlying system rather than treating them as additional requirements.
Regulation is particularly important for insurers, with requirements capable of varying between states and municipalities. Being able to explain how an AI-assisted decision was reached and demonstrate how it was governed can therefore be critical.
Systems that cannot provide an explanation for their decisions, maintain an audit trail or account for jurisdiction-specific privacy requirements could leave insurers exposed to legal and regulatory risks. For carriers adopting AI, this makes governance a consideration from the point of design rather than something to address after implementation.
Earnix has developed its AI Orchestration System (AIOS) around this model. The InsurTech says its 25 years of work with insurers including AXA, Generali, Mapfre and Matmut has informed the platform, which extends its decisioning capabilities across the insurance lifecycle.
AIOS is intended to support decision-making across risk evaluation, underwriting, claims and customer engagement, while providing greater visibility into how AI is being used and governed throughout the organisation.
Earnix argues that the next source of AI advantage for insurers will not simply be having access to the latest models. Instead, it will depend on how effectively those models are connected to proprietary data, existing processes and the regulatory obligations that shape the industry.
Copyright © 2026 FinTech Global
