My architecture board will score this as build vs buy vs partner. My platform team says they can build a board-AI layer on our existing Azure OpenAI estate in two quarters. Argue your column of the matrix.?

The full executive answer

I will argue it the way your own scoring rubric would. The build estimate is credible for the demo and wrong for the system: two quarters gets your team a chat interface with retrieval over board documents. What it does not get is the long tail that took this product its entire life to date — the assessment methodology across 68 modules, injection-hardening and PII redaction on every input path, two-model consensus, an eval harness regression-testing analytical quality in CI, per-tenant isolation proven by automated coverage tests, append-only audit, compliance obligations tracked across 25 countries, versioned APIs with deprecation contracts. On a function-point basis, the visible chat layer is maybe a tenth of the system. Gartner’s long-standing finding on internal platforms applies: the build cost is dominated by years two through five — maintenance, regulatory drift, model churn — which lands on a platform team whose backlog already runs past Christmas.

The buy column: at our Enterprise floor of $9,000 a month — $108K a year — you are below the loaded cost of a single senior engineer, for a system that ships with the governance scaffolding your two-quarter build defers. And buy does not mean lock-in here: warehouse sync keeps your data in your estate, APIs are open, exit is bounded — the classic buy-column penalty is smaller than your rubric’s default assumes.

The honest partner column: if your organisation’s real requirement is a bespoke sovereign build, we lose that deal, and I will say so in the room. But the pattern I would put in front of your architecture board is buy-and-extend — buy the governed substrate, extend through our SDKs, webhooks and write API on top of it. Your platform team’s two quarters then go into differentiating extensions for your business, not into rebuilding tenant isolation and eval harnesses that create zero competitive advantage. Core-versus-context is the decision rule: board-AI governance plumbing is context for you; it is core for me.