Five Takeaways from TMRE 2026—The State of Market Research (and where the industry is headed)
By Lucas Longacre
Lucas Longacre’s five takeaways from TMRE 2026 point to a market research industry being transformed by AI—but not one where humans are becoming less important. Faster tools, always-on insight systems, synthetic data, and AI-generated analysis are raising the stakes for data quality, validation, judgment, and empathy, making the researcher’s ability to distinguish what is true and meaningful from what is merely fast or plausible more valuable than ever.
The Market Research Event (TMRE) 2026 in Denver, Colorado, was overwhelming. For the first time, it was three events combined into one: The Market Research Event, Content Marketing World, and Lions all rolled into one super-event that took place over three days at the Denver Convention center. I barely had time to catch my breath between keynotes, fireside chats, panel discussions, and interviews for Signal & Noise.
On the plus side, I got in over 15,000 steps each day running from one side of the Denver convention center to the other—which is lovely, comfortable, and very professionally staffed by kindhearted workers. With all of the chaos, I left my personal laptop behind on a desk and, with the assistance of the staff, was able to track it down to security and retrieve it without missing a panel. Very grateful for that.
To sum up three days, dozens of sessions, and interviews with keynote speakers, I walked away with a lot of new insights and opinions. Here are the Five Takeaways that really stuck with me after the conference wrapped.
AI Knows What's Been Said. It Doesn't Know What's True.
This was the line from Pamela Forbus and Stephen Springfield at the Evidence Based Marketing Alliance session, and it's the one that all users of AI—both experienced and newbies—need to keep in the forefront when interacting with AI. Pamela emphasized this by asking an AI model (that won’t be named) for a brand growth strategy and told her to target her best customers. Sounds logical, right? That’s what the internet says on the subject. That's what most of the content online reflects.
But 75 years of marketing science—25 scientific journals and 100 proven hypotheses—proves it is a completely backwards assumption. The machine learned from our published works. We take the findings as truth. Then plug it back into the machine. That's model collapse. Garbage in, garbage out—at a massive scale.
Stephen Springfield has a name for what that produces: the Great Marketing Hallucination. An industry confidently acting on a consensus that was never true, now supercharged and self-reinforcing. Don’t believe everything you’ve read on the internet. Isn’t that advice that we all followed before AI?
AI needs to be trained on this assumption. Or at least, us practitioners of AI need to be a lot more critical of AI results. The session closed with a Yuval Noah Harari quote that goes hard: “If you flood the world with information, the truth will sink to the bottom.” That's not a technology problem. That's a leadership problem. And it’s our collective responsibility to seek truth.
“The machine learned from us. We're starting to learn from the machine. Soon the machine will learn from the machine.”
Speed Is the Accelerant, Not the Solution.
I sat down with Dr. Thomas Ramsøy, CEO of Neurons, after his session. For me, it was one of the best talks of the entire conference. He opened with a number that shocked me, but didn’t surprise me: ad and brand memory has dropped from 30 percent to 5 percent.
No, I didn’t make a typo. Five percent! His point wasn't that AI caused this. AI just supercharged it. Ninety-three percent of marketers are using AI to generate content faster. Ninety percent to speed up production. But if the creative isn't working in the first place, moving faster just gets you to wrong quicker and at a much larger scale.
What's actually missing, according to Ramsøy, is judgment; we must evaluate what we’ve made before we ship it. And he has a solution. He’s built a predictive model at Neurons (trained on 20 years of neuroscience data) that can score attention, emotional response, and memory likelihood before a dollar of media spend is committed. The model can score your campaign and generative AI can create multiple improved versions to bring back to designers, copywriters and developers before a campaign is launched. The real ROI isn't faster execution. It's fewer catastrophic launches.
“If the creative isn't working in the first place, moving faster just gets you to wrong quicker and at scale.”
Real Human Data Is Under Attack—and That's Everyone's Problem.
I caught up with Christopher Barnes, President of Escalent, between sessions and asked him to expand on what he'd shared in his fraud session. What he described was more serious than I expected. Societal fraud is up 30% year over year. AI avatars are already infiltrating research panels—not as a theoretical risk, but as documented reality.
Escalent has abandoned social media recruiting entirely because open links have become honeypots that are scraped and exploited within minutes of posting. They run 31 different quality checks across six ensembles before they trust a data set. And they’re still not certain they've caught everything. Barnes made the case that 75 real, carefully validated respondents are worth more than 120 where 20 are fraudulent and 25 are professional survey-takers on autopilot. Total survey error matters more than sample size, but procurement teams are still buying on N.
A recurring topic of discussion was synthetic data and Christoper had a lot of opinions. He pointed out that the moment you introduce synthetic augmentation, your standard statistical validity tests—T-tests, Z-tests, confidence intervals—are essentially meaningless. If the goal is to use AI to get more insight from fewer real humans, the quality of that human data has to be better than ever. Not worse. I loved his advice for the industry: share your problems openly.
He'll have a fraud conversation with a direct competitor because a fraudster who sharpens their tools on another firm will come for his next. He also encouraged me to write an article about industry best practices for AI Fraud and I accepted his challenge. If you want to be included in the article, reach out.
“The best fuel for AI is really good human data. Validate, validate, validate—and don't let scanning it and feeling right count as validation.”
The Shift from Episodic to Always-On Is Happening Now.
This theme showed up in at least four different sessions across the three days, framed slightly differently each time. Romani Patel from Microsoft put it most clearly: research used to be project-focused. It's becoming like electricity—infrastructure, always running, always available. The customer voice isn't a deliverable anymore. It's a resource.
Jordan Welling from Tecovas gave the most concrete example: when they went to build a new brand direction, they didn't commission a new study. They pulled 500 hours of recorded customer interviews already sitting in their system. The insight was already there. They just had to surface it.
Lev Mazin from aytm framed the original failure mode: a brand asks a question on Tuesday and gets the answer four months later.
By then the decision has already been made. The original motivation behind always-on intelligence wasn't to replace researchers or cut costs — it was to be faster than the decision cycle. What struck me across all three: the technology for this exists. The constraint is organizational. Most enterprises still run insight generation like a project — brief, field, analyze, deliver. The companies that crack the infrastructure model will have a durable competitive advantage over the ones waiting for the next RFP.
The Human Is Not Optional.
This was the through-line I didn't expect to be the defining theme of the conference, but it was everywhere. P&G launched a product that performed well in the US and got completely rejected in
the UK—because British people soak their dishes. AI surfaced the pattern. It took a human standing at a sink, watching, to understand why. That insight led to a different product positioning that drove double-digit growth.
Marc Pritchard's framing: AI gives you the “what” at incredible speed. Human empathy gets you the “why.” The why is where the brand idea lives. The healthcare consumerism panel built the same case from a different direction: stop saying “patients.” A person living with narcolepsy is not the same as a narcolepsy patient.
Ngozi Kamalu at Jazz Pharmaceuticals ran a study where participants expressed their disease experience through gifs and memes because that's how people actually communicate. The insight was more honest and more usable than anything a traditional survey would have produced.
Pamela Forbus at Mondelez delivered of the best lines of the conference: “If you haven't
been wrong in public, you probably haven't said anything of consequence.” Her challenge to every researcher in the room: AI just gave you an extra hour. What are you going to do with it? Her answer: use it to engage the questions that actually matter, the risks your business is currently ignoring.
And the Future of Insights Throwdown ended with what might be the most honest thing said all week: “You have to be human to understand humans, and your gut is you acting as a human.” Even the data purists in the room couldn't fully argue with that.
“You have to be human to understand humans, and your gut is you acting as a human.”
The Through-Line
The market research industry has more data, more AI tools, and more processing speed
than at any point in its history. And the most critical element across every session, every
panel, and every conversation I had in Denver was the thing that's hardest to automate:
human judgment, human empathy, and the ability to look at what the data is showing
you and know which signal is actually worth acting on.
That's a positive view of AI. The tools are genuinely powerful. The researchers who learn
how to use them well—with rigor, with governance, and with the discipline to validate
before they act—are going to be incredibly more productive.
To quote Aneesh Dhawan from knit.”It is the golden age of the human researcher. I can
get behind that sentiment. The question is whether the industry will do the work to earn
it."
· · ·
Stay tuned for full interviews with Dr. Thomas Ramsøy (Neurons), Christopher Barnes (Escalent),
Tracy Adams, Keith Smith, and Lev
------------------------------------------------
Lucas is a product and AI leader with more than fifteen years of experience across software, media technology, and emerging platforms. He is Head of Product at Inlightened, where he builds responsible AI systems that translate complex data into clear, usable products, modernizing expert intelligence and search in an increasingly AI-driven world.
Comments