AI & Marketing Research with Dr. Eva Wolf
Not another AI news podcast. This is a research radar — a twice-weekly briefing that surfaces peer-reviewed studies on AI and marketing, tells you what the evidence actually says, and helps you decide what's worth a deeper read.
AI & Marketing Research with Dr. Eva Wolf
AI Marketing Research: Generative Recommendations, AI Frameworks & Adoption
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Thanks for listening to AI & Marketing Research Radar by Big Plans Media.
I’m Dr. Eva Wolf, and I help marketers, educators, consultants, and business owners turn AI marketing research into practical strategy, smarter workflows, and real business opportunities.
More episodes: https://bigplans.media/ai-marketing-research-radar/
Consulting: https://bigplans.media/ai-marketing-consulting/
Big Plans Media — Where Big Ideas Meet Smart Marketing.
You're listening to Evita, an AI-generated research briefing avatar trained on the research framework and methodology of Dr. Eva Wolf, marketing professor, AI researcher, and founder of Big Plans Media. Every day, Evita scans emerging research in AI, marketing, consumer behavior, psychographics, and business strategy to identify the most relevant developments, opportunities, and risks worth watching. These daily radar reports are designed to help busy professionals stay informed without having to read hundreds of research papers themselves. And every Friday, join Dr. Eva Wolfe live for her personally recorded weekly AI Marketing Radar Roundup, where she breaks down the biggest stories, explains what actually matters, and shares practical insights and strategic implications for marketers, educators, entrepreneurs, and business leaders. Now, here's today's radar report. Here's the signal I can't ignore today. Your AI recommendation engine might be doing three things at once: lifting clicks, trapping customers in a bubble, and quietly failing every new product you launch. And most marketing teams have no idea which of those is happening right now. Today's papers point to the same pattern. AI in marketing is maturing fast, but the gap between what's technically possible and what's actually deployed safely is wider than most practitioners think. We screened 391 papers. Three cleared the full text bar and made today's radar. Quick caveat: this is a first pass research briefing, not a final academic review. Every paper today has full text access. I'll tell you what the papers suggest, what they don't prove, and which ones deserve a deeper read. Okay, let's get into it. Paper one. Here's the business question. If your AI recommendation engine has zero data on a new product or a new customer, what actually happens to your conversion rate? This is a full field survey from a team that includes researchers at Google DeepMind, Amazon, and UC San Diego. They mapped every major approach to what they call generative recommenders. Hmm, not systems that rank existing items, systems that can generate content, conversations, and personalized visuals. Here's the finding that matters most. Generative AI systems, the kind that can have a real back and forth with a shopper, are specifically better at the cold start problem. That's the industry term for when your algorithm has zero history to work from. New product drops, new customers, first visit. Traditional recommendation engines essentially go blank in those moments. LLM-based systems don't. They can reason about the product and the context without needing prior purchase data. And the monograph also covers multimodal generation, meaning AI that doesn't just suggest items, it generates personalized visual content, virtual try-ons, custom product ads tailored to an individual shopper. This isn't concept territory, it's production ready at leading platforms. But here's the catch. The authors are explicit. Current evaluation methods aren't designed to catch filter bubbles, bias problems, or manipulation risks at scale. Standard metrics will make your system look fine when it isn't. That's the part I keep coming back to. You could deploy a generative recommendation system, hit your click metrics, and still be quietly showing half your catalog to nobody, reinforcing demographic bias in your personalization the whole time. And your dashboard wouldn't tell you. That is not a UX problem. That is a revenue leak you can't see. Plain English payoff. Generative AI can fix your cold start problem and create personalized visual experiences, but it will also introduce risks your current analytics won't catch. Okay, here's where this becomes commercially interesting. Money move. Build or sell an AI recommendation audit, specifically one that checks for filter bubble concentration, demographic bias in who gets shown what, and cold start failure rates on new product launches. The monograph spells out exactly what to look for. Most brands have none of this instrumentation in place. Action Step. Before your next product launch, pull your recommendation engine's performance data on the first 72 hours of a new SKU going live. If conversion is low and impressions are concentrated in a narrow user segment, you're seeing cold start failure in action. Evidence check. This is a literature review, not an original experiment. It synthesizes existing work. It doesn't produce new data. And the version we're working from is a preprint. Treat it as a field map, not a controlled finding. Radar verdict. Test this week. The cold start insight is real and the audit framework is immediately actionable. But read the original before you redesign your recommendation stack. This next one is more conceptual. But if you've ever sat in a room trying to explain where AI fits in your marketing org and felt like you were guessing, you're going to want this framework. Paper two, here's the business question. With AI marketing research growing this fast, how do you know which investments to prioritize and what are you probably missing? Researchers ran a bibliometric analysis of 438 peer-reviewed articles on AI and marketing. They used multidimensional scaling to map the intellectual structure of the field. Essentially, which ideas cluster together and which are isolated. So the key finding: AI marketing research has been growing at roughly 27% per year since 2020, nearly doubling every three years. And the field has crystallized into three distinct domains. Strategic and service AI, using AI for business operations and customer service. Consumer AI, understanding and influencing how consumers behave. And conversational AI, chatbots, virtual assistants, dialogue systems. Hmm, here's what I find genuinely useful about this. It gives you a mental model for diagnosing your own org. Most marketing teams are accidentally operating in one domain while thinking they're doing all three. And the researchers flag something important. Generative AI is officially its own fast-growing cluster within the field. But it also has the least developed research base. The risks, bias, misinformation, data privacy, are real and understudied. The research is running about a year behind the hype right now. But here's the catch. This paper maps what researchers are studying, not what actually works. Citation patterns are not business outcomes. A topic can dominate the literature and still underdeliver in practice. Also, the venue here is brand new, no track record. I can't tell you how rigorous the peer review actually was. And that actually bothers me. Because the framework is genuinely useful, three clean domains, a clear map of where the field is heading, and I want to recommend it with more confidence than the venue allows. Plain English Payoff. AI marketing research divides into three lanes: strategy and service, consumer behavior, and conversation. Generative AI is the newest, fastest growing, and least understood lane of all. Okay, here's the business hiding inside the research. Money move. Use the three-domain framework as a diagnostic. Run a workshop where you map every AI tool your team currently uses against the three domains. The gaps you find, the domain with no tools, no experiments, no budget, that's your next pitch to leadership. Action step. Before your next campaign review, sketch out which of the three domains your current AI budget is concentrated in. If it's all in one bucket, you have a roadmap problem, not a tool problem. Evidence check. This paper tells you what's being studied, not what's been proven to drive revenue. The framework comes from citation patterns, not from testing what actually moves the needle. And the venue is new enough that independent verification of peer review quality isn't possible yet. Radar verdict. The three-domain framework is worth using as an internal diagnostic right now. Just don't cite it as empirical proof that any specific AI approach works. Stay with me here because this next one looks like a niche case study. It isn't. The barriers it identifies are the same ones showing up in boardrooms everywhere. They're just easier to see when you strip away the polished tech company PR. Paper 3. Here's the business question. If the benefits of AI in marketing are so obvious, why aren't most companies actually deploying it well? Researchers ran a mixed method study in Zambia's insurance sector. 100 survey respondents, 25 in-depth interviews across selected insurance firms. They looked at how AI tools like chatbots were being used and whether more AI use correlated with better marketing outcomes. Here's what they found. 63% of firms reported using some form of AI, but only 31% had a customer-facing chatbot. So most of what gets called AI adoption is back-end and basic, not intelligent, not customer-facing, just slightly more digital than before. And the firms that did use more AI scored higher on marketing effectiveness. A moderate positive correlation, not proof of causation, but a real signal. Now here's where this gets useful beyond the Zambia context. The three barriers holding adoption back are skills gaps, poor data quality, and unclear regulatory rules. I know, I'm telling you, those are not emerging market problems. Those are universal problems. I hear them from marketing teams in London, New York, and Sydney every week. Not budget, not interest, skills, data, and governance. Every time. And here's the finding that genuinely surprised me. Full automation isn't what's working on the ground. Human agents plus some AI, not robots all the way down. That came through clearly in the interviews. Staff confirmed it themselves. But here's the catch. Small sample, non-random. One country, one sector. Correlation only. Self-reported AI usage, which tends to be overestimated, and the venue has low citation history. This is directional evidence, not a proof point. Plain English payoff. The real blockers to AI adoption and marketing aren't budget or enthusiasm. There's skill gaps, messy data, and fuzzy governance rules. And that's true whether you're in Lusaka or Los Angeles. Okay, here's the monetizable angle. Money move. If you're a consultant or vendor selling AI tools to any industry running behind on adoption, stop leading with the technology. Lead with a data readiness audit. That's the actual bottleneck. And it's a billable diagnostic service before any AI sale even starts. Action step. Before your next AI tool pitch, internal or external, answer three questions out loud. Who on the team can actually manage this tool? Is the underlying data clean enough? And what's our governance policy if the AI gets something wrong? If you can't answer all three, stop. Fix those first. Evidence check. 100 respondents, one country, one sector. Correlation only. Don't build strategy on this. Do use it to validate what you're probably already seeing anecdotally in your own organization. Radar verdict. Use cautiously. The barriers it identifies are real and recognizable. But the methodology limits how far you can push the findings. At first glance, these three papers look completely separate. A technical monograph on recommendation systems, a bibliometric field map, a small case study from Zambia. But together they show something that's been bothering me all morning. The gap between what AI can do in marketing and what's actually being deployed safely and smartly is enormous. And it's not closing on its own. Paper one says generative AI can fix your cold start problem and generate personalized visual experiences. Paper two says generative AI is the fastest growing and least understood cluster in all of marketing research. Paper three says the blockers aren't the technology. Put those together and you get one clear picture. Not more AI, better infrastructure for AI, not more tools, better intake before the tools go live. The contrast running through all three papers is this. Not capability, readiness. The capability is there. The cold start fix, the conversational commerce, the personalized visuals. The readiness isn't. And here's the tension I keep coming back to. The teams moving fastest on AI adoption are also the teams most likely to skip the audit step. The bias check, the data quality review, the governance policy. And the monograph is explicit. Current metrics won't catch those failures. Your dashboard will look green while the problem compounds quietly. That is not a technology problem. That is a discipline problem. And the research can't fix it, only you can. Here's the playbook from today. One, pull your recommendation engine's data on new product launches. Look specifically at cold start performance in the first 72 hours. If conversion is low and impressions are narrow, you have a system problem, not a product problem. Two, map your current AI tool stack against the three domains Strategic, consumer, conversational. The domain with nothing in it is your next strategic conversation. 3. Before your next AI deployment, answer three questions out loud. Who manages it? Is the data clean? And what's the governance policy if it fails? If you can't answer all three, stop. Fix those first. Evidence check on all of that. One paper today is a pre-print monograph with no original empirical data. One is from a brand new venue with an unverified peer review process. One has a sample of a hundred people in one country. Use them to decide what to test, not what to blindly believe. Links to all three papers are in the show notes. Read the originals before making major decisions. Want the human expert take? Join Dr. Eva Wolfe every Friday for the AI Marketing Radar Roundup, where she extracts no nonsense, money-making tips, practical strategy, and real business opportunities from the week's research. Subscribe on Apple Podcasts, Spotify, YouTube, and wherever you listen to podcasts. This is Evita for Big Plans Media, and I'll be back in the next radar brief.