Nearly a third of organizations have already decided against buying software they could build themselves using AI coding tools. That finding comes from McKinsey’s State of AI 2026 report, which surveyed 1,719 respondents across 97 countries. AI coding tools have made building faster and cheaper than most executives thought possible just a few years ago. Organizations at every scale are treating it as a live option. The total cost of ownership is where the math changes. Why the Build Option Feels Compelling Right Now AI tools dropped the cost of writing code quickly and significantly. What used to require a full engineering team and several months of work now takes a small team and a few days. A working prototype in hours for almost nothing is hard to argue with in a budget meeting. Teams are seeing it happen, and once you’ve watched it work, the case for vendor software feels harder to make. A self-built system feels like a cost you control. You decide what to invest in, when to upgrade, and what to leave alone. Vendor software comes with ongoing license costs that grow every year. But what makes sense at first might not add up when you get to year two. What AI Actually Changed About Software Costs AI tools lowered the cost of writing code. Writing code was already the cheap part of owning software. The expensive part is everything that follows the first working version. The moment a system goes live, the real costs begin: Integrations with other tools, which take time to build and break without warning when upstream systems change. Security patches and compliance updates that can’t be delayed. Bug fixes for edge cases that only surface in production. Data quality and governance work that a vendor builds into its roadmap. Key-person risk when the engineer who built the system moves on. None of those costs dropped because AI can generate code faster. The maintenance burden, the integration work, the security surface area, the engineering hours consumed by upkeep when the team could be building new things: all of that stayed where it was. AI lowered the starting line costs and left everything beyond them exactly in place. What the Total Cost of Ownership Calculation Actually Shows Build-vs-buy comparisons usually start by putting the build cost next to the license cost. That’s the wrong comparison to make. A complete total cost of ownership for software includes the build cost, ongoing maintenance, integration costs, security and compliance upkeep, upgrade costs when underlying infrastructure shifts, and the opportunity cost of internal engineering capacity absorbed by support work. Run that calculation over three years, and the picture changes. The custom build that looked cheap at launch usually ends up costing more than the vendor option. It also does less, and has fewer people working to make it better. The third of organizations in McKinsey’s survey that have already skipped software purchases in favor of custom builds are now carrying those full ownership costs. The many more planning to follow are about to. Some of those decisions will prove out. Many will look different in 24 months than they do today. What This Means for Risk and Insurance Buyers For organizations managing risk, insurance, and safety programs, the build decision compounds in specific ways. The integrations required, the regulatory compliance requirements, the reporting and audit standards, and the data quality expectations are all core to whether the system delivers value, and they don’t get simpler to manage in-house over time. A purpose-built vendor platform carries a product roadmap shaped by years of domain-specific development, a support organization, an established security posture, and a client base whose real-world needs drive ongoing improvement. That accumulated investment is part of what the license fee buys. Building a system that reaches that level of depth from the ground up takes more than an AI coding tool and an engineering sprint. Origami Risk clients regularly point to domain functionality depth, implementation speed, and ongoing platform investment as what sets working with a purpose-built system apart from managing a custom build. The comparison over a three-year horizon covers total cost of ownership for software, functional breadth, and what the organization’s internal team can actually focus on each year. The build-vs-buy question deserves a full answer grounded in the real numbers. If you’re working through that decision, the ebooks below walk through the complete calculation for each track. Learn more: Download the build vs. buy guide for insurance carriers. Download the build vs. buy guide for risk management teams.