In Insurance IT Do We Build or Buy? The Answer Is Much More Nuanced

While A.I. can generate code in seconds, it can’t replicate decades of insurance experience. Additionally, even the best A.I. requires oversight and review.
By: | November 16, 2026
Topics: Risk Insider

A.I. has fundamentally changed how smart Insurtech companies look at writing code, automating workflows and rolling out new solutions. Technology that previously took significant time and dollars to develop can now be built in a fraction of the time at a lower initial cost. Every conference agenda right now has the same question on it: build or buy? The answer is: Both. The dynamic is not binary.

The organizations moving fastest are buying the foundation and building on top of it when appropriate. As teams navigate how to best leverage internal resources and where to bring in outside expertise, here’s what they need to be thinking about.

Look Beyond the Initial Build

The pitch for building always starts with cost. Yes, A.I. can lower the cost of building a system. I’ve seen it firsthand. But developing a system is just one small part of the equation. Owning and maintaining the system is where time and cost compound.

A true system of record requires an ongoing investment in maintenance, security, compliance and integrations. The actual cost of a build comes long after the code is written, in the ongoing responsibility to update the technology as business needs evolve, regulations change and new technologies emerge. The real question is, who runs it in three years?

We’ve seen most of our new clients migrating away from legacy systems. Just because you can use A.I. to write code, won’t make today’s great idea the new legacy system of the future. Once you factor in the dedicated resources necessary to maintain, innovate, constantly monitor security, patch, and make other updates, the business case for building your own system may not make sense. Without it, every new A.I.-enabled hacker is waiting to pounce on immature code.

Don’t Underestimate the Value of Domain Expertise

While A.I. can generate code in seconds, it can’t replicate decades of insurance experience. Additionally, even the best A.I. requires oversight and review.

Lots of A.I. code is not optimized for scale or efficiency. The smarter the tools get, the more domain knowledge matters. When you build your own platform, you also own the requirements. You supply them and you validate them. Your technology doesn’t benefit from practices learned from other edge cases and configuration patterns.

When I talk to customers, one benefit of Origami’s platform they consistently point to is the implementation experience behind the technology. They want solutions informed by lessons learned across hundreds of organizations, not just within the confines of their own four walls. Also the more bespoke you make something the less it will interoperate with a standardized world. A core platform integrates with dozens of platforms for various use cases, data enrichment, and more. Each one is monitored, maintained, and enhanced.

Make Governance the Imperative, Not the Afterthought

For companies operating in highly regulated industries like risk and insurance, governance is nonnegotiable. It’s also one of the factors that can make building technology internally more costly and time-consuming than anyone expects. You need to know that your systems have the proper access controls, validation and audit trails from the outset.

Before you build, ask whether you have the people to maintain those safeguards in-house. Every code change requires full regression to prevent accidental data leaks to unauthorized users. If a state changes a regulation, can you turn off a specific A.I. function by rule, today? If the honest answer is no, a certified platform is the realistic path. Go in eyes open.

It’s Not Build Versus Buy, It’s Build On Top

To tell an in-house tech team not to build their own solutions would be counterproductive, but so would telling them to build everything from the ground up. I would also be remiss not to clearly articulate that there is a big difference between an internal IT/Tech team, and a group of seasoned SaaS engineers building platforms/systems to be used.

While I’m sure many IT teams long for the days of the 1990’s with huge control, unwieldy power and budgets, I don’t think any business leader wants to return to those days.

No system stands alone anymore. Technology needs to be interoperable and flexible. But it must sit on a solid foundation. The real winners, the companies that will innovate faster, will be the ones that prioritize organized, governed data foundations first and build from there. But remember to ask yourself, what is your core competency? What is the opportunity cost of building X? While the math becomes a little easier with A.I., a distraction is still a distraction. So, adding proprietary IP or utilities to work with your core systems may make sense if it gives your business a differentiator. But when it is simply to “save a few short-term dollars,” ultimately that becomes penny-wise and pound-foolish.

With a trusted system of record as your foundation, you can experiment, connect new tools and extend workflows without rebuilding the base every time. Building with A.I. will keep getting faster and cheaper. The foundation underneath determines how much of that speed you get to keep. &

As Chief Product and Technology Officer at Origami Risk, Ryan Cantor leads the global Product and Technology organization, spanning product management, engineering, UI/UX, cloud operations, and internal IT. With a team of over 300 professionals, he drives innovation and execution across Origami’s high-growth, cloud-based SaaS platform serving the Risk Management and Core Insurance markets.

More from Risk & Insurance