What is AI and what does it mean for market research?
I was at the “StockTech” conference in Stockholm this week, and lots of the talk was about AI. Input from lots of gaming companies, Google, startups, and many other people in jeans and sneakers. Wasn’t my usual habitat but exciting to see just what’s around the corner in many business sectors. I thought I would just share some learnings that stuck with me, thinking about Research.
- AI isn’t that scary, in essence, we are asking computers to spot patterns and then by iteration (comparing early outcomes to the desired result) get more accurate in that task. To “learn” as it were. This is the computer building its own algorithm (solution) as it goes, rather than a human giving it an algorithm at the start. It’s why it’s sometimes called Machine Learning.
- One way to think about AI is not to wait and expect some big game changer to turn up but look at the mundane, repetitive elements of our work and get computers to learn to do those for us (so we can focus on what they can’t do which is interpret the unusual (not the usual).
- McKinsey predicts that 30% of most people’s jobs will be done by AI in a few year’s time. That’s much more encouraging than thinking about replacing humans (although that could mean fewer employees). Good news is that our jobs may be more interesting as a result.
- In research, a great deal of what we do is gather data then look for patterns. The “front end” of analysis is particularly open to machine learning – even more so if it’s a repeated task, e.g. customer satisfaction measurement.
- Big Data availability isn’t the only game-changer, its faster pattern recognition in any data.
- Finally, for now, Google made a powerful point – don’t think about trying to understand how AI works. Other people already do. Just ask yourself what you need help with. Then ask AI to do it. Focus on the outcome (need) not the “method”. A familiar challenge, but one market researchers should be good at.
As a company processing vast amounts of shopper survey data, it’s an area we are paying increasing attention to. Clients don’t want reports, they want to cut to the “so what” faster than ever before.