We’ve had some version of the same conversation probably a dozen times in the last few months. A business leader says they’re “investing in AI” for marketing. We ask what that means, and the answer is usually one of two things: their team is using ChatGPT to write content, or they’re paying for a handful of tools someone subscribed to after a conference talk made a compelling case.
Neither of those is an AI budget strategy. Neither one is actually wrong, but without a framework, these tools start to accumulate until someone pulls the invoice report and asks what all this AI spend is producing.
What if we didn’t lump all AI spend into 1 line item?
The thing we push back on first is the idea of an “AI budget” as its own line item. Lumping all AI spend into one category is about as useful as grouping paid search, brand work, and experimental content together and calling it “marketing stuff.”
I think we all need to move away from “how much are we spending on AI?” toward something more useful: what is each AI investment trying to do, and is it doing it? That question has a real answer. And it changes what you do next.
The three-bucket framework we use — Performance, Transformation, and R&D — is a great fit here as well, where AI tools span all three.
AI in your performance bucket
Performance is your near-term pipeline work. The AI tools that belong here are the ones helping you reach the right buyers faster, qualify them more accurately, and convert them more efficiently. This covers things like ad optimization, AI-powered lead scoring, AEO (Answer Engine Optimization, meaning your content shows up in AI when buyers are actively researching), and content generation for high-intent search.
Performance marketing has one job: generate pipeline in the near term. Any AI tool in this bucket gets evaluated against that: CAC (Customer Acquisition Cost), conversion rates, and pipeline contribution. If it doesn’t move those numbers within 90 days, it doesn’t belong here.
AI in your transformation bucket
Transformation marketing is the investment that builds your brand, your authority, and your presence with the 95% of buyers who aren’t actively in-market right now. AI tools show up here too; they just look different. Less automation, more amplification. Thought leadership content that earns presence in AI-first search. Systems that help your team show up consistently without burning out on production. Tools that keep you in the places where buyers are doing their research before they ever raise their hand.
The return timeline is longer, 6 to 18 months before it’s really visible. But that’s also why it compounds, right? The research phase has gotten longer and more independent. AI has changed how deep buyers dig before they ever surface to sales. Transformation investment is how you stay in those conversations even when no one’s talking to you yet.
AI in your R&D bucket
This is where the actual experiments live. And in a space moving as fast as AI is, getting this bucket right matters more than the other two right now.
We’ve talked about this. R&D is where you make small, disciplined bets on what might work before you know for certain. Our AEO practice started exactly here. We ran it on ourselves for months before it became a real service line. The AI chatbot we tested? Also R&D. It didn’t move the needle the way we hoped, took more leadership time than it was worth, and we cut it fast once we saw that. (Drippy still hangs out on the site if you need him.) One failed. One became core to how we serve clients. Both were worth running.
Right now, agentic AI for outreach and follow-up, account-level personalization, tools for synthesizing market intelligence faster than your competitors… these are all worth structured experiments. The ones that prove out get graduated into Performance or Transformation. The ones that don’t, you take the learning and move on.
The mistake is letting FOMO run the R&D budget. There’s a real difference between, “we’re testing this because the hypothesis is solid” and, “everyone’s talking about it, so we subscribed.” Jasper’s 2025 State of AI in Marketing report found that 51% of marketers say they can’t measure the ROI of their AI investments at all. That gap is almost entirely the second approach: tools bought without a defined hypothesis, held to no timeline, renewed on autopilot.
The practical starting point
Pull up your current AI-related spend and do a simple audit. Map each tool to one of the three buckets. Ask whether it’s being measured against the right timeline for that bucket. And if something doesn’t fit cleanly anywhere, that’s worth figuring out before you hit renew.
AI is a capability. How you invest in it, and what you hold it accountable to, is the strategy.
If the full picture of how to allocate across these three buckets is useful, we laid it out here: How to Manage Your B2B Marketing Investments.
