The insurance sector was among the first to move toward digital transformation and is now at the center of the evolution toward AI-native models. The potential of AI is enormous in a sector driven by data, but some challenges can slow down its large-scale adoption.
In this article, we analyze the state of the art of AI in the insurance sector (AI insurance), the main application areas, and the concrete steps to take, in light of the specific characteristics of the sector.
Key Points
-
The insurance sector is among the most active in AI investments, but only a minority (around 7%) has reached a truly pervasive and transformative level of adoption.
-
Scalability is slowed down by structural constraints: strict regulation, data silos, legacy systems, and architectures not designed for AI.
-
To overcome these limits, it is necessary to focus on key pillars: a domain-driven approach, modular architectures, and strong investment in change management and adoption.
AI in the insurance sector: an opportunity to seize
With investments exceeding 10 billion dollars in 2025, the insurance sector is undoubtedly among the most active in AI adoption, with levels of diffusion that — according to BCG — are comparable to those of the TMT sector (Technology, Media & Telecommunications).
At the root of this acceleration is not only a technological push, but also a sort of native predisposition of the sector: insurance companies have vast information assets and long-standing experience in data-driven decision-making models, resulting in an approach and internal capabilities already oriented toward data.
This is combined with a tangible economic return. According to McKinsey, insurance companies that have strongly adopted AI (so-called leaders) have generated shareholder returns six times higher than followers. At an operational level, leading companies have recorded concrete improvements in claims management and premium growth (10–15%), as well as a significant reduction in customer acquisition costs (up to 40%).
From experimentation to large-scale adoption: the challenges for insurers
Despite this momentum, data shows that around two-thirds of companies are still in a pilot phase, and only a very limited share (around 7%, according to BCG) has managed to bring AI to a widespread and transformative integration. McKinsey refers to this condition as “pilot purgatory,” where many experiment but few succeed in integrating AI into business models.
The reasons for this gap span multiple dimensions.
-
Lack of an enterprise-level strategy
As in many other sectors, AI often grows through local initiatives, typically linked to individual functions and limited use cases. What is missing is a vision that connects individual applications to a common strategy capable of generating results across the entire value chain. -
Legacy infrastructure and technological complexity
Many players operate with core systems layered over time and difficult to integrate with modern AI and data analytics solutions. This slows down the entire evolution process. -
Data-related challenges
If there is one thing the insurance sector does not lack, it is data. However, quantity does not automatically mean usability. The effectiveness of AI depends on the ability to govern these assets: ensuring quality, security, and compliance, building shared and accessible data platforms, and overcoming historical fragmentation. -
Skills and change management
The introduction of AI is not only a technological issue, but also an organizational one. Many companies face a shortage of internal skills and resistance to change, especially when AI impacts established processes and operational roles.
AI use cases in the insurance sector: where value is concentrated
The insurance sector is broad and complex, structured around multiple segments and business lines. Each area primarily invests in use cases consistent with its operating model and strategic objectives. At the same time, it is clear that there is a common foundation at the process level, from which cross-functional and reusable applications emerge.
Technological components tend to converge: intelligent data extraction systems, predictive engines, GenAI solutions for document summarization, conversational tools for case opening and management, and risk scoring models. Their modular nature allows companies to reuse assets and expertise across different domains, accelerating AI scalability and avoiding the need to start from scratch for each new application.
That said, different insurance segments are focusing on specific applications.
-
Life Insurance
AI is transforming risk assessment and policy definition, making them more accurate and dynamic, also thanks to the integration of synthetic data that complements and enhances existing information bases. At the same time, AI enables the creation of increasingly personalized offerings aligned with the customer’s risk profile and needs.
-
Health Insurance
In the health segment, AI makes it possible to move beyond average statistical models, introducing a much more precise and dynamic assessment of health risk. This enables the creation of coverage aligned with the specific customer profile.
-
Property & casualty (P&C)
Here, AI can act along two main directions: on the one hand, building increasingly accurate risk models and simulating complex scenarios; on the other, automating claims management and strengthening anti-fraud systems in the retail segment, with direct impacts on operational efficiency and settlement times.
AI insurance: i pilastri di un percorso di successo
Analysts converge on one key point: the real success factor is not the technology itself, but how it is integrated into the business. Without a structured approach, the risk is what we already observe today: a proliferation of isolated initiatives, disconnected from each other and lacking an enterprise vision, which struggle to translate into large-scale adoption.
Summarizing market indications, several pillars of an effective approach emerge.
Domain-driven approach
It is essential to move as soon as possible from individual use cases to an end-to-end transformation of entire domains (claims, underwriting, risk, customer management…). This enables synergies across data, models, and processes, generating deep and measurable impacts.
Business–technology alignment
AI initiatives must be driven by clear and shared business objectives at the executive level, with concrete and measurable KPIs. Without this alignment, the risk is accumulating technological projects without real impact.
Modular architecture
At a technical level, large-scale AI adoption requires the development of reusable components such as data extraction engines, predictive models, and GenAI solutions, applicable across multiple domains. In addition, appropriate organizational structures are needed: cross-functional teams and the ability to rapidly develop and deploy solutions across different areas.
A solid data platform
Data must become a truly shared, governed, and real-time usable asset. This requires overcoming historical silos and modernizing data infrastructure, going beyond the limitations of legacy technologies and outdated data management models.
Change management
It is essential to rethink processes, roles, and operating methods, supporting people in adopting new solutions. In what is often mistakenly considered a purely technological journey, many initiatives fail precisely due to poor adoption, starting from top management roles.
Kirey’s support for a concrete transformation
At Kirey, we support companies across all sectors in their evolution toward truly data-driven models enhanced by artificial intelligence solutions. We approach this journey by working on data, architectures, and processes to build the foundations for a solid and scalable transformation.
This experience is combined with strong specialization in the finance market, which we have always supported in digitalization, cloud journey, and AI transformation initiatives. We have deep knowledge of business dynamics, regulatory constraints, and competitive challenges that characterize the sector, enabling us to design, implement, and manage effective AI-ready solutions integrated into core processes and oriented toward value.
Contact us to discover how we can turn AI into a concrete lever for evolution and competitive advantage.
