Unlocking scalable AI in insurance from the core 

AI is already reshaping how insurers underwrite, assess risk, manage claims, and engage customers. But most firms are not yet realizing its full value, not because of a lack of vision or investment, but because their systems were not built for it. 

Outdated core systems, some dating back decades, were not built for AI integration. In a market defined by speed, intelligence, and precision, legacy infrastructure is a costly bottleneck. 

Real transformation requires rebuilding the foundation  

To unlock AI’s full potential, insurers must modernize these systems through strategic, modular changes that build toward long-term transformation.  

This demands more than new tools: it requires purpose-built infrastructure for continuous AI integration. However, many insurers attempt to layer AI capabilities onto outdated architectures, creating more problems than solutions.  

With 71% of insurers describing their current underwriting systems as “complex,” integrating modern AI tools with rigid legacy infrastructure generates friction at every stage, from deployment delays to compliance risks and integration bottlenecks. 

The specific challenges legacy systems create for AI implementation include: 

• Compromised data quality: Legacy platforms often rely on inconsistent, outdated, or unstructured data, which undermines AI models that depend on real-time, high-quality inputs for critical functions such as fraud detection and personalized pricing. 

• Scalability constraints: Older systems lack the computational power to handle the massive data volumes required for advanced AI applications, including image recognition and natural language processing, at scale. 

• Siloed architecture: Decades of department-focused IT development have created fragmented systems with poor data governance, limiting enterprise-wide data visibility that effective AI requires. 

Without the proper technical foundation, insurers will miss out on more than just automation. They’ll lose competitive ground on speed, insight, and innovation to more agile competitors.  

Laying the technical foundation for scalable AI in insurance

While a complete system overhaul or total re-platforming may be out of reach for many insurers, a modular approach allows insurers to embark on incremental modernization. This strategy enables seamless interoperability between existing systems and emerging AI capabilities while aligning with evolving business priorities. 

From SAP Fioneer’s perspective, an AI-ready infrastructure includes: 

• Unified data platforms: Centralized access to standardized, structured data, from policy documents to call center recordings, enables effective AI model training and real-time decision-making across all business functions. 

• Cloud-ready scalability: Dynamic compute power adjustment based on AI workload demands ensures optimal performance during peak processing periods while controlling costs during periods of lighter usage. 

• Robust integration architecture: Seamless connectivity between legacy systems, new AI applications, and third-party services accelerates implementation while protecting existing investments.  

• Built-in governance and compliance: Automated audit trails, role-based access controls, and encryption ensure responsible AI deployment while supporting regulatory transparency requirements.  

• Low/no-code tools: Democratized tools enable business users in claims, underwriting, and customer service to build and customize AI-driven workflows without technical expertise. 

As AI capabilities mature, the next wave of AI in insurance goes beyond optimizing processes; it will autonomously execute them. This shift to agentic AI could have profound implications on core insurance functions in the coming years: 

• Intelligent underwriting, AI agents could autonomously review submissions, flag anomalies, and recommend real-time pricing adjustments, accelerating decision-making and improving accuracy. 

• End-to-end claims automation: Beyond current process improvements, AI agents could manage the entire claims journey, from digital claim intake to fraud detection and automated payments, thereby minimizing human intervention while enhancing customer experience.  

• Proactive policy management: AI agents could continuously review and update policies, identify coverage gaps, and automatically initiate amendments or customer notifications. 

• Autonomous financial operations: AI agents could automate payment processing, monitor account risk, and manage disbursements, ensuring efficient cash flow and compliance at scale. 

SAP Fioneer’s insurance platform is built for this evolution and natively integrates with SAP’s Business AI ecosystem, including Business Technology Platform (BTP), Joule, and Business Data Cloud (BDC).  

The platform’s modular and scalable design supports rapid deployment across both SAP and non-SAP environments, while leveraging proven security and compliance frameworks. Embedded AI capabilities automate business processes and deliver contextual insights, enabling organizations to innovate rapidly while maximizing operational value. 

The mandate to modernize 

Insurance is fundamentally about managing risk. Currently, the most significant risk is falling behind technologically as the industry around you undergoes transformation. 

AI’s full potential should not be held back by legacy infrastructure, siloed data, and outdated processes. Integrating AI is not a one-time upgrade; it’s an ongoing evolution that requires continuous investments in technology and people. Modernizing core insurance systems is a prerequisite for shifting from traditional reactive operations to agile, innovative, and customer-centric business models.  

However, technology alone doesn’t guarantee success. Sustainable transformation requires cultural change, employee upskilling, organization-wide development of AI literacy, and technical modernization. 

Insurers who approach modernization as a business growth enabler, rather than a back-office upgrade, will lead the industry in innovation, customer trust, and operational efficiency.