AI is the new crystal ball in research — Or is it just guessing?

Hey there, 

Sharekh here! 👋

Welcome back to The Research Mag! 

I’ve been thinking a lot about our last issue on synthetic data, especially after a coffee chat with a fellow founder here in the Bay Area. We talked about how some people calling generative AI to be the biggest thing to have ever happened in technology since the internet, and I’d agree (on some days). Then we talked about how literally every company now has been releasing an “AI” upgrade to whatever products and services they’re selling.

Which makes sense. AI can make things way, way easier on their customer’s end to do certain tasks, automate workflows, get the software to (literally) do their work for them. And it’s a great addition to any product! Well, mostly. But here’s the thing: AI (including synthetic data) isn’t without its issues. In fact, some of the same risks and limitations we covered in synthetic data are creeping into other AI applications.

But I’m not a pessimist, so let’s start with the good stuff.

Chances are, you already know the basics of how AI is making a huge difference in the market research industry, but let’s recap:

• Sentiment Analysis: AI-powered tools are combing through customer feedback in real-time, analyzing not just what people say, but how they feel. It’s like getting instant access to your customers’ emotions.

• Predictive Analytics: Brands are leveraging AI to predict future customer behaviors based on past data, helping them stay ahead of trends.

• Customer Segmentation: AI is no longer just slicing people by demographics—it’s diving deep into behavioral data to help brands personalize experiences like never before.

These are impressive, but they’re only scratching the surface. Now let’s jump into the fun stuff. Something that’d make you go, “Oh, that’s actually really cool!”

• AI-Generated Personas—Real enough to fool you

  Market research often relies on building customer personas based on survey data. But what if those personas could react like real people? AI is now creating interactive digital personas that can be used to simulate how a target audience might respond to current events, product changes, or marketing campaigns.

  Example: A brand might train AI on past customer data to build digital personas that reflect their audience’s responses to hypothetical situations—without needing to conduct a survey.

• AI-Enhanced Perception Mapping—Finding hidden insights faster than ever

  Traditionally, perceptual mapping (a way to visualize how consumers perceive different brands) was a time-consuming, resource-heavy process. But AI is changing that. By analyzing vast amounts of data in real-time, AI can create perception maps that help brands understand how their products are viewed compared to competitors—without needing a large survey sample.

  Fun fact: Researchers used synthetic data to generate a perceptual map for the US automotive industry. The results were strikingly similar to insights gained from a traditional survey of 530 real consumers, but were generated in a fraction of the time.

• Facial Coding with AI—Reading your face like a book

  This one is wild. AI-powered facial coding analyzes customers’ facial expressions during product tests or ad viewings to gauge their emotional reactions—without them saying a word. This goes beyond traditional surveys by capturing raw, unfiltered emotions. It’s being used by everyone from movie studios to consumer goods companies.

  Fun Fact: Disney used this tech during movie screenings, collecting 16 million data points on audience reactions, which helped them tweak scenes based on real-time facial feedback.

• AI for Predicting Trends—Is It a Fad or Forever?

  Ever wondered if a viral trend is here to stay or just a flash in the pan? AI is analyzing social media trends, purchase data, and online conversations to predict whether a trend will last or fade away. This allows brands to decide whether to go all-in on a trend or stay cautious.

  Fun Fact: Some brands are using AI to predict how long TikTok challenges will stay relevant, helping them figure out whether to invest in a short-term campaign 

These sound very exciting, and we too at CleverX are pushing some AI features very soon. But as much as I love what AI can do for market research, it comes with significant risks.

Firstly, there could be bias in data: And this one is very serious. We’ve talked about this before, but it’s worth repeating. If AI is trained on biased data, it will perpetuate those biases. In market research, this can lead to skewed insights that misrepresent your audience.

Privacy Concerns: With AI needing large datasets, privacy is a huge concern. Sure, synthetic data helps, but how much personal information are you comfortable using in your AI models?

The Human Element: AI gives you trends, predictions, and insights, but it doesn’t replace the why. For that, you still need human analysis to provide context and interpret the findings.

So, where do we go from here? AI is a fantastic tool—no one’s denying that. But if there’s one thing we’ve learned over these last few issues, it’s that we shouldn’t rely on it too much. So while AI will help you prepare for most of what’s coming, don’t throw away your human intuition just yet. We’re still going to need that. The human element.

That’s all for now! I’ll come back to you soon with more interesting stuff happening in research!

If you liked reading this issue, please leave us your feedback, as well as ideas as to what you’d like to know more about!

You can also reach out to me directly here.

Best,
Sharekh,
The Research Mag
Founder @CleverX
Connect with me on X and LinkedIn