Is AI the future of market and social research?
Read almost any opinion piece on trends within the research industry and you’ll find Artificial Intelligence (AI) and machine learning in the top few, so why do I take the contrary view? Perhaps I should explain myself a bit more clearly.
Around 15 years ago, when I was a fresh-faced exec at Ipsos MORI, they ran a ‘Dragon’s Den’ competition for staff to give elevator pitches on innovations they thought the company should invest in. I pitched the idea of using electroencephalograms (EEGs or, more simply: brainwaves) to uncover emotions and reactions that might not otherwise be verbalised.
At the time, the technology was still too young and experimental, and it simply wasn’t available to end-users like us in a functional form. Fast forward 15 years and whilst EEG consumer tech is out there, it is by no means a mainstay of the research industry. But why?
A sledgehammer to crack a walnut
Ultimately, as exciting as EEG tech is, it’s overkill for the vast majority of real-world research questions where that level of data capture is superfluous. We can arrive at much the same insights in far less time and for far less money through good research design and skilled moderation or question writing. The edge that EEG research provides simply doesn’t apply to enough problems that clients have, and that’s why it remains out of the mainstream, providing extremely high amounts of data in mainly academic settings.
The trendy radiator spanner
So, given the blog title, let’s get back to AI. If EEG research is a sledgehammer, then AI is a radiator spanner: a tool that is very helpful…but only in the context of one type of job.
If you have a very large amount of customer data – perhaps online reviews or feedback emails – then the right AI tool will help you distil it down effectively into key themes and sentiments. There are still caveats around this though, in that you lose nuance, and no AI tool can solve the garbage-in, garbage-out problem (will some companies pay AI bots to leave “customer reviews” and then have another department analyse those reviews with a different AI tool? I’m betting yes). What’s more, I’d never advise basing future actions off only the input of those people who take the time to leave a review or post something on social media – it’s simply not a balanced sample.
The reality is that most companies have research needs far broader than the very specific application where AI is of benefit. So, while AI has successfully captured the zeitgeist, it’s far from being the panacea for all research problems.
The right tool for the right job
Here at WA Research, we pride ourselves on offering the whole range of research tools (including AI) but talking to you in terms of the problems they help solve and the outcomes they generate, rather than the mechanisms by which they work.
By keeping a laser focus on the problem and desired outcome, we’ll then select the right tool for the job at hand, regardless of how trendy (or not) that tool is. Feel free to get in touch to discuss the business problem into which you need insight.