How AI Can Help Companies Understand What Customers Really Want

The Philadelphia-based Barnes Foundation recently undertook a project using AI to categorize its vast art collection, with unexpected results. The foundation found that machines could make surprising judgments about art that the human eye would not typically make.

For example, when it asked its computers to locate images of "cherubs" in the collection, the results brought up many pictures which did not have the angels in them -- but contained features that the computer perceived as "cherub-like."

The machines saw the artwork differently than humans did, creating undiscovered connections that scholars had never considered before.

This exercise, and many others like it, has started to provide insight into how we may view art on a subconscious level that our rational human minds will have missed. 

“It’s this magic that we want to keep,” said the project’s leader, Shelley Bernstein. 

Ironically it’s the ability of AI to see more as a machine than a human that will start to help companies examine consumers’ subconscious interactions with increasingly vast amounts of content across a growing number of touchpoints. This is happening with the intent of finding new patterns and subtext that the conscious human mind may not notice.



Why is this important? Contrary to popular belief, most people buy products emotionally, not rationally. Research from Harvard Business School professor Gerald Zaltman found that we make about 95% of our buying decisions subconsciously.  To ensure consumers buy your products, it’s essential that the product you’re selling connects emotionally with the viewer. 

And to this point, machines are not just analyzing content, but starting to read the emotions of people viewing your content. This feature is in its early days but MIT professor Erik Brynjolfsson says machines can even notice certain subconscious micro-expressions on people’s faces which may be missed by our rational self. 

This "emotion AI" is becoming a more vital tool for marketers. The Boston firm Affectiva conducts advertising research by getting people to watch digital content and measure their facial expressions and emotions while viewing it. The data it collects enables them to predict a consumer’s likely behaviour. 

AI is still in its early stages for companies to truly understand what everyone wants. Nevertheless, it’s only getting better at reading our words, judging our pictures and analyzing our faces. Moreover, as AI increasingly comprehends the subtext of people’s emotions, the stronger the tools that brands have to attract new customers to buy their products.  

And it’s not just the development of AI that will help contribute to this new approach. The rise in popularity of API-first architected experience platforms, such as platforms like Contentstack, where content is decoupled from presentation layer, will enable companies to employ significantly more progressive approaches when determining how their brand is influencing its audience. 

We want to believe that we are thinking rational people who, on occasion, tangle with emotion then flick it out of the way and go back to thinking.

That’s not the truth. Really, we are emotional beings who occasionally think. If AI can help companies understand these emotions better, and the technology is appropriately regulated and ethical concerns over its use are sufficiently addressed, this can empower companies to better deliver the products and services their consumers truly want.

2 comments about "How AI Can Help Companies Understand What Customers Really Want".
Check to receive email when comments are posted.
  1. John Grono from GAP Research, December 18, 2019 at 3:39 p.m.

    And of course the algorithm that 'brought up many pictures which did not have the angels in them -- but contained features that the computer perceived as "cherub-like."' may have just got it wrong and because the human brain being much more pliable used its creative elasticity to alos perceive tham as cherubs.

  2. Ginger Cookie from Consultant, December 23, 2019 at 1:58 p.m.

    So true John, and to further your point, realistically, that wanna be "cherub-like" picture may not net out to the visitor/viewer staying more focuse, interested, curious to seek further in the saturated content marketing mill...thus, the algorithm is off...and its wasting time/content that humans/individual have zero interest on.

Next story loading loading..

Discover Our Publications