GATE ARCH:StructuralBuy-Side Audit
For CIOEvery initiative, its own ruler
“Because a measure written and reported by the party that built the thing is a claim, not a check — whether that party is a vendor or your own team.”
Iceberg Tribe Advisory • AI Decision Architecture (Build & Buy)
Every piece here audits an irreversible decision an institution faces across the five enterprise AI gates — whether commissioning an in-house build, approving an internal go-live, or procuring from an external vendor. Not trending topics, not a content calendar.
When an audit reaches an enforceable standard — a warranty review for vendor contracts, or an Internal Build Checklist and Initiative Register for systems built in-house — it points directly to Instruments: runnable diagnostic tools, not another conceptual post.
The Five Decision Gates
Irreversible milestones across the enterprise AI decision lifecycle — both in-house builds and vendor procurement
“Because a measure written and reported by the party that built the thing is a claim, not a check — whether that party is a vendor or your own team.”
“Because vendors routinely invert data protection restrictions (such as DIFC Regulation 10) into affirmative procurement mandates.”
“The Capability Grade Ladder: a claim may not be cited above the grade its evidence reached.”
“Because enterprise AI agreements often grant protective contractual clauses while withholding enforceable rights.”
“Distinguishing vendor-supplied assumptions from institutionally verified figures when boards evaluate AI business cases, tracing whether reported numbers originate from the counterparty being paid.”
“Because vendor and consultant build-or-buy models omit opposite cost categories by design, the buy-versus-build decision depends on whose arithmetic authored the total cost of ownership.”
“Because a category is a compressed argument, adopting a vendor or analyst's vocabulary quietly commits an institution to their boundaries, metrics, and options before any formal decision is made.”
“An outcome shown once under the reporter's own controlled conditions is evidence about the report — when the firm that measures the outcome, writes the case study, and runs the benchmark is the same firm at every step, the proof is an internal methodology rather than independent validation.”
“Appointing a chief AI officer resolves the optics of a vacancy without resolving the governance gap — because a seat given internal champion incentives and no stable instrument cannot act as an independent check.”
“Authorship of a measure defaults to whoever has both the incentive and the capability to build it — here, the same strategy practice that built the diagnostic, ran its calibrating survey, and sells the redesign it recommends.”
“Doctors diagnose and prescribe under external licensing board standards — AI vendors prescribe remedies for unmanaged consumption without an external check on fit.”
“Because the problem arrives pre-framed in categories vendors and analysts built — and a tool can only see the shape of context its own instrument was built to detect.”
“Because the check that matters happens after signature, and by then everyone close enough to run it has a decision to protect.”
“Because whoever picks the benchmark, runs the demo or keeps the score has a stake in how it comes out — and the buyer's own side of the table is not exempt.”
“Because whoever builds the ROI model has a position on how it turns out — and the person who signs it inherits one.”
“Because whoever names the problem also shapes which solutions qualify — and nobody who usually handles that sentence is free of a stake in it.”
“Because whoever builds the cost model has a stake in the answer — and so does everyone else positioned to check it.”