AI & Marketing Research with Dr. Eva Wolf
Evidence-led briefings that translate peer-reviewed studies and important preprints on artificial intelligence, generative AI, marketing, advertising, consumer behavior, and business strategy into practical insight. Dr. Eva Wolf explains what the evidence actually says, what deserves a deeper read, and what marketers, consultants, educators, and business leaders can do next.
AI & Marketing Research with Dr. Eva Wolf
AI Marketing Research: Gen Z Trust, Ad Forecasting & LLM Ads
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 Wolfe, 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. What if the brands that win the next decade of advertising aren't the ones who spent more, but the ones who figured out that transparency isn't a compliance checkbox, it's a conversion lever. Today's papers point to one underlying shift. AI is changing how brands build trust, how they forecast revenue, and how they'll eventually buy ad space inside conversations. We screened 388 papers. Three cleared the full text bar and made the radar today. Quick caveat: this is a first-pass research briefing, not a final academic review. 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 brand uses AI to personalize ads, emails, or recommendations, does telling Gen Z how it works actually make them more likely to buy? Jay Prakash surveyed 100 Gen Z respondents in Kerala, India. The focus was transparency in AI-driven marketing. How clearly brands explain what data they collect, why they personalize, how decisions get made. And the finding was stark. Gen Z consumers who saw clear AI explanations reported significantly higher trust. Transparency alone accounted for roughly 55% of the variation in how much they trusted a brand. Let me translate that. It's not that Gen Z hates AI. It's that they hate being manipulated without knowing it. Explain the machine and they reward you. Hide it and they walk. That is not a UX problem. That is a revenue problem. But here's the catch. This is a hundred people, convenience sampled, one district in India, self-reported, published in a journal with essentially no citation record. I'm not dismissing the finding. Not because they're idealists, but because transparency signals you're not hiding something. Okay, here's where this becomes commercially interesting. Money move. Offer a Gen Z trust audit as an agency service. Review a brand's AI-driven touch points. Rewrite the chatbot intros, the recommendation footers, the ad disclosures in plain language. Then measure the impact on engagement and repeat purchase. Not a rebuild, a layer. Action step. Pull your current AI disclosure copy right now. Whatever's in your email footer, your app onboarding, your cookie banner. Read it like a skeptical 22-year-old. If it sounds like a terms of service document, it is not doing the trust work this research says it should. Evidence check. n equals 100. Convenience sample, one region, self-report only, correlational design, directionally useful, not a mandate. Radar verdict. Use cautiously. The signal is real, but the sample is too small and too narrow to generalize. Treat this as a prompt to test, not a proof to act on. This next one looks technical. It is technical, but it's hiding the answer to a question every performance marketer has every single quarter. What would happen to my sales if I changed my ad budget? Your current tools probably can't answer that. This paper is about fixing that. Paper two. Here's the business question. Can AI tell you what would happen to your sales if you changed your ad spend, not just what happened based on what you already did? Researchers at Alibaba built a system called CEDAR, two-stage machine learning framework. Stage one is a transformer model that learns how merchant decisions, ad budgets, discounts, directly cause sales to move. Stage two is an LLM that reads messy external event descriptions, holidays, platform promotions, viral trends, and adjusts the simulation for outside shocks. They tested it on 32 million product trajectories on Alibaba's B2B marketplace. In offline testing and live production, Cedar beat every standard forecasting baseline on simulation accuracy and delivered real improvements in how merchants planned their ad budgets. Let me translate that. Most forecasting tools tell you what will happen based on patterns from the past. CEDAR tells you what would happen if you made a different decision than the one you made before. That's a fundamentally different tool. And here's the piece most teams miss. Separating what your spending does from what the market is doing anyway. Black Friday lifts everyone. Your ad spend lifts you specifically. Mixing those two signals together is exactly why your post-campaign attribution feels unreliable. But here's the catch. This was built and tested entirely inside Alibaba's B2B infrastructure. 32 million product trajectories. Rich historical data most brands simply don't have. And it's a preprint, accepted to KDD 2026, but not yet through full peer review. Genuinely surprising. Not that the model works, but that no one had formally built this separation into a production system before, given how obvious the problem is once you name it. Plain English payoff. Your forecasting tool tells you what will happen. CEDAR is a blueprint for building a tool that tells you what would happen if you spent differently, which is the question you actually need answered before you commit a budget. Here's the business hiding inside the research. Money move. If you're an agency or a Martech builder, there's a counterfactual budget planning service waiting to be productized here. Help clients stress test their media plans before major retail events by separating what the event does from what their spending does. Sell it as scenario simulation, not just forecasting. Action step. Before your next major campaign, run a simple mental audit. Can your current analytics tools answer what would happen if we spent 30% less on paid search next month? If the answer is no, you're forecasting, you're not simulating. That gap is worth surfacing to your data team or vendor before your next campaign review. Evidence check. Preprint, not yet peer-reviewed. Alibaba scale data that most brands can't replicate. The online production experiment details are thin in the available text. Good concept, narrow applicability right now. Radar verdict. In concept, not the model itself. The idea of separating spend-driven lift from event-driven lift in your own planning. That analytical habit is free, and this paper makes a strong case it's the right one. Okay, stay with me here. This last one is the most speculative of the three, and also the one I think you'll be turning over in your head longest. It's about where advertising goes when the interface is a conversation. Paper three. Here's the business question. When people stop Googling and start chatting with AI assistants, how does advertising actually work? And can it be made to work fairly? Kim, one researcher, independent, working in mechanism design, proposes a new auction framework built specifically for LLM chat interfaces. Instead of matching ads to keywords, the system measures how close the current conversation is to an advertiser's description of their ideal customer in embedding space, geometrically. And the claim, verified using computer-checked formal proofs, is that this system is truthful. Advertisers have no incentive to lie about what a customer is worth to them. Bidding honestly is always the dominant strategy. I'm telling you, that is not true of today's keyword auctions. The paper also shows that traditional keyword auctions are just a broken, oversimplified version of this more general structure, which is a sharp way of saying the whole thing was built wrong from the start. This is purely theoretical. No user study, no revenue comparison, no empirical data of any kind. Independent researcher, no institutional affiliation, published on Zenodo. The formal proofs haven't been independently validated. This is the most speculative paper in today's radar. And yet, Perplexity is already running ads. Chat GPT is building toward monetization. The question this paper answers, how should ads work inside a conversation, is going to get answered by someone. This is one serious attempt at the math. Plain English payoff. The next ad surface might let you buy space in a conversation by describing your ideal customer in plain language, not by picking keywords. And if the math in this paper holds up, that system could be more honest and harder to game than anything running today. Okay, here's the monetizable angle. Money move. If you're building an AI product with a conversational interface, this paper is a blueprint for monetizing it with ads without modifying your model. A plug-in revenue layer for AI startups who need income without compromising quality. Worth bookmarking for your next product roadmap conversation. Action step. Not a persona deck, not a keyword list, a sentence. Our customer is someone mid-conversation about renovating their first home. Practice that framing now. When LLM ad platforms launch, that's the targeting input. Evidence check, zero empirical data. Unreviewed preprint from an independent researcher. Formal proofs haven't been independently validated. This is a conceptual contribution, not a validated system. Radar verdict, watch list. The idea is directionally important and the math is serious. But there's no evidence it works in practice yet. Monitor as LLM ad platforms develop. At first glance, these three papers look completely separate. Trust surveys, demand forecasting architecture, auction theory. But together they show a single pattern. AI is forcing marketers to be honest about what they don't know. And the brands that get ahead will be the ones who build systems that are transparent by design, not by accident. Paper one says Gen Z trusts brands that explain their AI. Paper two says the biggest forecasting mistake is conflating what your decisions do with what the market does anyway. Paper three says the next ad auction can be mathematically truthful in a way keyword auctions never were. Not more AI, more honest AI. That's the through line. Not more personalization, more legibility, not more targeting, more trustworthy architecture. Here's the tension I keep coming back to. The transparency finding in paper one is weakly evidenced but directionally credible. The forecasting architecture in paper two is powerfully evidenced, but currently inaccessible to most teams. And the auction theory in paper three is intellectually compelling but has zero real-world validation. So the pattern is clear. The tools aren't all here yet. Which means the window for being early is actually open right now. Here's the playbook from today. 1. Audit your AI disclosure copy. Read it like a skeptical 22-year-old. If it reads like legal boilerplate, rewrite it in plain language and test whether clearer disclosures move trust or engagement metrics before your next campaign review. 2. In your next budget planning cycle, separate the question, what does our spending do? from what does the market event do? If your current tools can't answer those independently, name that gap explicitly to your team or vendor. 3. Write a two-sentence meaning-based description of your ideal customer. Not a keyword list, a sentence. File it somewhere. When LLM ad platforms launch targeting by conversational context, you'll want to have already thought about this. Evidence check on all of that. Paper one is a small regional self-report study. Use it to prompt a test, not to mandate a rebrand. Paper two is a preprint from a platform with infrastructure most teams don't have. Paper three has no empirical validation at all. 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 Avita for Big Plans Media, and I'll be back in the next radar brief.