Searching for the Why in AI

Editor’s Note: This post is part of our Big Ideas series, a column highlighting the innovative thinking and thought leadership at IIeX events around the world. Jason Smikle will be speaking at IIeX North America 2019 in Austin, TX. If you liked this article, you’ll LOVE IIeX North America. Click here to learn more. 

---

A lot of companies are asking right now, how can we use AI to lower costs and make better business decisions. It’s no secret that AI is great, but only at a few key things. It’s great at predicting. It’s great at helping us with repetitive tasks and uncovering patterns we didn’t even know to look for. And it will only get better every year.

But what AI and the engineers building machine learning models struggle with are turning their findings into stories and emotion that can create action and buy-in internally.

A lot of times these models cannot detect the real and raw emotion tied to a product or brand that is getting in the way of a sale. Unless these human emotions are being shared and stored in a data set in real time, it’s going to be really hard to predict why your sales are down. And there are attempts to do this of course, but from what I’ve seen we’re not there yet. It’s certainly not at the level where these models are mainstream and easily accessible for anyone to use.

Of course, there are exceptions. Social media sentiment will tell you why Starbucks sales are down when a barista calls the police on a customer for no good reason, or when a sneaker malfunctions on one of the best athletes in the world. And there are correlations many times between the number of people talking about your product and product sales. If you’re a new brand and people are beginning to discover, love and share your brand with the world, you can predict an upward trend of growth as new buyers flood the market.

But for most product designers, nobody is tweeting why they choose one deodorant over another, and the average person trying to make it in their world isn’t leaving product reviews about that new garbage can they just bought for their kitchen.

So our public data sets are full of extremes, and they don’t catch all segments, which gets in the way of our overall understanding of consumers.

That’s why research is still really important. Talking to people about the small mundane things that take up a small part of their attention and deep diving to understand the drivers of a consumer is essential to crafting this story.

One of the best things that AI will do for business leaders will be to uncover patterns you hadn’t even considered. But once a pattern is discovered, you’ll want to understand the reasoning behind the pattern. That’s where research comes in.

Research has to answer the question of why so that you can be confident in your decision making.

The question of “Why?!”, which is defined as “for what reason or purpose” is what drives us as humans. Discovery and growth, whether it be personal or in a professional setting, is at the core of who we are as a people. Understanding the motives behind patterns, emotions, and decisions to help us grow is what drives the business forward. Acting properly on this information will increase your sales.

And this is something AI simply isn’t great at. At least for the time being. So for now, we still need research.

Good research tools can uncover the why behind the patterns AI discovers. And once you are discovering the reasons behind the patterns, you’ll need to shape them into a story, whether it be on PowerPoint, a video or even a perfectly crafted word of mouth story that goes viral internally. These are the stories that will drive executive action and decision making forward.

So how can research work at the speed of AI? It doesn’t do anyone any good if your “why story” is delivered a month after the discovery. The story needs to be discovered and crafted at the speed of AI.

The biggest challenge now for researchers is to have the tools that will help them gather insight into why at the same speed that AI moves. This is what our clients have demanded, and it is what our company will continue to build. Being able to gather real qualitative data that gets to the heart of an issue in real-time is how researchers can be complementary to what AI can do.

AI is trendy right now, and its capabilities are only going to get better every year. I still believe we won’t see the full power of AI for another 10 years (think of a webpage in 1999 compared to 2019).

But for now, it’s important for executives that are integrating AI to understand its power and limitations. AI paired with a real-time understanding of the heart of the consumer is the holy grail to successful product design. When research and AI work together, business wins.