Qualitative Research with Humans > Fake AI “Research”

Focus groups and interviews with real humans in your actual target audiences can give you value productized AI "customer research" never could.

AI tools can help a human researcher identify patterns in research findings, but the AI shouldn't replace the human research participants or the researcher. Despite what some well-funded AI product companies are telling you.

Why?

AI responds to prompts, using an established data set. If you ask an AI tool to tell you what Customer Segment A thinks about Product Feature 3, it will likely make up something plausible, and you will have an answer in--as I've seen advertised--"minutes, rather than days or weeks."

But will you have a reliable answer you can base business decisions on?

Probably not.

Because even if you get a plausible answer about what hypothetical customers would think of Product Feature 3, or the things YOU prioritize, you will not hear what your actual target customers prioritize, and what THEY want to talk about. Or why they think the way the do, and how their perceptions and emotions as humans shape their actions as consumers.

Case in point: I recently conducted a focus group about subscription software purchasing decisions. The client wanted to learn what potential customers thought about very specific features THE CLIENT thought should be the most important selling points.

As a researcher, I knew to broaden the conversation to yield real learning, beyond just what the client was focused on.

We ended up learning that the customers had different priorities from what the client expected. When they made purchase decisions, they paid little attention to the features the client assumed were big selling points.

Their rapport with each other in the focus group, and with me as a moderator, resulted in their sharing very personal stories, strategies, and frustrations that characterize their customer experiences and shape their buying decisions--all new insights to the client. Pure gold, from a research standpoint!

The client realized they needed to reassess their marketing strategy, and maybe even their product design.

These insights couldn't have come from an algorithm.

Only by listening to real humans can we be enlightened to the complex interplay of intellect, emotion, needs, and desires that shape (buying) decisions and (customer) experiences.

The value and the beauty of qualitative research with real humans is, we never know exactly what we'll learn. Some of the most useful insights are answers to questions we didn't know to ask.

P.S. Shortly after I wrote this post (originally on LinkedIn), Jakob Nielsen published an article about synthetic user research and research using "digital twins" based on data from research previously conducted with real individuals. His conclusions are, as I would expect, that synthetic users (based on aggregated data) aren't reliable, and digital twin research (based on the individual data of actual people) is fraught with ethical issues and hasn't been proven reliable. His article:
https://www.nngroup.com/articles/digital-twins/