Most enterprise AI conversations land on the same three choices: buy a platform, build in-house, or orchestrate a mix of both. Frameworks like the 3C model, which scores an initiative on Capability, Complexity, and Criticality, exist because most AI investments are still working to earn back their cost (Forbes / Plug and Play 2026 Enterprise AI Strategy Pulse Survey). The model still holds up. But one of its three axes has started to behave differently from how it was designed to.
Where the grid behaves differently
The 3C logic has always been simple. Low Capability pushes you toward Buy, because building was assumed to need a resourced engineering org. The framework itself is fine. What is aging is the assumption baked into it, that building requires that kind of org in the first place.
AI-assisted development has collapsed the cost of turning a well-documented process into working software. The hard part was never the coding hours. It was understanding the process well enough to encode it correctly. So Low Capability no longer forces Buy. It changes who can build, and how fast.
It starts with the standard stuff
The shift shows up first in workflows nobody argues about: procurement scoring, visitor management, travel and expense. These are Low Complexity almost by definition, since the rules are already written down, and mid-tier criticality at most. For years the default here was an off-the-shelf module, because building anything custom was not worth the engineering investment.
That math has already shifted at the market level. In Retool’s 2026 Build vs. Buy survey of enterprise builders, 35% had already replaced a purchased tool with something custom-built, and 78% expected to build more within the year. Separately, comparable workflow tooling that used to take six to twelve months is now increasingly reported at two to eight weeks. None of it needed a bigger team. It needed someone who understood the workflow to direct the build.
Then it reaches the floor
The harder test comes when Criticality climbs. Shop-floor safety SOPs, for instance, are usually mature and well understood, so Low Complexity, but they tie directly to compliance and continuity, which puts them squarely in High Criticality. By the 3C grid, that combination has always pointed to Buy, because the stakes felt too high to risk an in-house build.
That is no longer automatic. Once the capability barrier drops, a Low Complexity, High Criticality process becomes buildable on the same terms as procurement or visitor management, often with AI built into the workflow itself: computer vision watching for hazards on the floor, or simpler anomaly reporting from workers’ audio notes. Nothing about the stakes changed. What changed is the cost of encoding a well-understood process correctly, and that has fallen far enough to make the build viable.
What still holds
Complexity and Criticality have not moved. They still tell you how much governance, testing, and review a build deserves. A safety workflow still needs more scrutiny than a travel approval, no matter who builds it. What has moved is capability, and it has moved for a far wider set of organizations than the framework assumed when it was written. Complex, cross-system integrations still deserve careful orchestration. True commodities still make sense to buy.
Whether this kind of build works across unrelated functions that share nothing but well-written SOPs is not really the question anymore. The market data says it does. The open question is internal: which of your own Low Complexity, High Criticality processes are still on Buy out of habit rather than math, and what would it take to find out?
Sources: Retool 2026 Build vs. Buy Report;
Zylo 2026 SaaS Management Index;
Forbes, Enterprise AI ROI production survey.
