Milena Traikovich is a powerhouse in the demand generation space, specializing in the delicate art of turning technological potential into measurable revenue growth. As a seasoned expert in marketing technology and enterprise AI strategy, she has spent the last several years helping global organizations navigate the messy transition from manual workflows to autonomous systems. In our conversation today, we dive into the stark realities of the current martech landscape, where the rush to fund AI is frequently outpacing the ability to prove its worth. Milena provides a grounded perspective on why so many AI initiatives are stalling in the pilot phase and how leaders can bridge the gap between “faith-based” investing and undeniable financial impact.
With roughly 40% of martech leaders currently unable to demonstrate a clear financial return on their AI investments, what do you believe is the primary disconnect between spending the budget and proving the value?
The disconnect is truly a systemic issue rooted in how quickly these budgets were mobilized compared to the maturity of the measurement frameworks available. We are seeing two-thirds of enterprise organizations rushing to establish dedicated AI budgets, yet they are often flying blind because 40.2% of leaders haven’t yet identified the right financial proxies to track. It is a high-pressure environment where the “fear of missing out” has led to a surge in experimentation, but without the foundational data to back it up, many teams are left flailing in the boardroom. I see a lot of visceral anxiety among managers who have been given millions to spend but find themselves relying on “faith-based” justifications rather than hard attribution. When you look at the numbers, those who can actually provide proof are 56% more likely to see budget increases, while those who can’t are stuck at a much lower 37.5% growth rate. It’s a wake-up call that the honeymoon phase of “experimental funding” is ending, and the era of accountability is finally here.
Data shows that AI is having a much more significant impact on internal operations at nearly 70% compared to customer experience at 60%. Why is it so much easier for organizations to optimize their back-end workflows than it is to revolutionize the customer-facing side?
It essentially comes down to the “safety” of the internal sandbox versus the unpredictable “wild” of the public-facing market. For about 69.3% of organizations, AI is making a substantial dent in the martech stack through internal functions like data management, reporting, and creative variation—things that live behind the firewall where a mistake won’t break the brand. Internal teams feel a sense of sterile comfort when using co-pilots for analysis or video generation because they can control the variables and the output. Conversely, the 59.9% impact on customer experience reflects a much more cautious, almost hesitant approach to journey optimization and real-time decisioning. There is a deep-seated fear that an unchecked agent could offer a wrong discount or provide a hallucinated answer to a loyal customer, which is why we see high-value areas like loyalty optimization and campaign creation still struggling to find their footing. Marketers are essentially prioritizing the “plumbing” of their operations because it’s easier to govern, even if the customer never sees the immediate benefit.
Only a tiny fraction of companies—about 1.6%—allow fully automated AI content to reach customers, while 43% still require heavy human intervention. How does this “human-in-the-loop” necessity affect the promised productivity gains that justify these expensive AI tools?
This is the great paradox of modern enterprise AI: the time we supposedly save on production is being cannibalized by the grueling labor of verification. When 43.3% of brands refuse to let anything out the door without a human editor’s eyes on it, and another 24.4% limit generative AI to internal use only, the “efficiency” we were promised begins to feel like a mirage. I’ve seen teams where the creative process is technically faster, but the “exhaustion of the edit” is real—staff are spending hours fact-checking AI-generated claims or smoothing out the robotic tone of a draft. This creates a bottleneck where the high cost of the technology isn’t actually lowering the cost per lead because the human headcount hasn’t shifted; it has just changed roles from “creators” to “glorified proofreaders.” If the underlying workflow is shaky or the data being fed into the model is flawed, the AI just produces mistakes at a higher velocity, forcing humans to work even harder to catch them.
With 95% of organizations putting AI agents on their roadmaps, but only about 10% achieving full-scale production, what are the hidden “brick walls” preventing these pilots from becoming permanent fixtures?
The gap between a 95.3% roadmap adoption rate and a 10.8% production rate is a sobering reminder that planning is easy, but execution is a battlefield. The “hard part” that many leaders underestimate isn’t getting an agent to generate a response; it’s ensuring that the response is consistently true and compliant across millions of interactions. We are seeing agents for copywriting and data analysis move ahead because they are relatively “low-stakes” and isolated. However, when you try to scale agents for attribution, offer decisioning, or audience selection, you run into the “brick wall” of fragmented data streams and legacy silos. Many organizations are finding that their current data infrastructure is too shaky to support an autonomous agent, which essentially turns the agent into a liability rather than an asset. It is a frustrating plateau for many martech teams who have the vision, but lack the “data hygiene” required to let the AI actually take the wheel.
In the current climate, where AI funding is plentiful but “proof” is the harder currency, how should marketing leaders change their approach to the annual budget meeting to ensure their initiatives survive?
The strategy has to shift from talking about “potential” to talking about “proxies.” Since 40.2% of leaders are struggling to show a direct financial line, they need to start building better bridges between AI activity and accepted financial metrics. You can no longer walk into a meeting and claim that AI is “transforming the culture”—you have to show how it is lowering the cost of customer acquisition or increasing the velocity of the sales cycle. I advise leaders to focus on the “visible” AI—the tools that make work tangibly better and faster—rather than the “invisible” AI that just creates more noise. Those who can prove value are seeing a 56% budget increase because they’ve turned the technology into a revenue driver rather than a cost center. It’s about moving away from “faith-based” marketing and returning to the discipline of hard numbers, even if the tools we are using are brand new.
What is your forecast for the future of autonomous marketing agents as we move toward the end of the decade?
My forecast is that we are about to enter a period of “mass consolidation” where the thousands of tiny, specialized agents we see today will be swallowed by larger, integrated orchestration platforms. By late 2027, the focus will shift entirely away from “generative” capabilities and toward “agentic” capabilities—systems that don’t just write emails but actually execute entire cross-channel strategies with minimal oversight. However, this will only be possible for the small percentage of brands that have cleaned up their data act; the rest will continue to struggle in that 10% production limbo. We will likely see a significant jump in the 1.6% of fully automated content as trust in “guardrail technology” improves, but it will be a tiered landscape. The winners will be the companies that didn’t just buy the AI, but invested in the human expertise needed to govern it, ensuring that their autonomous agents are actually serving the customer rather than just “flailing faster.”
