The AI ROI Divide: Why Marketing’s AI ROI Numbers Don’t Add Up

The hidden costs of AI in martech (and why they distort ROI)

ROI is a ratio of value output over cost input. Martech has been a major contributor to the cost side of the equation for years. Even before every platform started adding AI-enabled features, martech spend added up to $131 billion in 2023. But current estimates put martech costs at $215 billion by 2027.⁵

Pre-AI, your marketing techstack was a bicycle pump. In the age of AI-enabled-everything, it’s an air compressor.

Surveys estimate nearly two thirds of martech budgets now go to AI,⁶ and 61% of marketers now say overall cost is their top martech challenge.7 

But the costs run deeper. The IBM Institute for Business Value found the average cost to compute — including cloud and infrastructure — has soared roughly 90% in the past two years,⁸ and most executives say generative AI is why.

While leadership isn’t wrong in blaming rising costs on AI, there are several factors contributing to AI expenses:

• Multiple tools offer AI for the same tasks (i.e., lead scoring, content generation, SEO), so many teams pay for the same AI-enabled features across several tools and only use them in one (or worse, not at all).
• Vendors continue to add new AI features faster than marketing teams can adopt them, so platforms get more expensive while usage stays flat.
• Some platforms automatically enable AI-powered features that trigger additional compute, storage, or API use, raising costs even when teams aren’t intentionally using them.
• AI tools often rely on behind-the-scenes cloud processing every time you use them, stealthily increasing infrastructure costs.

This is perhaps the biggest question surrounding AI marketing ROI: are AI-related costs outweighing AI-related gains? If so, how long will it take for the tide to turn favorably?

That will depend on how quickly marketing teams establish AI governance to rein in excessive martech creep and adapt workflows to capitalize on automation opportunities. We deep dive into the importance of strategic AI guidelines to balance the strengths of humans and technology in another article: The Pitfalls of All-or-Nothing Approaches to AI.

Why most teams aren’t using the AI they already pay for

While martech budgets are ballooning so dramatically they could feature in the Macy’s Thanksgiving Day Parade, most teams aren’t even making use of the AI capabilities already embedded in their techstack.

• Their AI modules are licensed, but not activated.
• Marketers don’t know which tools have AI features.
• They continue with manual workflows despite options to automate.
• Their organization lacks centralized tracking for AI adoption and usage.
• Teams receive little to no training on AI tools.
• They buy AI tools without evaluating overlap with existing platforms.

The CMO Survey found that while many teams report modest gains from AI, most aren’t using it for advanced functions such as predictive analytics or media optimization, even if their tools offer those features.¹

That’s not necessarily because teams don’t think to use AI for those purposes; in many cases, they can’t: 74% of U.S. teams say they’re not getting the most out of AI because it can’t access the right data.⁴

This further muddies the water on AI returns because you can’t measure ROI on AI that marketing teams aren’t using (or aren’t using consistently). It ties back to a central, recurring point in the data: most companies are giving teams access to AI, but not restructuring how they work to fully incorporate the technology.

To actually use the AI tools you have instead of just collecting them in an expensive techstack, your team needs to know what tools they have, how to use them, and have an infrastructure that supports leveraging the features you’re paying for.

Want consistent value from AI? Build it into your team’s workflows rather than leaving its use to individual decisions. AI should add convenience, not hurdles.

The marketing leadership dilemma: rising expectations, unclear outputs

Marketing leaders are feeling added pressure on three fronts since the AI surge:

• The C-suite expects clear ROI and efficiency gains.
• The CFO wants justification for rising martech budgets.
• The marketing team needs guidance on which tools to use and how.

You’re trying to meet all these demands immediately, so dashboards are multiplying and you’re overwhelmed with AI vendors stuffing your inbox, promising results. And even after buying the tools and comparing data, a gut-wrenching question lingers: “Are we actually getting ROI from AI, or are we assuming we are?”

Unfortunately, many marketing teams are in the assumption camp because they haven’t yet made foundational adjustments to optimize their processes for AI integration. Updating workflows for AI and setting up KPIs to measure results are key steps both McKinsey and Atlassian outline to see real financial impact.^{2 4}

Leaders can breathe easier about answering tough questions when they do the boring work: setting up an environment for automation to produce measurable ROI — not just buying AI tools and hoping.

Measuring AI marketing ROI (a practical framework)

If there’s an upside to the mismatched numbers on AI ROI so far, it’s that leaders know not to consider only revenue and conversions or only productivity — both matter.

This list can help you view the full picture of the impact of AI on your marketing efforts:

1. Tool utilization: is your team using the AI features in your martech stack?

Look at activation rates, usage frequency, and adoption (at the user level and feature level).
If your team doesn’t know the AI features included in your licenses, there’s a good chance they’re not even activated. You can’t measure ROI on features no one’s using, but you’re definitely paying for them.

2. Workflow impact: which manual tasks has AI replaced?

Map tasks before and after AI adoption. Compare time per task and end-to-end cycle duration.

It’s not a matter of whether AI is faster — you know it is — but whether your team has stopped doing tasks the manual way. If you’re running AI alongside traditional processes “just to be safe,” you’re doubling work instead of replacing it.

3. Performance impact: have your results changed?

If you’re A/B testing, look at AI vs non-AI segments to track and compare KPIs (i.e., conversions, cost per lead, revenue).

4. Cost alignment: do your tools offset spending or add to it?

Assess function overlap, redundant licenses, cloud consumption, and AI-related surcharges.

Tools that save you time could be costing you elsewhere. Weigh whether the AI content generator reducing your copywriting hours is worth a potential $2K/month spike in your cloud computing bills.

5. Redundancy: do you have multiple tools with the same features?

How many of your tools offer AI-driven lead scoring, content creation and optimization, or automation? Most marketing teams don’t even realize their techstacks have two, three, even four tools that do the same work. For example: if you have HubSpot, Marketo, and Jasper, you’re paying for AI content generation three times.

6. Capability alignment: have you adapted your workflows and trained your team for AI?

To cross the bridge from AI contributions that are neat to fundamentally shifting processes for scale and ROI, marketing needs to revise workflows for automation and teach humans to fully take advantage of it. We can’t emphasize it enough: workflow redesign is the top predictor of AI ROI.