AI-powered solutions in the insurance industry with Managed AI: Why the insurance industry is facing its biggest turning point.

The end of "digital paralysis": How managed AI is saving the insurance industry

What functioned for decades as a stable business model based on risk aggregation and incremental innovation is now facing a perfect storm of technological debt, exploding costs, and regulatory pressure. The numbers speak for themselves: While insurance fraud claims around $122 billion annually worldwide, paradoxically, up to 90 percent of IT investments by traditional companies are spent solely on maintaining outdated systems – a “digital paralysis” that stifles innovation.

But the price of stagnation is no longer just a matter of lost efficiency; it's becoming an existential threat. In an era where fraud patterns are becoming more dynamic and customers expect real-time experiences, simply managing policies is no longer enough. The industry's answer lies in the strategic implementation of managed AI solutions. These technologies are no longer an optional gimmick, but rather the crucial lever for overcoming the gigantic "legacy trap," radically accelerating processes like claims handling, and assessing risks more accurately than ever before.

The following analysis examines the economics of this transformation in detail. From the historical reasons for the IT monoliths at industry giants like Allianz to the pitfalls of the new EU AI Act: We investigate how insurers are managing the balancing act between strict regulation and necessary automation. Learn why managed AI is more than just software – it's the infrastructure for tomorrow's competitiveness – and which strategies will determine the winners and losers of the coming decade.

Suitable for:

• UNFRAME.AI: Managed AI Solutions for Insurance

How insurers are automating away their future or shaping it cleverly

The insurance industry is at a critical turning point where technological, economic, and regulatory forces are converging and forcing fundamental change. The insurance business model, built over decades on manual processes, decentralized data structures, and incremental innovations, is coming under increasing pressure. The reality is unequivocal: the insurance sector currently loses approximately $122 billion annually to property and casualty fraud, with Germany alone facing losses exceeding €6 billion per year. At the same time, 70 percent of insurance companies' IT budgets are spent maintaining outdated systems that are increasingly collapsing under their own complexity. Two-thirds of insurance providers worldwide have so far failed to scale artificial intelligence beyond pilot projects and integrate it into their day-to-day operations.

This situation describes not simply an efficiency gap, but a survival problem. Managed AI solutions for the insurance industry are therefore not a technological gimmick or an optional modernization, but a strategic necessity that determines the competitiveness, profitability, and ultimately the long-term market viability of insurance companies. This report analyzes the economic drivers, institutional players, and market mechanisms behind this transformation process. It highlights how managed AI systems, as integrated solution platforms, enable insurers to overcome legacy systems, detect and prevent fraud in real time, accelerate claims processes, and scale personalized customer experiences.

Suitable for:

• Unframe AI transforms AI integration for enterprises in record time: Tailor-made solutions in hours or days

From electromechanical data processing to digital paralysis

To understand the current situation in the insurance industry, it is necessary to look at its technological development. Allianz, for example, was the first insurance company in Europe to introduce the IBM 650 mainframe computer in 1956. This was a breakthrough that revolutionized data processing and enabled insurers to efficiently manage millions of policies. In the following decades, these systems were continuously expanded and adapted to meet new requirements. Each new function was not integrated, but rather layered: insurance administration, claims processing, billing, and customer management emerged as systems that were partly isolated and partly intertwined.

This was historically understandable and economically sound. Until the end of the 20th century, such monolithic systems were the standard business model not only in insurance but in virtually all major financial institutions. At the time, they enabled scalability and profitability. However, these systems were not primarily designed for flexibility, rapid iterations, or external integration. They were optimized for stable, predictable processes.

The turn of the millennium and the following two decades then revealed the downsides of these decisions. As financial services worldwide came under pressure due to mergers, new regulations, and the rise of InsurTechs, insurers became increasingly reliant on systems they themselves no longer fully understood. In some cases, the technical dependencies are now so complex that no one in an insurance company has a complete overview of its own software architecture. Some changes that would seem trivial, such as adding a second email address to the system, incur costs in the six-figure euro range because they require adjustments in hundreds of places within the system.

Investments in IT illustrate the scale of the problem. German insurers alone increased their IT investments to a record €6.2 billion in 2024, primarily to address existing problems rather than to invest in future innovation. A large portion of these funds, estimated at 70 to 90 percent, is spent simply maintaining legacy systems. This represents a classic example of economic inefficiency: companies are paying ever-increasing sums to maintain the same level of functionality while their competitiveness declines. Technical debt is growing exponentially, while investments in innovation and growth are stifled.

Analysis of the key factors: Systemic inefficiencies and the incentive structures of the transformation

The insurance business is based on asymmetric information, risk aggregation, and premium logic. Insurers collect data on risks, assess these risks, and calculate premiums based on this assessment. For this assessment, they combine historical data, external information, and established calculation models. Traditionally, these were manual or semi-automated processes. An underwriter needed years of experience to perform these assessments consistently. Claims handling was similar: A claims adjuster had to read documents, compare facts with the policy, identify potential indicators of fraud, and then make a decision.

In this context, managed AI solutions act like a catalytic transformer. They enable these cognitive tasks to be handled not only faster, but also more precisely and in a more scalable way. But the economic leverage goes much deeper:

First, fraud prevention is paramount. Globally quantified losses due to insurance fraud in property and casualty (P&C) insurance amount to approximately $122 billion annually. In Germany, the estimate is over €6 billion per year, and this figure is constantly rising. Conventional fraud detection relies on rule sets: Suspicious patterns are defined by experts and then hardcoded into systems. The problem is that fraudsters adapt to known patterns, develop new techniques, and become more creative. Machine learning-based fraud detection works differently: It recognizes anomalous patterns that have never been described by humans before. McKinsey analyses show that state-of-the-art fraud detection increases the detection rate by 15 to 20 percent, while simultaneously reducing false positives by 20 to 50 percent. This has immediate economic consequences: Less fraud means fewer claims payouts. Fewer false positives mean fewer unnecessary investigations and faster verification for honest policyholders.

Secondly, there has been a massive increase in efficiency in claims processing. A major Dutch insurer that implemented AI-based claims processing achieved automation of 91 percent of all eligible motor claims. The average processing time per claim decreased by 46 percent, and customer satisfaction (measured as the Net Promoter Score) increased by 9 points. A Nordic insurer that introduced document intelligence solutions achieved correct data extraction and interpretation for 70 percent of incoming documents in real time, instead of manually and with delays. This was not only technically significant but also economically transformative: Claims adjusters were able to free themselves from routine tasks and instead focus on complex, high-value cases where human expertise truly adds value.

Third, dynamic risk assessment through AI enables a radical improvement in pricing accuracy. While traditional underwriting was based on a few variables (age, driving history, postcode), AI models can analyze and combine hundreds or thousands of data points in real time. This allows for more precise premiums that reflect actual risk, rather than average premiums that subsidize a large portion of the customer base. An Allianz case study demonstrates how the AI ​​system BRIAN (Underwriter Guidance Tool) uses data integration and semantic analysis to deliver risk-based recommendations that inform underwriters faster and more effectively.

Fourth, AI-driven personalization massively improves customer acquisition and retention. Generative AI and large language models make it possible to communicate with insurance customers in natural language, offer automated self-service solutions, and provide individualized product recommendations. A customer advisor who typically handles 100 inquiries per workday can double or triple this capacity with AI assistants, while simultaneously increasing the quality of the advice.

However, these levers only work under specific institutional conditions. Most insurers have been unable to realize these effects because their legacy systems are not flexible enough to support rapid integrations. An AI project at a traditional insurer can take years because every new interface creates hundreds of existing dependencies. This is the key reason why two-thirds of insurers worldwide have not yet scaled AI beyond pilot projects.

The current situation: Data-driven inventory and challenges

The global AI market for insurance is growing at an exceptional rate. In 2024, the AI ​​market in insurance was valued at approximately $6.44 billion to $11.33 billion, depending on the source. Forecasts for the coming decade are dramatic: the market is projected to grow to between $45.74 billion and $246 billion by 2031-2035, with an average annual growth rate of between 32 and 33 percent.

These figures are not mathematical abstractions, but rather expressions of real economic transformations. Insurers worldwide are investing massive sums in AI technology, talent acquisition, and transformation projects. The largest insurers, such as Allianz, Munich Re, and Zurich, have established investment units, labs, and research partnerships. Zurich announced the opening of a new AI lab in 2025 in collaboration with the University of St. Gallen and ETH Zurich to transform the insurance business model itself. Allianz is building an enterprise-wide data platform to integrate data from all sources and thus enable AI applications.

But these investments are not unlimited. Resources are typically tied up in legacy systems. German insurers spend roughly €5.9 to €6.2 billion annually on IT, but 70 to 90 percent of that goes toward maintaining existing infrastructure. This means that only 10 to 30 percent of this sum is available for genuine innovation and future investments. Small and medium-sized insurers have even fewer resources. Therefore, the rapid, integrated delivery of AI solutions from a single source is a massive advantage.

The most pressing challenges are as follows. First, the technical complexity of integration: Every insurance company has a unique landscape of legacy systems, each with its own APIs, data structures, and business logic. A true solution provider must offer not only AI algorithms but also a configurable connector framework that adapts to this diversity. Second, the regulatory complexity: With the EU AI Act, which came into force in August 2024 and will be fully applicable from May 2026, high-risk AI systems, especially those for risk assessment and pricing, are subject to strict requirements regarding governance, documentation, transparency, and bias minimization. Third, the issue of data protection and trust: Sensitive customer data, health information, and financial details must be handled with the highest level of security. Insurers cannot simply outsource this data to external cloud providers without incurring regulatory risks. They need solutions that run on-premises or in controlled environments and offer audit trails and full transparency.

Managed AI Platform - Image: Xpert.Digital

Here you will learn how your company can implement customized AI solutions quickly, securely, and without high entry barriers.

A Managed AI Platform is your all-round, worry-free package for artificial intelligence. Instead of dealing with complex technology, expensive infrastructure, and lengthy development processes, you receive a turnkey solution tailored to your needs from a specialized partner – often within a few days.

The key benefits at a glance:

⚡ Fast implementation: From idea to operational application in days, not months. We deliver practical solutions that create immediate value.

🔒 Maximum data security: Your sensitive data remains with you. We guarantee secure and compliant processing without sharing data with third parties.

💸 No financial risk: You only pay for results. High upfront investments in hardware, software, or personnel are completely eliminated.

🎯 Focus on your core business: Concentrate on what you do best. We handle the entire technical implementation, operation, and maintenance of your AI solution.

📈 Future-proof & Scalable: Your AI grows with you. We ensure ongoing optimization and scalability, and flexibly adapt the models to new requirements.

More about it here:

• Managed AI Platform

Case studies from practice: Comparative analysis of different transformation strategies

To illustrate the practical implications of this analysis, case studies with different approaches are helpful.

The Nordic insurer, which implemented AI-based document intelligence, illustrates the path of its phased, process-specific deployment. The company had decades of experience with paper and digital documents in claims processing. The manual process was highly complex: A claim would come in, external documentation would be photographed or scanned, employees would read it manually, and copy the data into various systems. Error rates were significant. With EY Fabric Document Intelligence, this workflow was transformed. Seventy percent of documents are now correctly interpreted in real time, and data is automatically extracted and transferred to the backend systems. This solution wasn't a completely new development, but rather an integrated component built on top of existing claims management systems. The result: Significantly faster claims processing, reduced errors, and employees who could focus on more analytical, customer-oriented tasks.

A major Dutch insurer is demonstrating an even more radical approach: the complete automation of traditional claims decisions. This company had a very clear hypothesis: approximately 91 percent of all motor claims follow standardized decision logics and could be fully automated if a system learned this logic. The insurer trained an AI agent that modeled the decision patterns of experienced claims adjusters. The agent was designed to classify claims, review claim conditions, and automatically approve simple cases. This implementation was technically challenging because it required connecting dozens of legacy systems. But the business case was so compelling that the investment was justified. After full implementation, the average claims processing time decreased by 46 percent, 91 percent of all eligible motor claims were automated, and customer satisfaction increased by 9 NPS points. However, this wasn't a complete automation of human labor, but rather a smart division of labor: the agent handled the routine tasks, while humans took on the complexities.

Allianz, as a global player, is pursuing a company-wide data integration and AI strategy approach. The company recognized that AI projects are only sustainably successful if the data quality is high and the data is accessible organization-wide. Therefore, Allianz is investing heavily in its Allianz Data Platform, data governance, and Chief Data Officer positions within its individual operating units. This is a long-term transformation path, but it addresses the core problem: Good AI needs good data, and good data needs organizational structure and culture.

In contrast, Zurich is pursuing a research and innovation approach through its new AI Lab. Zurich recognized that simply applying existing AI technologies is insufficient to achieve genuine business model transformation. The company has partnered with leading universities to develop new AI technologies and methods. The Lab focuses on agent-based AI systems that operate more autonomously and can make complex decisions. This is a game for the future, not about realizing efficiency gains today.

The comparison reveals several key insights. First, there is no single silver bullet approach. Process-specific AI solutions (such as Document Intelligence), full process automation (like the Dutch insurer), enterprise-wide data strategies (Allianz), and fundamental research (Zurich) are all valid and address different economic challenges. Second, speed of implementation is a critical competitive factor. A system that can be implemented in months, not years, offers economic advantages. Third, integration with existing systems is crucial. Insurers who pursue AI as an isolated project have limited success. Those who integrate AI into their existing technology landscape scale more effectively.

Suitable for:

• Managed AI Enterprise Solutions with Blueprint Approach: The Paradigm Shift in Industrial AI Integration

Future development paths and potential disruptions

Based on the analysis conducted so far, several likely development paths can be outlined.

The most likely scenario is a progressive fragmentation of the insurance industry. Large players with resources, such as Allianz, Munich Re, and Zurich, will successfully scale AI and data integration, thereby consolidating their competitive advantages. They will also remain innovative under regulatory oversight because they have the resources for compliance. Medium-sized and smaller insurers will face a dilemma: either they invest heavily in AI and modernization (which will reduce their profitability in the short term), or they fall behind technologically and lose market share. Many will opt for outsourcing or strategic partnerships with AI platforms (such as managed AI solutions providers). This could lead to consolidation, with the largest insurers attracting the best AI talent, while smaller insurers turn to distributors or pursue niche strategies.

A second likely scenario is the emergence of new insurance models fundamentally built on AI and data analytics. New InsurTechs, or tech companies entering the insurance sector (like Google in insurance), have less technical debt and can integrate AI into their architecture from the ground up. They could quickly dominate niche vertical markets. This puts pressure on established insurers to not only optimize their existing processes but also rethink their business models. Zurich has recognized this and is investing in research into new business models.

A third scenario is the progressive regulation and formalization of AI standards. The current EU AI Act is just the beginning. Further regulations are expected to follow, whether regarding explainability, bias minimization, or the creditworthiness of AI systems. This could lead to a situation where only specialized, highly regulated AI solution providers with genuine security and compliance certifications will be successful. Generic AI tools from tech giants could become inadequate for regulated industries such as insurance.

A fourth scenario, less likely but not impossible, is a backlash against AI automation in insurance, driven by public debate about job losses or discrimination. This could lead to political pressure to limit or ban AI in certain contexts. However, this is unlikely, as the economic benefits are too great.

Potential disruptions that could upend these pathways:

1. Massive data breach at a major insurer fundamentally damages trust in AI systems
2. Demonstrated discriminatory effects of AI systems in high-risk cases (such as a case like the Amazon hiring scandal, but in insurance), triggering a regulatory backlash.
3. Breakthrough in agent-based AI or autonomous AI decision systems that will further transform insurance models
4. Combined effects of climate change and improved risk assessment through AI, leading to massive market distortions (for example, when AI recognizes that certain regions are much riskier than previously assumed)

Strategic implications: The need for a coordinated transformation

Empirical analysis paints a clear picture: Managed AI solutions are not optional for insurers, but essential. Current inefficiencies are so drastic, competitive forces so strong, and regulatory requirements so clear that inaction is tantamount to giving competitors a competitive advantage.

For policymakers, this means that the regulatory framework (EU AI Act, GDPR, national insurance laws) must be strengthened, but also combined with practical support for smaller insurers. Without support, a two-tier insurance industry could emerge, in which large insurers remain innovative and force smaller insurers to either acquire or exit the market.

For insurance executives, the strategic implications are clear. Piloting individual AI projects is not enough. Insurers must:

1. Develop a company-wide data strategy that treats data as a critical asset.
2. Progressively dismantle or modernize legacy systems, don't endlessly install patches.
3. AI should not be viewed as an isolated project, but as an integral component of the operational architecture.
4. Integrate governance and compliance into project implementation from the outset, not as an afterthought.
5. Making strategic decisions about Make vs. Buy vs. Partner: When does it make sense to develop your own AI solutions, and when is a Managed AI Solutions provider the right choice?

For investors and stakeholders, the fundamental insight is that insurers who successfully navigate this transformation can expect competitive advantages, higher profitability (through fraud reduction, cost efficiency, and improved pricing accuracy), and stronger customer relationships. Insurers who fail to do so will lose relevance in an increasingly competitive and regulatory landscape.

The central thesis of this analysis is therefore not that AI is a technological option, but that AI is a strategic necessity that will determine the viability of insurance companies in the coming decade. Managed AI solutions, properly configured and anchored in governance, are the economic instrument to transform this necessity into reality.

Konrad Wolfenstein

I would be happy to serve as your personal advisor.

contact me under Wolfenstein ∂ Xpert.digital

call me under +49 89 674 804 (Munich)

LinkedIn
 

Our global industry and business expertise in business development, sales and marketing - Image: Xpert.Digital

Industry focus: B2B, digitalization (from AI to XR), mechanical engineering, logistics, renewable energies and industry

More about it here:

• Xpert Business Hub

A topic hub with insights and expertise:

• Knowledge platform on the global and regional economy, innovation and industry-specific trends
• Collection of analyses, impulses and background information from our focus areas
• A place for expertise and information on current developments in business and technology
• Topic hub for companies that want to learn about markets, digitalization and industry innovations