Cutting through the AI noise in MRX with Dobrinka Vicentijevic

At this year’s Marketing Summit in Belgrade, attendees had the chance to hear Dobrinka Vicentijevic speak on how technology is transforming the shopper and retail landscape—and it’s safe to say, no one left the session indifferent. But then again, that comes as no surprise.

For those who may not know her, Dobrinka is a seasoned Shopper Insights Director with extensive experience in market research. She leads high-value projects worth between 3 and 4 million euros annually—with impressive client retention rates. Beyond her many industry awards, she’s more than just an MVP at EyeSee; she’s the go-to person when you need truly golden advice.

We took the opportunity to sit down with Dobrinka for a quick chat about the real impact of technology on market research. If you’re curious about spotting red flags in tech-enabled research solutions, you’re in the right place. Enjoy this short but insightful interview.

What risks do brands face by relying only on traditional research?

Dobrinka: First off, any research is better than no research at all. That said, traditional surveys often focus too much on rational, logical thinking, and consumers sometimes have a hard time putting their emotions or subconscious motivations into words. These methods can miss the more subtle responses that tools like biometric tracking, implicit testing, or behavioural analysis are much better at picking up. So, if businesses only stick to traditional research, they might end up making decisions that aren’t as accurate.

With so many tech-driven research solutions flooding the market, how can brands separate real value from empty promises? What red flags should they watch out for?

Dobrinka: There are many red flags these days, but if a provider lacks any of these characteristics, it should be a major red flag:

• Proven Track Record – If they possess zero case studies, testimonials, or real-world examples that show the technology can deliver real results, run.

• Transparency – Be wary of tools that are vague or unclear about how they actually work, especially how their models are trained or what kind of data they’re using. If they can’t explain it clearly, that’s a problem.

• Data Quality Assurance – Be cautious of solutions that promise large amounts of data but lack mechanisms to ensure its validity (issues with bots, outliers, noise in data, improper handling of sensitive information, data duplication, etc.).  

Hype vs. reality: What AI-powered research breakthroughs have truly impressed you, and what trends make you roll your eyes?

Dobrinka: To be fully honest, I love the use of AI predictive models that are properly trained on large sets of data (big data) and can make highly probable behavioral predictions. I also appreciate some types of generative AI, especially LLMs (like ChatGPT), as they are very useful and extremely helpful tools in analyzing and summarizing text content.

On the other hand, I wouldn’t say I roll my eyes, but I’ll admit, I’m pretty sceptical about the quality of today’s automated data collection & automated reporting and especially insights tools. Maybe one day they’ll be amazing, but right now, we’ve tested quite a few, and most of them, if not all, have “hallucinated”, jumped to the wrong conclusions, and given recommendations that just don’t hold up. Honestly, they remind me of self-driving cars: they sound great in theory and can do specific, isolated tasks really well, but throw them into a complex real-world situation, and things tend to go off the rails.

How much has technology really reshaped consumer behavior?

Dobrinka: Technology has changed everything, even our bodies. We’re so glued to our screens, especially our phones, that it’s actually starting to reshape us physically. Ever heard of “text neck”? It’s that pain you get from constantly looking down at your phone. And “texting thumb”? Yep, that’s a thing too.

But it’s not just our posture that's taking a hit; tech is also changing how we think, behave, and what we expect as consumers.

These days, we’re more (mis)informed than ever and way less patient. We want personalized deals, smooth experiences, fast checkouts, and of course, same-day delivery. We don’t want to wait, deal with irrelevant ads, or pay more than we think is fair. And if we’re not happy, we’ll definitely speak up about it.

We’re constantly bouncing between online and offline worlds, and honestly, it all kind of blends together now. That creates a big challenge for brands and retailers who are trying to keep up, tracking us across platforms and figuring out how their campaigns actually affect what we buy.

What’s one research myth you wish brands would finally stop believing?

Dobrinka: One mindset I’d love to see go away is the idea that “more data always means better insights.” Sure, having a lot of data can be helpful, but it’s really the quality of the data that makes the difference. Too often, brands fall into the trap of thinking that if they collect tons of responses through long surveys and from every possible source, their problems will magically be solved. But what actually works is focusing on smart, predictive data (especially the kind that’s implicit and behavioural) and making sure it directly ties back to the business questions they’re trying to answer.

Thank you, Dobrinka.  

Interested in learning more about market research tips and tricks, like the many levels to get the sample size just right? Go here.  

‍