Skip
Current Issue
This Month's Print Issue

Follow Fast Company

We’ll come to you.

3 minute read

How Artificial Intelligence Is Helping Brands Create Ads Just For You

IBM and the Havas Group have launched Havas Cognitive to give marketers Watson's cognitive computing power.

How Artificial Intelligence Is Helping Brands Create Ads Just For You
[Photo: Flickr user Robert Couse-Baker]

Consider all the information about your likes, dislikes, wants, needs, and behavior that are out in the world. Whether it's Facebook, Twitter, apps, games, or any other digital interaction, it amounts to a giant pile of data just about you. So, in this world of big data and targeted advertising, why do you still see diaper ads when you don't have a baby?

It's a silly example, sure, but the point is, as much as brands know about us, there are still far too many irrelevant marketing messages poisoning our eyeballs. But a new partnership between IBM and Havas Group aims to change that. The two companies have launched a new division called Havas Cognitive, which uses the cognitive computing power of IBM Watson to help brands develop marketing campaigns and products better tailored to individual consumers.

These organizations have worked together for more than 20 years, but the idea for the new venture came about last year when a group of Havas developers participated in the first-ever IBM Watson Hackathon. The team finished in second place with an app called NYC School Finder, that used the IBM Watson Personality Insights to analyze students’ personalities through writing samples to finds schools with similar "personality" traits.

"Our potential brand applications really came out of this, where it took a writing sample from your kid, analyze it for a personality construct, and then find the best school match, to connect kids to the high school best-suited to their personality," says Havas global head of marketing innovation Jason Jercinovic. "So we thought, hey this might be a good way to help our clients better know their customers on a more individual level."

Marketers are spending increasingly more on social listening tools, e-commerce data, and more to build smarter communication mechanisms, but often lack the technology to most efficiently use it.

IBM Watson Ecosystem director Jodie Sasse says Watson technology and cognitive computing is capable of making sense of all this data, quickly and continually. "It's a system that understands natural language, it understands the nuances of what we say, why noses run and feet smell, that kind of thing," says Sasse. "It understands idiosyncrasies of how we speak, and it also reasons to sort through massive amounts of information, and come to hypothesis based on the information. Lastly, it learns over time. That's what's really unique about the technology, and then as you apply it to marketing and advertising, it helps marketers to better understand and engage their customers."

Havas Cognitive has been working now in beta with a number of brands for the past six months, including TDAmeritrade, Adidas, and Red Bull. For TDAmeritrade, they utilized a similar approach used in NYC School Finder to find the brand's Most Confident Fan.

"What they did was leverage one of our capabilities called Personality Insights—psycholinguistic analysis of writing—to assess a level of confidence in their fanbase," says Sasse. "They then highlighted the most confident fans based on a writing sample from their social media posts."

Jercinovic says that same kind of analysis can be used in different ways. "For TDAmeritrade, that came through as a confidence ranking, for other clients, it might be about, how do you alter the tone? We're doing tests now in which we figure out how to change individual copy of digital media based on knowing who we're talking to, what they like, don't like, and other factors," he says.

Chasing the rainbow to individual advertising is an ongoing challenge, one that Jercinovic says is based on the trend that traditional audience segmentation is dead. "The demographics that put people in a nice little bucket is gone," says Jercinovic. "You need to treat people differently. We're all chasing this goal of one-to-one segmentation on an individual level, we're not there yet, but getting closer."

loading