The insurance industry is haunted by a persistent narrative: unless every system is rewritten, replatformed and reimagined by next Tuesday, the enterprise architectural house of cards will collapse.

Some advocate halting all in-flight capital projects and immediately pivoting to all-AI, all the time, regardless of whether the tech is ready for practical use in complex transactions and can survive an audit. Too many of those who repeat this mantra have never run a technology team, implemented a core system, underwritten a policy or settled a claim.

An honest look at most insurers tells a more grounded story. Much of today’s technology is not obsolete—it is simply mature. And maturity, when managed correctly, is not necessarily a liability. Many carriers have successfully wrapped and reskinned their incumbent systems, giving the appearance and interoperability of “newishness” without gutting their core systems.

Frankly, the current fastest path to leverage the power of AI is not taking years to replace core systems with ground-up AI-native capabilities. It’s the far less glamorous approach of adding AI capabilities where available and upgrading in place.

What It Means to be Modern

Modernization is not synonymous with buying something new. It is about meeting functional, architectural, operational and economic expectations. Many insurers have already made enormous strides, and the idea that everything built more than 10 years ago must be ripped out and replaced is neither practical nor strategically sound.

We need a more nuanced benchmark for what insurance systems should be considered “modern enough.” Here are what I see as the five basic criteria. And yes, I am broad-brushing and simplifying here, but roll with me:

1. They must be integratable (via APIs and messaging buses) and securable, so they can safely exchange data cleanly with third‑party solutions, InsurTechs, analytics platforms and downstream processing engines.
2. They must deliver the core functional needs of the business without daily heroics or unwieldy manual compensating processes.
3. They must run on a supported technology stack with enterprise-class vendor support and a clear operational path forward.
4. They must operate at a reasonable cost relative to the value they deliver.
5. Knowledgeable human resources and documentation must be available. Luckily, AI can assist greatly when documentation is lacking.

When systems meet these criteria, they are not “legacy”—they are incumbent and stable. Replacing stable systems simply because they are not new opens the door to enormous waste.

The Cost of Stack Chasing

It’s a truism that replacing core systems is fraught with danger. According to research from both McKinsey and BCG, large‑scale core system transformations fail more than 70 percent of the time. Almost a quarter of the time, they become full write‑offs. These are astonishing statistics for any CIO or COO.

Part of the issue is cultural. “New” feels strategic, innovative and BOLD. “Enhancement” feels tactical, incremental and bland. But the functional outcome is what matters. A 12‑year‑old policy administration platform that is API‑enabled, stable, scalable and cost‑effective is materially more valuable than a brand‑new system that consumes capital and resources for years before stabilizing and delivering measurable differential business value. Over‑modernization is also an unspoken drain on capital, talent and opportunity. Every dollar/euro/yen spent replacing something that already works is capital NOT spent advancing underwriting capabilities, digitizing workflows, securing customer data or improving policyholder experience.

As a wise colleague recently told me: “The pace of change in technology is so fast that clock of obsolescence starts ticking the moment it hits production. Chasing tech stack is futile.”

How Much More Do You Need?

Lest we forget, the modern insurance environment has already benefited from nearly two decades of consistent technology investment, with trillions spent globally on insurance technology over the last two decades. The majority of carriers already operate on architectures built within the last 10-15 years.

Now, consider the tradeoffs between a $300 million system replacement versus a thoughtful modernization program. The latter might only achieve 80 percent of the benefit, but at 10-20% cost, with business value delivered years earlier. This isn’t a theoretical scenario. I’ve seen it play out repeatedly with insurers challenging the assumption that new is always better, embracing their inner engineer and taking on difficult technical work.

There is also a deeper strategic reason to embrace the concept of “modern enough.” Insurance systems do not operate in isolation. They sit within an ecosystem of distribution tools, workflow engines, underwriting workbenches, CRM platforms and, increasingly, AI‑supported processes. An insurer’s ability to adopt innovations—especially point solutions and InsurTech components—depends less on whether their core systems are “new” and more on whether the architecture is integratable, flexible and API‑rich.

If the system can integrate easily, support cloud‑based components, run securely, and adapt its data structures and workflows when business needs evolve, it can anchor a modern insurance operation—even when its internal codebase predates the smartphone.

Choose Your Journey

Modernization is not a destination but an ongoing set of architectural decisions. The goal is not “perfect modernity,” whatever that means, but to align technology capabilities with business value. When insurers define “modern enough” around flexibility, operability, functionality and cost efficiency, they free themselves from the tyranny of unnecessary replacement cycles.

If you’re a carrier, here’s my homework for you: Which of your systems meet the five basic criteria, and which do not? Where are the architectural bottlenecks that impair integration? Where is cost misaligned with value? And most importantly, where would modernization-in-place deliver the best risk‑adjusted return?

The insurance industry’s real competitive advantage won’t come from universal adoption of brand‑new platforms. It will come from thoughtful architectural stewardship, measured investment and the ruthless prioritization of business value over novelty.

Successful companies won’t be those who modernize the most but those who most intelligently modernize enough.

This article was previously published on LinkedIn by Robert Pick, the Group Deputy Chief Information Technology Officer at Tokio Marine Group and the Executive Vice President and Chief Information Officer at Tokio Marine North America. It is being republished here with his permission.