AI in Marketing 2025: Seymour Duncker on Strategy, Personalization & Pitfalls

 1.What marketing functions or workflows have benefited the most from AI so far and which areas remain largely untapped?

While the hype around AI often runs ahead of reality, digital advertising has proven to be one domain where the lift has been real and measurable. At Decision Counsel, we can certainly attest to that. Ad production—especially for programmatic platforms—has seen clear gains thanks to the rules-based structure of digital channels. Marketers can now spin up thousands of ad variations from just a few core creatives, rapidly testing at scale to find what works best. More broadly, the earliest wins with generative AI are concentrated in “maker” workflows. According to Jasper’s State of AI in Marketing 2025 survey, nearly 60% of marketers already use AI for everyday content tasks, including copywriting, ideation, SEO tweaking, campaign testing, and desk research.

Meanwhile, most marketing teams still have a long way to go: only 29% rate their AI maturity as “advanced.” More transformative applications—like enforcing brand governance, automating full campaign workflows, or delivering true one-to-one personalization—remain underdeveloped. And short-form video still has yet to see generative AI break into mainstream production. The next wave of adoption will go well beyond content generation, and will be about finding ways of embedding AI deeper into higher-level decision-making, audience needs, and the full customer journey.

2.From your vantage point, what are companies most commonly getting wrong in their approach to AI-powered marketing?

The most common misstep companies make is expecting magic. Especially in creative teams, there’s often a misplaced hope that AI tools will act like a “one-click” fix—tap a button and watch campaigns write themselves. But creativity doesn’t work that way, and neither does AI. The real work lies in understanding the creative process deeply enough to thoughtfully integrate new tools—augmenting ideation, accelerating iterations, and ultimately optimizing the entire workflow.

Jasper’s survey also found that 67% of marketers now cite “lack of education and training” as the number one barrier to adoption—up from 64% last year. Other leading blockers include “lack of awareness or understanding” (56%) and “lack of strategy” (43%), with “lack of resources” also rising. Concerns about “unknown risks” are decreasing—down to just 25%—showing that fear is fading. The real obstacle isn’t the fear of AI, but a lack of readiness to utilize it effectively. Closing these gaps in literacy, strategy, and governance is the clearest way forward.

3.Was there anything you expected to happen with AI in marketing that hasn’t materialized or that’s taken a different form than anticipated?

Generative AI is seemingly everywhere, but still far from fully integrated. While individual use cases like faster copywriting or sharper research have delivered tangible wins, most marketing teams still haven’t figured out how to connect these wins across the full value chain. Critically, few have linked AI efforts to ROI in a way that feels systematic or scalable. In other words, the novelty has worn off. What’s missing now is the operating model—the repeatable processes, playbooks, and cultural alignment—that elevates isolated success into enterprise-wide transformation. Sixty-three percent of marketing teams already use generative AI, 78% of which report improved outcomes, according to Jasper. However, only 43% of adopters have formal enterprise-level AI programs in place. Most teams only began serious experimentation in 2024, and full-on pilot failures are rare—just 3%. The tools are working. What’s lagging is the strategic framework to turn experimentation into a durable advantage.

4.What risks do brands face when personalization becomes overly automated or intrusive? How can they avoid that trap?

There’s a fine line between personalization and intrusion—and brands are stumbling across it. When AI-driven personalization becomes overly automated or impersonal, it doesn’t deepen loyalty; it triggers backlash. Simply labeling a product as “AI-powered” can dampen consumer enthusiasm, with broader concerns about AI, such as data privacy and job displacement, also coming to mind. Instead of feeling seen, customers might start feeling watched, and that’s a trust killer. Duolingo and Audible each learned about these sorts of pitfalls the hard way earlier this year. After all, personalization should be a tool for empathy, not efficiency at all costs.

Avoiding the over-personalization trap means finding a way of blending hard systems with a soft touch. That starts with privacy-by-design: only collect what you need, get clear consent, and explain the value exchange in easy-to-understand language, not legalese. Then integrate real-time sentiment tracking, straightforward opt-outs, and emergency “kill switches” that halt personalization flows when a user shows signs of discomfort. Brands that get personalization right anchor it in the C-suite. As McKinsey advises, a “leadership triad” of CEO, CMO, and CFO should jointly own the AI agenda—so goals, ethics, and accountability stay intertwined from day one.

5. What are the biggest challenges marketers face when implementing AI from data, tech stack integration, to internal alignment?

The biggest AI implementation hurdles often are less about adopting new tools and more about overcoming the drag of old ones. Most marketing tech stacks were designed for batch email campaigns and clickstream analytics, not for AI-driven workflows powered by real-time vector embeddings or agents. Data silos and technical debt are real and widespread: 86% of enterprises say they must overhaul their systems to deploy AI agents effectively. And even where the desire is there, many teams are still buried under the weight of maintaining legacy infrastructure. Even with the right tools, transformation stalls if people and systems aren’t ready. Many marketing teams aren’t yet AI-literate and still wary. AI’s biggest gains, however, come from deep workflow integration. The art is in pushing forward without getting stuck in the weeds of outdated systems. The solution lies in incremental embedding, thoughtful upskilling, and change management that balances the potential of AI with the inertia of the past.

6. Is there anything you don’t think AI will solve or shouldn't try to in the marketing space?

AI should never solve for the heart and soul of a brand. AI can and will increasingly help and aid in that process, certainly, but ultimately that’s the marketer’s job, and it requires an elevated sense of empathy, nuance, perspective, wisdom and judgment that, as far as I can see, only humans and the human experience can provide. Without true heart and soul—the kind you get from a human marketer—branding will always fall flat, and no amount of AI innovation can reproduce the trials, tribulations, fear, loathing, and joy of a person pouring themselves into a brand.

AI is a tool, not a crutch. And it can very easily be misused or overused. Discernment is key, and as AI advances, so will the importance of the ability to discern when AI is appropriate and when it is not.

Get in touch with our MarTech Experts.