Digitalization and AI in Insurance

Insurance market-wide observations from the perspective of a professional software engineering and consulting firm.

The Insurance industry needs modernization. The current accelerated digitization makes it hard for laggards to keep up. There is a clear efficiency gap that needs to be bridged. The market's connectivity is low-tech, hugely inefficient, and the market lacks the will to adapt to modern technology standards - particularly among traditional insurers that have expanded organically or through acquisitions.

On the other hand, emerging digital-first insurance companies need to establish a clear technology strategy, consider the  “build vs buy”  approach to fast-tracking AI, and plan how to move from “old” to “new” operating models. Ultimately, they aim to enable the market and the brokers with modern technology-based solutions, seamless integration capabilities, and instant, algorithmic execution.

VirtusLab recognizes these different approaches to digitalization, taking into account that our customers present varying levels of maturity in terms of technology, data, and business process knowledge. To bridge this gap, we propose a structured set of technology patterns and services designed to make insurers more interactive and connected. This includes building strong data platforms with integration architectures and fostering a partnership-driven approach to build or buy models.

From our experience, Insurance companies generally progress through three key maturity stages in their transformation journey:

1. Building the solid data foundation: with the aim to consolidate and centralize internal IT systems, building them to be more scalable, connectable, and manageable in the market. This also positions the business for future innovation and ensures more enhanced data tech, and enables AI capability.
    
2. Improving insights and decision making: as an insurance company's capabilities mature, the data it utilizes and the technology it employs become more intuitive and useful. This evolution leads to better insights and automated processes. The advanced technologies that have been implemented facilitate smoother business operations, significantly enhancing the ability to gather, store, and analyze data.
    
3. Introducing predictive analytics and paving the way towards Agentic AI: to enable data-driven decision making and automate human labor tasks. As a result, it enables insurance professionals to spend less of their valuable time working with highly fragmented information and focus more on analyzing complex risk profiles.

This article explores the intricacies of navigating the insurance market from the perspective of an external service provider. While not intended as a comprehensive industry analysis, it shows the role of professional software engineering partners in helping insurers embrace modern technology to accelerate innovation. 

Insurers know their business but are often constrained by legacy systems and traditional thinking. This is why a partnership between insurers and technology providers is an interesting lever to consider.

Let’s delve into this.

Building a Data Foundation is now a prerequisite!

The insurance space is undergoing technological modernization. Due to changing market conditions and the rise of AI, things that were previously ignored, such as governance, automation, software development lifecycle (SDLC), and data management, are now at the forefront of change. The industry is prone to disruption with AI in many areas of the business, such as submission ingestion, claims management, risk & pricing models, legal, customer experience, and the list goes on.

Today, success depends on how quickly companies can make decisions, flip underlying solutions, and stay flexible – underlying technology platforms should support that. Both brokers and carriers still consider data as intellectual property (IP). Without secure storage, strong integration, and reliable data pipelines, none of the above will work.

The most notable challenges insurance companies face at the bottom level are:

• Dozens or hundreds of legacy IT systems that need to be consolidated create operational inefficiencies and present a high cost of ownership.
• Fragmented data across multiple external systems, internal databases, and local spreadsheets, which never make its way to more comprehensive data systems.
• Complex or non-existent integrations between existing systems and data, often constrained by inefficient data transformation and ingestion workflows.
• Poor data quality and unstructured data make it difficult to drive business insights.
• Ad-hoc and unstructured reporting capabilities are a big bottleneck and are slowing down business decisions.
• Difficulties in accessing the data across subsidiaries or during mergers and acquisitions (M&A), undermining due diligence and group-level transparency.

To have good AI, it’s essential to build a good foundation of information architecture, much like a sturdy house requires a strong foundation. A solid data foundation ensures that AI models learn from accurate, comprehensive information.

Iterative improvements to decision-making

As an insurance company's capabilities mature, the data it utilizes and the technology it employs become more intuitive and useful. This evolution leads to better insights from existing data, automated processes, and joined-up systems. The advanced technologies that have been implemented facilitate smoother business operations, significantly enhancing the ability to gather, store, and analyze data. 

The opportunity to democratize data across the business and interact more efficiently ensures rapid completion of insurance workflows and transactions. Moreover, it helps expand new business models and enables cross-sell opportunities. As a regulated market, there is a pressing need to report transactions and comply with regulations, particularly around premiums written, data, security, and so on. 

The modern solutions developed enable this process to occur seamlessly. More importantly, these advancements make the business more effective in its day-to-day tasks, such as underwriting decisions or resolving claims. This is also paving the way for further innovation in the AI space.

For instance, a company can transition from manual data re-keying tasks to more advanced self-service reporting capabilities concerning existing data, simplifying specific business intelligence needs, thus enhancing effectiveness. This is a critical stage to ensure that the business can remain competitive.

Based on our experience, the most notable challenges where software engineering and consulting partners offer the most support for insurers include:

• Combining widely spread systems to allow for the elimination of redundant or unnecessary systems, reducing complexity, and improving operational efficiency.
• Instilling an engineering culture and approach to building self-service systems and capabilities.
• Tackling manual tasks that are burdensome and tedious, and automating them.
• Making data accessible to the wider business through advanced technology and tooling.
• Moving from traditional release models to iterative models with speedy prototyping and solutions that are demo-ready at an early stage.
• Unlocking data assets for further experimentation with Data Science, ML, and GenAI.

The aim is to bridge the gap between skills, knowledge, and expertise effectively and enable companies to make the most of their data assets, being in control of them, and to have a scalable capability for the future, with the view to adopting AI as a key capability or function. 

We believe that positioning our clients for future innovation and making them self-sustainable in technology requires a two to three years of engagement lifecycle. Beyond that point, insurers can rebalance their outsourcing strategy.

AI adoption in Insurance

Historically, the high-value use cases, such as pricing optimization or risk management, mainly make use of developing “classic AI”. In other words, they use supervised Machine Learning capabilities, where models are being trained on structured, pre-labelled data. This still requires a conventional approach to data: from data gathering, through data transformation, to building underlying data platforms that enable further AI initiatives.

On the other hand, the commoditization of Generative AI tools (GenAI) lowered the entry barriers and opened the door for new use cases in insurance, ranging from extracting insights from unstructured data, providing more comprehensive risk summaries, to the interpretation of regulatory documents, and much more.  

Insurance professionals don’t need to become AI experts to bring the value it offers. They have far more important domain-related skills, and the entry barrier to using AI is falling each day.

When it comes to AI adoption, not only in the insurance domain, we usually start questioning the fundamentals: 

• Can AI help meet your objectives and challenges?
• Are you already digital-first, or are you currently on a modernization journey?
• What AI are you already using in the organization? 

This happens even before we move forward with identifying AI opportunities or assessing current challenges with their data and infrastructure. Smaller, specialist insurance companies need a stronger business case for AI as there is less room for experimentation; instead, they need to start realizing the value of AI faster. In contrast, larger companies tend to have more funding available as part of their ongoing innovation programs, where they can afford to run multiple AI-pilot projects.

Once you have answered the questions above and positioned your starting point, here is a list of key considerations to start your AI adoption:

• Practical use cases often start with improving data quality and reducing manual re-keying. Large language models (LLMs) can process unstructured documents, detect inconsistencies, and spot correlations across datasets that were previously hidden. This happens before companies start investing more in embedding AI into the decision-making process or leveraging AI agents for workflow automation, which is usually year two to three on the roadmap.
• Over time, start taking friction out of the system by leveraging AI - for example, in case underwriting takes too long, ingestion of the unstructured and messy data can be automated using LLMs.
• Insurance businesses tend to adopt more on-point solutions, for example, for submission handling. In contrast, end-to-end platforms that fundamentally disrupt existing operations or introduce entirely new practices face greater resistance and are more difficult and more costly to adopt.
• Insurance companies need to revisit their governance and data management frameworks to take into account the unique characteristics of AI systems. For example,  when introducing evaluation frameworks (to prevent inaccurate results) and managing relationships with third-party service providers, company data has to be protected from being exposed or used for training.
• There is still a strong need to educate insurance leadership on AI to demystify it, and go through the AI adoption process with them together. Especially around data privacy and security aspects.
• Finally, AI-first companies need to think about how to industrialize their digital business, especially in an industry that is not tech. The other challenge is to enable brokers to transact via modern submission portals and follow an API-first approach.

Market consolidation & Private Equity continue eating the world

The past two to three years have been difficult for private equity (PE) globally, with fundraising remaining tough. However, as economic conditions show signs of improvement, it leads to increasing consolidation in the market, especially in the Insurance ecosystem. These observations come from direct conversations with insurance leaders and from VirtusLab’s own long-standing relationship with a PE firm.

Insurance is an attractive sector for PE-driven consolidation. Its dependence on economies of scale makes it particularly well-suited to consolidation strategies: larger players benefit from operational synergies, greater negotiating power with reinsurers, and broader distribution networks with cross-sell enabled. Importantly, consolidation also unlocks richer datasets, which in turn enable more accurate underwriting, pricing, and improved portfolio management. As a consequence, there is pressure to “increase” and “consolidate” for insurers.

From a technology perspective, several key observations stand out:

• Economies of scale through consolidation and standardization of technology - larger, consolidated entities can reduce duplication, centralize infrastructure, and decommission legacy IT systems. Unified and properly governed technology platforms not only cut costs but also create stronger data foundations. We usually see a more top-down approach in terms of technology adoption and architectural choices.
• Streamlined M&A integration depends on standardized playbooks for integrating technology, data, operations, and culture. For traditional insurers, there is often a need to replatform as tech is outdated. Having a blueprint to onboard companies quickly in a consistent and governed fashion is becoming crucial.
• Buyers seek platform efficiencies and AI readiness as a core part of due diligence. Modern, integrated platforms accelerate post-deal value creation and open the door to AI-enabled efficiencies.
• Sellers gain higher valuations through embracing technology and attract more interest from PE buyers. The key deciding factors include best class data privacy and data management, non-monolithic structure (microservices), and seamless data migration.
• More money is spent on technology, where the main use case is price optimization, as people buy based on price. Also, companies start leveraging technology more in the claims journey. There is also an emerging interest in Agentic AI use cases to automate burdensome tasks such as manual data entry and let insurance experts focus on more meaningful work.
• Moving operations offshore has become a common lever to reduce expenses (OpEx).

Taken together, it creates a need for streamlining the process of merger and acquisition (M&A) to accelerate post-acquisition integration and shorten the overall investment cycle. In most cases, this will involve going through the “messy middle”: multiple legacy systems, fragmented data, and overlapping operations that must be somehow integrated. 

Beyond PE’s core drivers of maximizing profit and expanding market share, ultimately, insurance is something everyone needs. When the technology part is done well, it helps to deliver meaningful social value by offering customers a sense of security - the assurance that if something goes wrong, they won’t face it alone.

Opportunity cost of inefficiency

The ultimate metric in insurance is loss ratio, which relates to how well insurers manage their risk portfolios. The more comprehensive the data assets and the more streamlined the technology, the greater the efficiency insurers achieve. 

Among insurers, Managing General Agents (MGAs) stand out as prime candidates for efficiency improvements because they sit at the crossroads of providing specialized expertise & distribution and often operate with leaner infrastructures and legacy systems. This combination makes them both prone to inefficiencies but attractive for investors looking to modernize and scale.

Insurance remains a "grey hair" business, according to NAMIC data, with nearly half of its workforce expected to retire by 2028. So there is a need to attract a new wave of talent to the industry that was historically lagging behind in innovation. The next generation of workers thinks differently: they expect better tools, modern UX, and a digital-first approach that aligns with their working style. The traditional manual re-keying of data and fighting with legacy, fragmented systems is no longer an option to remain sustainable.

Insurance is deemed as a necessity buy; however, that dynamics change due to how the world operates today. From the insurer's perspective, they need to think more about the ecosystem and buyer persona - it's about the market they penetrate and how they operate within this market, making the customer viewpoint insurance quite important. The priority is no longer just offering coverage, but enabling effective distribution while considering next steps for both horizontal and vertical growth. Missing these opportunities because of friction points caused by inefficient technology solutions can be costly.

Beyond the operational efficiency and attracting a new wave of insurance talent, there is a significant opportunity cost in the valuation of the business itself. Insurers that are technology underserved tend to receive lower valuations. Private equity buyers and strategic investors increasingly evaluate technology maturity, including:

• Best-in-class data privacy and data management
• Non-monolithic structures (microservices) that can scale and integrate efficiently
• Seamless data migration to consolidate systems more easily

Technology maturity is no longer a “nice to have”, it’s a significant leverage that shapes talent attraction, customer experience, distribution, and ultimately, the enterprise value.

How to navigate digitization in Insurance?

Successfully navigating digitization in insurance requires more than simply chasing the latest technology advancements. The key is not to overshoot technical maturity, but to start where the company’s culture is today. Transformation often involves going through a messy middle, especially in relation to the PE-owned environment - they buy promising but not technologically mature companies to expand their geographies, portfolio, or data assets. Companies that move with a clear blueprint of what “good” looks like are far more likely to succeed.

Insurers need to understand where they differentiate and build an ecosystem and internal capabilities to leverage that point of differentiation. Buy off-the-shelf solutions where capabilities are commoditized and offer no additional value.

There is still a need for talent to build, and also talent to use modern technology. The technology is at the heart of it, but it’s mostly about processes and how people use it. 

On the AI front, success starts with a strong data foundation and a solid business case for AI. Many insurers have experimented with pilots; the challenge now is to operationalize AI at scale, embedding it into underwriting, claims, and customer journeys.

It’s no longer tech for tech, but tech for business.