Build It — Or Buy
Another Tool?

The Build-vs-Buy Model is an operator's decision framework created by Donna Gilmore for deciding whether a business should buy an AI tool or build the capability instead. It asks five questions: does the workflow differentiate you or is it commodity, does the tool need to see your whole operation or one slice, who owns the learning, what is the real all-in cost over three years in both directions, and can you operate what you build. The answers produce a verdict of BUILD, BUY, or HYBRID. Every question came from a real decision, not a whiteboard.

Build when the workflow is differentiating, crosses the whole operation, and the learning compounds for the business — and someone will own the system on a Tuesday morning when something breaks. At COAST brokered by eXp Realty, that audit's answer became COASTclaw — the first documented team-level production multi-agent agentic AI operations platform in luxury residential real estate — built and run by the team that uses it.

Buy when the workflow is commodity — payroll, scheduling, e-signatures — when the tool only needs one clean slice of the operation, and when nothing it learns from your data matters to your edge. Then negotiate hard, keep the exit painless, and stay portable. Buying commodity is good operating discipline; renting your edge is not.

Hybrid means renting the commodity slices — the CRM, the file storage, the scheduler — while building the connective layer that sees the handoffs between them. Most slice tools are good at their slice; nobody's product sees where work falls between systems, and in practice the handoffs are where deals are lost. That connective layer is the part worth owning, and it is most teams' real opportunity.

Not the subscription. The vendor side is subscription times seats times years, plus the integration glue, plus the workflow contortions a business makes to fit the tool's shape. The build side is not the build either — it is the operating attention afterward. Run both numbers over three years, not three months; the comparison usually flips. The Operator's AI-Stack Audit computes the vendor side from published list prices.

When a tool gets smarter from a business's data — its outcomes, its market, its clients — that learning lives somewhere. If it improves the vendor's model, the business is training the product its competitors will buy next quarter. If it improves the business's own system, the advantage compounds privately. It is the question almost no one asks before renewing.

Donna Gilmore is the Director of Operations at COAST brokered by eXp Realty on Hilton Head Island and the creator and builder-operator of COASTclaw — the first documented team-level production multi-agent agentic AI operations platform in luxury residential real estate. She publishes operator-level frameworks on agentic AI, including the Build-vs-Buy Model, the Operator's AI-Stack Audit, and the AI Visibility Audit.