Not long ago, getting a business question answered quickly was costly: you needed an in-house analyst, a well-maintained warehouse, and a dashboard for every recurring request. Then AI tools made instant answers essentially free. Strangely, the analytics budget didn’t shrink. Speed dropped to zero cost. Defensibility became the expense. TL;DR: Turning data into an answer used to require analyst time, data engineering, and tooling overhead. AI has driven that marginal cost down to almost nothing for any team that wants to ask a question. The money that once paid for speed didn’t vanish. It shifted into three buckets: verification work, the fallout from acting on incorrect numbers, and the infrastructure that prevents both. The new core metric is cost-per-trusted-answer: what you spend so you can confidently stand behind a number in a meeting without re-deriving it by hand. The most efficient way to buy defensibility is at the data layer: governed, shared metric definitions; computations executed in tested code instead of a language model; and a system that is honest enough to admit when it doesn’t know. A marketing leader planning next year’s budget should focus on the cost of defending a number, not the time it takes to produce one. Introduction Today, nearly every AI tool can respond to an analytics question in seconds. The marginal cost of a quick answer is effectively zero. So why does the analytics budget look almost identical to what it was two years ago? The explanation is uncomfortable.…