How Outdated Core Systems Are Quietly Draining Insurance IT Budget

There is a budget conversation happening inside most insurance companies that rarely makes it onto a strategy slide. It is not about underwriting losses or claims inflation. It is about the compounding cost of keeping old technology alive — and what every dollar spent on it fails to produce.

Research from PwC found that, on average, 70% of an insurer's annual IT budget is spent maintaining legacy systems. Earnix That figure deserves scrutiny. If a mid-size carrier spends $100 million annually on IT, $70 million is effectively a maintenance tax — paid not to build anything new, but to keep aging infrastructure from collapsing. The remaining 30% is what that organization has left to compete with insurtechs that built their entire stack from scratch.

Executives trying to understand the full scope of this problem — what it costs operationally, where the drag actually shows up, and which modernization paths reduce risk without halting operations — will find a detailed breakdown in nCube's analysis of insurance legacy system transformation strategies. What follows is the business and financial case that makes that analysis worth acting on.

This is the core problem with how the industry has framed the insurance IT modernization conversation. It treats legacy systems as a technology issue. They are, first and foremost, a capital allocation problem. Understanding where the money actually goes — and what it costs to leave those systems in place — is the starting point for any serious budget review.

The Budget Trap Nobody Audits

Legacy systems do not announce their costs in a single line item. They spread them across the organization in ways that are easy to rationalize individually and devastating in aggregate.

The most visible drain is direct maintenance: licensing fees for unsupported platforms, specialist contractors for languages like COBOL that few engineers under 50 have worked with, and the recurring cost of patches that buy time without solving anything. Research from Celent found that insurers often allocate between 70% and 80% of their IT budgets simply to maintain legacy systems — leaving a meager 20–30% for innovation, digital projects, and growth initiatives. Decerto

The less visible drain is operational friction. Nearly half of insurance executives surveyed said that launching even a minor product update takes between nine and sixteen weeks, and 41% reported lacking real-time data access. Insurance Business America Those delays are not free. 

Every week a product sits in a development queue is a week competitors can price against you, a week a distribution partner waits, and a week customers looking for a faster quote go elsewhere.

Then there is the compliance exposure. Deloitte research shows that technical debt costs the insurance industry over $30 billion annually in maintenance and lost productivity. Iankhan Regulatory requirements around data privacy, audit trails, and real-time reporting continue to tighten — and legacy architectures, built before these standards existed, require increasingly expensive manual workarounds to remain even nominally compliant.

What makes this pattern so persistent is that each cost looks manageable in isolation. The COBOL contractor is a small line in the IT budget. The nine-week product launch is just "how things work here." The compliance workaround gets absorbed by operations. None of it triggers a capital review. All of it compounds.

Where the Costs Actually Live

Policy Administration: The Core Engine Running on Fumes

Policy administration systems are typically the oldest and most deeply embedded platforms in any carrier's stack. They also touch every revenue-generating process: quoting, issuance, renewals, endorsements, and billing. When these systems are slow or rigid, nothing downstream moves cleanly.

Launching a new product on a legacy policy administration system can take 18 to 24 months and require multiple specialized programmers. By the time the product is ready, the market opportunity may have already passed. Decerto For carriers trying to respond to embedded insurance trends, usage-based models, or competitive pricing pressure, that timeline is not a minor inconvenience — it is a structural disadvantage baked into every product decision.

The financial consequences extend well beyond delayed launches. McKinsey's benchmarking data shows that companies with modernized, integrated IT achieve 41% lower IT costs per policy and 40% higher operational productivity than those running fragmented systems.

Genasys Technologies For a carrier with several million policies in force, the per-policy cost differential is a meaningful drag on profitability that does not appear in the legacy system's maintenance invoice but shows up clearly in expense ratios over time.

Data Silos: The Hidden Tax on Decision Quality

Legacy architectures store claims, policy, billing, and underwriting data in separate systems that do not communicate in real time. The cost of this fragmentation is not merely technical — it directly degrades the quality of decisions being made across the business.

Underwriting teams working from stale data price risk less accurately. Finance teams reconciling numbers across disconnected systems introduce errors that compound during audits.

Customer-facing staff without a unified policyholder view give inconsistent service. McKinsey documented one insurer running more than 300 active IT systems, with over 40% slated for decommissioning — and cost ratios twice the market average for its size. Genasys Technologies

For any insurer with ambitions to use AI in underwriting or claims, the data silo problem is not a future challenge to address during modernization. It is the reason modernization cannot be deferred. AI systems require clean, unified, accessible data. A model trained on fragmented, inconsistent inputs does not reduce loss ratios — it systematizes existing errors at scale.

Talent Costs: The Expense Sheet Nobody Sends to the CFO

The workforce implications of legacy systems rarely surface in IT budget reviews, but they represent a real and growing financial exposure. Engineers with deep experience in legacy insurance platforms are retiring faster than they are being replaced, and the premium for retaining them has risen accordingly.

Fewer than 5% of insurers are expected to realize direct, tangible benefits from AI in the near term, with legacy systems and a shortage of AI talent cited as the primary obstacles. Insurance Business America The connection is direct: organizations locked into maintaining old systems cannot easily attract engineers who want to build new ones.

The talent required to keep a COBOL-based policy system running and the talent required to build AI-assisted underwriting tools are almost entirely different populations — and competing for both simultaneously strains recruiting budgets in opposite directions.

What Modernization Actually Recovers

The business case for addressing insurance IT modernization is not primarily about technology. It is about reclaiming capital that is currently being consumed without producing a return.

According to McKinsey, modernizing legacy systems can reduce IT costs per policy by 41% and increase operational productivity by 40%. Astera A phased approach — targeting the highest-friction systems first, running new components alongside existing ones to reduce disruption risk — allows carriers to demonstrate those returns before committing to full-scale replacement.

The choices you have depend on what each system does and how deep it is integrated. Some platforms are best served by switching to cloud infrastructure with minimal edit, releasing budget from on-premise hardware, and simultaneously gaining time for the deeper work. Others need refactoring: change the code structure without changing the business logic, lower technical debt, and improve maintainability.

Only a small proportion usually the most fragile, highly regulated, or those that cause integration problems will be worth replacing altogether.

One thing that unites all these different ways is here: timing is everything. Starting off with the systems that constitute the biggest maintenance cost, are the main blockers for the most valuable integrations, or the ones that pose the clearest compliance risks will bring you quick successes that will not only bring internal confidence but also strengthen the business case for the next step.

The Cost of Staying Put

Framing insurance IT modernization as a capital expenditure decision — rather than an IT project — changes the risk calculus significantly.

The Earnix 2024 Industry Trends Report found that 74% of insurance companies still rely on outdated technology, while the Adacta 2025 State of Insurance Legacy System Modernization Survey found that nearly half of respondents cited system obsolescence as the primary reason they finally moved toward modernization.

A3Logics Note the sequencing: the majority moved when obsolescence forced their hand, not when the business case was optimal. Waiting until a system fails a compliance audit, loses vendor support, or creates a material incident before treating modernization as urgent is a common and consistently costly pattern.

Only 10% of large insurers have modernized more than half of their systems, despite the clear financial and operational benefits of doing so. UniRidge The gap between acknowledged problem and taken action is not primarily about budget availability or technical complexity. It is about how the cost of inaction gets framed internally. When legacy maintenance is treated as a fixed operational cost rather than a drag on capital efficiency, the urgency disappears from the analysis.

The more useful framing is this: every dollar held in legacy maintenance is a dollar not deployed toward customer experience, underwriting accuracy, distribution partnerships, or product speed.

The cost is real — it just does not arrive as a single invoice. It accumulates quietly, quarter by quarter, in lower margins, slower launches, and widening gaps between what the business can do and what the market is demanding.

That is not a technology problem. It is a profitability problem with a calculable answer