A client forwards you their OpenAI invoice with a one-line question: “Is this normal?”
You don’t know. And here’s the uncomfortable part — neither do they.
That conversation is coming to every firm, whether or not you have an answer ready. Businesses now pay AI providers like OpenAI and Anthropic directly, and that spend behaves nothing like the software you’re used to advising on. It isn’t a predictable per-seat subscription. It’s usage-based, it moves fast, and it’s scattered across provider projects, API keys, and individual people. The bill arrives as a single number that keeps climbing — with nothing to explain what’s driving it.
This is the fastest-growing line in business software spend
Ramp, reporting on its own customer base, found that AI token spend grew 20.7× between June 2025 and 2026 — outpacing every traditional software line it tracks. Unlike seat-based tools, this spend scales with usage, so it can balloon across teams in weeks, not budget cycles.
Your clients are living inside that curve right now. Most of them can’t see it.
Why this lands on you specifically
You already own the “where is the money going?” question for the rest of their stack — the duplicate subscriptions, the forgotten renewals, the tools nobody uses anymore. AI spend is simply the newest category in that same audit, except it’s the one currently sitting outside your view.
That’s the opening. Not the engineering team, not a developer cost tool bolted onto their infrastructure — you are the advisor they already trust with their books, the one they’ll actually ask. The firms that treat AI spend as the next advisory line rather than a technical problem to punt back to the client will be the ones those clients lean on hardest over the next year.
The question isn’t whether your clients will ask about their AI spend. It’s whether you’ll have something to say when they do.
The good news: bringing AI spend into the same audit you already run is closer than you think. More on that this week.
