I ran a quick poll on LinkedIn recently asking whether people wanted to see how others have actually gotten ROI from AI.
85% said yes. Another 6% said yes and they’d demo their own results.
So roughly 91% of respondents wanted to see real proof of AI ROI, but only 6% were willing to show theirs!
That silence said more than the poll result ever could because it reflects exactly where most GTM teams are operating right now.
Full of AI activity and still unable to answer the one question that actually matters: what did any of this produce?
Walk into most Marketing Ops or Revenue Ops teams today and you’ll surely find AI running somewhere. New automations, new tools, new prompts embedded in workflows that didn’t exist six months ago. The energy is genuine and the effort is real.
But when someone asks what it produced in pipeline terms, in revenue terms, in a way you’d put in front of a CFO without flinching, the room gets quiet.
This isn’t a performance problem. It’s that most teams introduced AI into workflows without designing a methodology for measuring how AI affects the outcome. The tools changed, but measurement didn’t.
That gap is where all the marketing attribution proof goes to disappear.
Here’s the pattern I keep seeing.
A new AI tool gets bought, workflows shift, and activity picks up. But no one documented where things stood before the tool went live and marketing attribution models don’t get updated. Ownership of what gets measured also stays blurry.
Six months later, when someone asks what the AI investment produced, the team tries to build a retroactive case out of incomplete data and inconsistent signals.
That’s a measurement architecture problem that predates the AI tool purchase by years.
Basically, AI didn’t create weak marketing attribution infrastructure. It just made the absence of it undeniable. When AI is doing things across your funnel (scoring, routing, sequencing, writing) and you can’t trace what it changed and what that change produced, you’re not flying blind because of AI.
You were already flying blind and AI just made it obvious!
The hard thing is that most orgs are scaling AI investment faster than they’re building the foundation required to evaluate it. Which means marketing attribution becomes something teams try to reverse-engineer after the fact. And reverse-engineered attribution rarely holds up.
There’s a specific set of conditions that need to exist before you can tie AI activity to real outcomes. Most teams are missing at least half of them, and no one talks about this part.
You need clean inputs — If the data feeding your AI tool is inconsistent or incomplete, what the tool produces is unreliable from the start. AI doesn’t fix bad data. It scales the consequences of it.
You need baselines set before launch — You cannot measure lift if you didn’t document where you started. This sounds obvious. It almost never happens, because there’s always pressure to deploy quickly and show value fast. Those two things are in direct tension with each other.
You need defined ownership of the outcome — Who is accountable for the metric this AI investment is supposed to move? I don’t mean tool admin; I’m talking business outcome. If that’s unclear before the tool goes live, the marketing attribution will be unclear, too.
You need a thread that connects AI activity to pipeline — A clean, traceable line from what the tool did to what moved downstream. In B2B, where buying cycles are long and buying committees are large, that thread is hard to build. But without it, you’re left with a story you can’t fully defend.
None of these are surprising or unique requirements. This is foundational MOps work, but the problem is that it’s getting skipped in the rush to grow AI investment.
The dominant question in most AI conversations right now is: what are you testing?
That question points toward activity. It gets answered in tool demos and quarterly updates.
The more useful question is: what changed in the system, and what did it produce?
That question points toward marketing attribution design. It gets answered by teams who built the right infrastructure before the AI investment went live, not after.
Those two questions lead to completely different operational behavior. And right now, most teams are answering the first one while hoping the second one takes care of itself.
It doesn’t.
Interest in AI ROI is high but proven AI ROI is still rare. The gap between those two things is almost always a marketing attribution gap, not a technology gap. The measurement layer is the bottleneck here, not the tools.
Mike Rizzo, is the Community-led Founder of MO Pros and MarketingOps.com. With his extensive background in marketing Mike helped build, launch, manage and optimize Mavenlink's first-ever user community and Client Advisory Board programs. He is the founder of MO Pros - The Community for Marketing Operations Pros that is growing with an average of 110 new members every month. He is also the co-host of the podcast called Ops Cast by MO Pros. As the founder of MarketingOps.com, Mike is taking a community-led approach to building resources that are purpose-built for MO Pros.
A while ago, Mike Rizzo asked a pointed question on LinkedIn: are Marketing Operations professionals actually seeing ROI from their AI investments? The poll got tons of responses. At least 91% of people wanted to see how companies were measuring AI ROI but only 6% of those people were willing
Let’s clear something up: MOpza isn’t over but it is evolving. And it’s coming back in 2026. We’ve seen the questions, the comments, the speculation. So let’s be transparent. Because if there’s one thing this community thrives on, it’s clarity and connection. You Spoke & We Listened When we shared
MOps-Apalooza 2025 brought more than four hundred marketing operations, revenue operations, and GTM professionals together at The Westin Anaheim for three days of hands-on learning, live workshops, and community conversations. This year’s program made room for everything from AI agent building to leadership development and tech stack decision-making. Attendees came
