When I talk to people about artificial intelligence, or AI, they’re often optimistic about how they think AI will improve their lives in the future. What they aren’t thinking about is sweaters. Specifically, how based on previous online behaviors an e-tailer knows that you’re in the market for a warm, blue sweater instead of a long-sleeved gray shirt. Historically, businesses required human interaction in order to understand exactly what a customer needed. Today, that’s no longer the case. Thanks to lots of data and AI, companies now know exactly what you want, when you want it and how you want it.
Some may fear AI, but I think of it as doing for minds what the Industrial Revolution did for muscles: machines taking on difficult and repetitive tasks and improving output. As in the case of an online retailer, AI will change the way marketing is done by turning to machine learning in order to discover patterns in data on human behavior. We don’t have to look far into the future to see how AI will make every marketing campaign smarter and more personalized; we’re already there.
Analytics that look at, not for
To understand how AI will change the role of the marketer, we have to first look at what it does well. AI finds patterns in data, which we humans do poorly by comparison. Here’s why: first, there’s too much data to pour through, and second, we’re biased. We only ask what we know to ask for. By contrast, machines aren’t necessarily looking for something—they’re just looking.
By looking through massive amounts of information, AI discovers its own design as it goes. The machine finds surprising information in patterns, and not just in individual shopping patterns. It’s able to capture large and impactful trends like social patterns that define a community’s habits.
So why is the time for AI now? Only in recent years do we have the volume of data needed to find these patterns, and the economies of scale to be able to store it.
Intelligent journeys that drive results
Marketing has traditionally been linear and deterministic. But with AI, journeys become intelligent and dynamic, and marketing can be predictive for each customer. The ability to capture, store and retrieve a boundless amount of data opens up avenues to reach customers with personalized messages and experiences. And, since the system learns, it gets smarter and more accurate with further iterations.
Let’s take a look at the famous auto company Peugeot and how it has been able to use AI to drive impressive marketing ROI. Of the millions of people who come through its doors, visit its websites, even look at an ad on the open web, only thousands of them buy a car. We can ask what sequences of events people followed to buy, or not. There are thousands of ways in, from downloadable content to walk-ins. Using data and AI, Peugeot created 800 micro-segments that enabled content personalization across 2,200 microsites. By following the buying journey of those who did buy, Peugeot saw how each path was different. With the power of AI, companies like Peugeot can tease out the many different variations to sell and market accordingly.
The bionic marketer of the (near) future
What about the creativity and art in marketing and advertising? Design remains a huge component of successful marketing. AI certainly has the power to help determine what sorts of digital ads, for example, a consumer is likely to click on—from color preferences to style and price. Many experiments with AI-created trailers, tweetbots and other interaction experiments have taught us that we still need both human and machine. The AI algorithm can interact, iterate and optimize for success, but design still lies in the art of human creativity.
AI isn’t an either/or for marketers
AI removes the tedium and guesswork of running a marketing campaign, which includes creative, email, social media and more, by generating insights from haystacks of data. But the marketer still designs and runs the campaign. What I see is people and AI working together to be more efficient than either one alone.
Creativity aside, the beauty of AI is that even if marketers don’t have specialized IT skills, they can still act on high-level insights. For example, iteration cycles start to become much faster as AI kicks in, so I think of this process as a layer cake.
The first layer is a huge quantity of data that AI uses to figure out what to probe for insight. If you perform a large-scale data-driven experiment with a piece of content to test engagement (like A/B testing), you’re going to get a lot of data back. This is the second layer. AI can then process that information quickly, and the iteration cycle gets shorter. In the third layer, AI works through nuances in the data so that it’s not just A or B, but a mix of both. By the fourth layer, AI has optimized the system, working out inefficiencies, maximizing speed and providing a clearer picture of customer preference and behavior.
What tomorrow already looks like
Everything I’m describing here—the massive data gathering, the importance of trust, the ability to analyze, iterate and predict, and intelligent journeys—are separate pieces of what is already shaping up to be a seamless experience.
As marketers work hard to get ever closer to the customer at every touch point, AI gives them the total view of what’s possible and what makes sense, reducing the friction to customer satisfaction even as their expectations are constantly rising.
Leslie Fine (@lesliefine) is vp of data and analytics at Salesforce.