AI in Arts Marketing: One Framework Paper, Three Big Claims

# Research Radar Brief — AI & Marketing | Episode radar-2026-05-21 **Date:** 2026-05-21 **Episode type:** Research Radar Brief **Papers screened:** 75 **Papers selected:** 1 **Theme:** AI and marketing > This is a first-pass research briefing, not a final academic review. Summaries are based on available abstracts and metadata. Read the original papers before making decisions. --- ## Papers Covered ### 1. Smart Cultural Curations: A Multidisciplinary Study on AI-Enhanced Marketing, Talent R -- This is a first-pass research briefing, not a final academic review. Read the original papers before making major marketing or business decisions. AI & Marketing Research Radar is produced by BigPlans Media. Subscribe wherever you listen to podcasts.

Transcript

SPEAKER_00 0:12

Here's the uncomfortable question this week. Two papers on AI and marketing, both published in 2026, neither one has a single original data point. So what are we actually learning?

SPEAKER_01 0:24

That's the thread. Both papers cover AI in marketing. Neither runs an experiment, surveys a customer, or measures a campaign.

SPEAKER_00 0:33

We screened 75 papers this week. Two made the radar.

SPEAKER_01 0:37

Quick caveat. This is a first-pass research briefing, not a final academic review. We'll tell you what the papers suggest, what they don't prove, and which ones deserve a deeper read.

SPEAKER_00 0:50

And today that second part matters more than usual. Okay, let's get into it.

SPEAKER_01 0:56

Paper one. What is AI actually doing inside marketing analytics right now? And does it help businesses make better decisions?

SPEAKER_00 1:07

Which sounds like a question that should have a clear answer in 2026.

SPEAKER_01 1:11

You'd think it's a literature review. The researcher pulled together existing studies and synthesized what they found.

SPEAKER_00 1:19

So no original data, no experiment, just here's what other people found.

SPEAKER_01 1:24

Exactly. And four things keep coming up: predictive analytics, machine learning to forecast who buys next, who churns, what demand looks like.

SPEAKER_00 1:35

Okay.

SPEAKER_01 1:36

Customer segmentation, sorting buyers by actual behavior, not demographics.

SPEAKER_00 1:41

Which is how you stop sending the same email to everyone on your list.

SPEAKER_01 1:46

Right. Then personalization. The Amazon and Netflix model show people things they're actually interested in.

SPEAKER_00 1:55

And the fourth.

SPEAKER_01 1:56

Automation, chatbots, triggered emails, ad bid adjustments, the stuff that runs without a human touching it.

SPEAKER_00 2:05

None of that is new, we've been talking about all four for years.

SPEAKER_01 2:08

We have. And that's the honest limitation here. This paper doesn't add new evidence, it summarizes what's already documented, and the venue isn't a top-tier marketing journal.

SPEAKER_00 2:21

So why does it make the radar?

SPEAKER_01 2:23

It's useful as a map. If you're on a team where people still argue about whether AI segmentation is worth the investment, this lays out the landscape in plain language.

SPEAKER_00 2:36

A framing tool, not a proof.

SPEAKER_01 2:38

Exactly.

SPEAKER_00 2:39

Okay, here's what I care about. The segmentation finding. Not because this paper proves it, it doesn't. But the pattern across the cited work points the same direction consistently.

SPEAKER_01 2:51

Behavior-based grouping beats demographic-based grouping for targeting. That thread runs through the underlying studies.

SPEAKER_00 2:59

That's the piece worth acting on. Plain English payoff. If you're still segmenting your email list by age or location, you're using the wrong signal. Switch to what people actually did. Try this by Friday. Go into your email tool and set up one behavior-based segment. People who clicked but didn't buy in the last 30 days. That's it. One segment. Send them something different than everyone else and watch what happens.

SPEAKER_01 3:26

Evidence check. No original data here. Every finding is borrowed from cited studies, and the review doesn't assess how rigorous those studies were. You're getting a narrative summary, not a quantitative synthesis. Treat it as direction, not proof.

SPEAKER_00 3:45

Radar verdict. Use cautiously. The four applications are real. Prediction, segmentation, personalization, automation. But this paper doesn't give you new evidence. It gives you a readable orientation.

SPEAKER_01 4:03

Paper 2. Same territory. AI marketing versus traditional marketing, but framed as a strategic question. Why are businesses shifting and what's driving it?

SPEAKER_00 4:14

And the answer is measurability.

SPEAKER_01 4:17

That's the core argument. Traditional marketing, TV, radio, print, is expensive and hard to trace. You spend on a billboard, you can't easily connect it to a sale. AI-driven channels track in real time. You know what worked.

SPEAKER_00 4:36

Which every digital marketer already knows, the billboard problem isn't new.

SPEAKER_01 4:41

It's not. And this is another conceptual article. No surveys, no experiments, no citations to primary research, we could verify.

SPEAKER_00 4:51

So what's actually worth taking from it?

SPEAKER_01 4:53

The hybrid framing. The argument that the best strategy isn't all AI or all digital. It's AI for targeting and automation. Traditional for brand storytelling and reaching audiences who aren't highly connected digitally.

SPEAKER_00 5:09

That I actually agree with. Right. The question isn't should we do AI marketing? The question is which part of our problem does AI actually solve? And it doesn't solve brand trust with a 65-year-old who reads a newspaper.

SPEAKER_01 5:34

The paper also covers automation as the efficiency case for AI. Less manual work, fewer mistakes, always on.

SPEAKER_00 5:42

Which is real. But again, this paper isn't measuring it, it's describing it. And that's where I'd push back on leaning too hard on either of these papers. They're readable, they're coherent, but coherent isn't the same as evidenced.

SPEAKER_01 5:59

That's fair. The logic holds. The data isn't here.

SPEAKER_00 6:04

Plain English payoff AI wins on measurement and targeting. Traditional wins on trust and reach with offline audiences. Use both and know which job each one is doing. Try this by Friday. Look at your current channel mix. Pick one traditional spend item. Print, radio, a sponsorship. Ask honestly, is there a digital equivalent that does the same awareness job with better measurement? If yes, pilot the switch. If your audience is offline, leave it alone.

SPEAKER_01 6:33

Evidence check. No data, no visible citations to primary research. The claims about AI's cost effectiveness and ROI advantages are asserted, not measured. Don't use this to justify a budget reallocation to a CFO. The logic is sound. The evidence isn't here.

SPEAKER_00 6:53

Radar verdict. Two papers, same topic, both 2026, and neither one ran a study. That's not an accident. And in that gap, you get review articles and conceptual frameworks that describe what people are already doing.

SPEAKER_01 7:37

Which isn't useless, but it means you can't look to these papers for proof. You look to them for vocabulary and orientation. The tension between the two is interesting. Paper one focuses on AI tools as a way to improve what you already do. Better segmentation, smarter targeting. Paper two asks a bigger question. Should you still be doing traditional marketing at all?

SPEAKER_00 8:03

And the answer from paper two is yes, but differently. Don't replace traditional hybridize. Here's the tension I'd name though. Both papers treat AI adoption as a direction without acknowledging that most small businesses don't have clean data, don't have ML engineers, and can't afford the tools that make AI segmentation actually work.

SPEAKER_01 8:28

That's a fair push. The landscape these papers describe assumes a certain infrastructure.

SPEAKER_00 8:34

So the real question for most teams isn't should we do AI marketing? It's what's the smallest version of this we can test with what we have.

SPEAKER_01 8:44

And both papers, for all their limitations, point toward a few places worth starting.

SPEAKER_00 8:51

Segmentation and automation. Those are the two that show up in both papers and have the most accessible entry points. Here's the playbook from this week. One, set up one behavior-based email segment. People who clicked but didn't convert. Send them something targeted. Measure open rate, click rate, conversion. That's your baseline experiment. Two, audit your channel mix. For each traditional spend item, ask whether there's a measurable digital equivalent that reaches the same audience. If yes, pilot it. If your audience isn't online, protect the traditional spend.

SPEAKER_01 9:26

Evidence check on the whole week. Both papers are conceptual reviews with no original data. Neither is in a top-tier venue. Use them to decide what to test, not what to believe.

SPEAKER_00 9:39

The actions are still worth doing. The papers just aren't the proof. Your own results are the proof.

SPEAKER_01 9:46

Links to both papers are in the show notes. Read the originals before making major decisions.

SPEAKER_00 9:53

See you Thursday. And if you've already been running behavior based segmentation and you've got results, I want to hear what you found. Send it our way.