Artificial Intelligence may be the most bandied about term of 2017. For consumers, AI is powering everything from virtual personal assistants and real-time translation to GPS navigation and self-driving cars. In business, AI is under the hood of everything from ride-sharing fleets to aerial analysis of shopping malls to credit scoring.
Ad and marketing tech, of course, are no exception. As the CMO’s role grows to include everything from ad tech to customer relations and corporate strategy, it becomes clear that unique and valuable insights—the kind analyzed by AI from mountains and mountains of data—are key. “The term AI is really fraught with multiple definitions,” says Wilson Raj, global director of SAS, “but I think what’s really happening is the data revolution.”
After all, the explosion of minable data and the ability of machine learning and AI to make sense of that data means that AI is a no-brainer for marketers to maximize their reach. Likewise, consumers accustomed to flawless Netflix recommendations and one-hour, real-world Amazon deliveries now expect advertising to display similarly predictive powers. That’s because surprising, delightful experiences aren’t just the purview of laptops and smartphones.
Will the real AI please stand up?
What do we mean by AI? At a minimum, it’s technology that enables machines to perform more like humans, with the ability to understand language, recognize what’s in front of them, translate and make decisions—and, in a perfect world, to do this in a way that looks and feels human. But what does this mean for marketers?
“It’s important to be on those platforms [that use AI] so that we understand the benefits, so we understand the customer experience,” says Pat McLean, CMO of TD Bank. “We believe that customers still have an expectation that AI is delivering a personal and meaningful interaction for them and it’s not just transactional.”
For something to be AI, it needs to be able to learn on its own. Much like humans, machines need real-world examples to do this. In this case, the “training food” is data, and the more data a machine receives, the smarter and more accurate it gets with whatever human task it has been designated to do.
Data, data, data
Typically, targeted advertising has relied on a few data points: age, demographics, location, gender and purchase history, for instance. Marketers decide what groups or geographical locations they want to target, then set it and forget it. Today, marketers know how valuable their data is, and they’re demanding it be used to make their messaging resonate. Crunching all those seemingly disparate data points can only be done by AI, which gets better and better at finding the right moment to serve an ad.
To do this, AI gleans deeper insights around, say, something as complex as the weather, making it possible to tailor creative and messaging in real time.
“We’re finally getting to critical mass in terms of the data we get off of these IoT machines and sensors around the world, which can be used by business, for business,” says Jordan Bitterman, CMO of The Weather Company, now part of IBM Watson. “Where we drive more sophistication is when we combine that data with other IoT data, and that data could be data that comes from cars, from washing machines. So whether you’re in auto, retail or CPG, there are always ways we use that IoT moving forward.”
For financial services giant BlackRock, which manages over $5 trillion in assets, AI makes it possible to reach investors at critical moments. “Gone are the days where a brand can plan for a major moment like the Super Bowl and have somebody sitting in the room ready to send out a tweet,” says Jason Hill, BlackRock’s global head of media. “When you’re in an industry that’s tied to the markets like we are, every day is the Super Bowl. There’s always something happening out there that’s going to send markets in one direction or another, that’s going to have people thinking about their investments.”
One person, many screens
One area where AI-informed advertising can help is in the marriage of television and digital. For the average consumer, seeing the same ad over and over across devices is a particularly annoying experience. AI can solve that issue.
“The holy grail of targeting in a very fragmented world—and this is a very fragmented world we’re living in right now—is that you’d be able to have one view of the consumer and you’re not hitting that consumer with an ad on a mobile device, then the same ad on the Web later or a tablet later,” says Bitterman. But with the advent of Internet-connected Smart TVs and with traditional TV data being added to online behavioral data, collaborations with the ad tech world can start to address this issue, seeing people across devices and screens, rather than just the devices.
So what’s next? Prediction
Besides allowing for one view of customers across devices, AI allows ad and marketing tech to provide the same predictive and proactive executions that can be found in other sectors. For example, to train its AI, Rocket Fuel uses logistic regression to identify “conversion” moments—a point at which a consumer is likely to be influenced by an ad—by processing a wide variety of real-time and historical data from first-party sources and third-party segment data. Then, for dynamic creative, which enables the delivery of ads specifically tailored to a person, Rocket Fuel uses deep learning neural nets able to analyze additional nuances around actual shopping behavior, depending on what’s in an ad.
By mixing these two methods and a lot of great data, AI starts to move into the predictive realm: In other words, delivering a personalized ad that’s as welcome as, say, a Google Now alert telling you it’s time to leave for your next appointment.
“The amount of data that’s available for analysis is really amazing, but beyond amount, it’s the speed at which we’re getting that data,” says Raj of SAS. “So in order to cope with that level of volume and velocity, you definitely need machine-learning algorithms that go beyond just analysis, but actually start working through what we call the predictive realm to be able to auto-segment, to be able to understand the context of the customers, whichever part of the journey they’re at, and to be able to provide those recommendations in pretty much real time or near real time.”
Jump in marketers, the water’s fine
Regardless of a brand’s needs or implementation of AI, one thing is clear: The time to jump in and embrace AI is now.
According to Eric Duerr, CMO of Rocket Fuel, the practicalities to get this done require a three-pronged strategy. First is a data strategy that addresses how you are going to involve your first-party and CRM data with third-party players—and remember, the more data, the better the AI. Second, you need a tech strategy that involves deciding which marketing cloud you’re going to connect with. It also helps that more and more companies—from Adobe to Salesforce—have major AI initiatives today. Third, you need a services strategy related to the people who are going to come in and operate the keyboards and understand how AI works. Finding the right talent is a challenge, for sure, and at least it bodes well for anyone who thinks AI doesn’t need human help.
