Artificial intelligence is coming to marketing, because it’s the only way to turn the huge amounts of data companies are collecting on their customers into meaningful interactions with them.
That was the view of AI that emerged during the conference sessions at this year’s Dmexco event in Cologne—that the technology is the crucial piece in the puzzle of big data and marketing, the final element needed both to turn data into insight, and then to deliver on that insight in the form of personalised marketing at scale.
Adobe’s vice-president of platform & products, Suresh Vittal Kotha, discussed AI in terms of companies’ need to focus on customer experience. He highlighted its ability to help marketers understand the context in which they were communicating with individual customers, orchestrate meaningful customer journeys across multiple touchpoints, and do this at a speed that makes the personalisation imperceptible to the consumer.
“Experiences must be adaptive, intelligent, and able to track customers’ multichannel footprints,” he said. The way to do that at scale, he argued, is with the help of artificial intelligence.
This view was echoed by Marcus Ruebsam, senior vice-president at SAP Hybris. He stressed the complexity of the task facing brands in being consistently relevant across customer journeys, complexity that he argued could be broken down and dealt with by AI.
Let’s Work Together
Contrary to the usual characterisation of AI as a bunch of robots coming to take marketers’ jobs, the stress at Dmexco was on the need for humans and machines to work together. Ruebsam talked about how machines will take on tasks that are routine but complex, and pointed out that machines, even intelligent ones, had to be trained by humans. Meanwhile, Limor Schweitzer, chief executive of robot development and distribution company RoboSavvy, reassured his audience, saying that while AI would automate many things but “anything humanistic is safe.”
A concrete example of AI being used to augment human efforts emerged in the session featuring Mark Thompson, president and CEO of the The New York Times Company. He described how The New York Times was using AI to help filter the comments coming in to the paper’s website, a flow that is too big for their human moderators to cope with. An AI trained to recognise what is acceptable in a comment and what isn’t now makes the first cut, and the comments it lets through are then reviewed by the human moderators.
For some companies, the value of AI is its novelty. German retailer MediaMarkt Saturn has a robot in one of its stores in Berlin, where it welcomes customers, asks them which products they are interested in, and provides them with information. The benefits, according to Saturn’s head of digital transformation and projects Sonja Moosburger, are mainly in the robot’s novelty value—its attraction for children, for example. But also she circled back to the idea that robots can take over basic tasks, while more complex ones, such as judging context and emotion, were still better performed by people.
The subject of job losses to AI came up in several sessions, with RoboSavvy’s Schweitzer noting that the replacement of human workers with robots was nothing new. But as Dentsu Aegis Network’s chief strategy and innovation officer Nigel Morris pointed out, the beginning of the AI age is coinciding with a massive talent shortage and often fierce competition for people with the right skills.
Talking to Morris, LinkedIn co-founder and vice-president of product management Allen Blue characterised this as a “hollowing out,” with those hit by changing demand for skills needing to either move up the value chain, or switch to another sector.
This was a point also picked up by The New York Times’ Thompson, who talked about the business’s need for specialist digital skills. He explained how, rather than losing people, the organisation had swapped out employees for new people with the new skills needed, reaching a changeover of more than 80% of the paper’s advertising staff in the past six years.
In parallel with changing their skills profiles, AI is also forcing companies to think about their underlying organisational infrastructure. Joe Zawadzki, CEO of programmatic marketing technology company MediaMath, talked about the need for companies to break down silos and rethink data flows to enable them to employ AI. Infrastructure, he said, needs to be “bulletproof.”
The other point that was made repeatedly during the conference sessions was that even though AI’s precursor, machine learning, is already widely used in current marketing—in search, or in recommendations, for example—we’ve barely begun to tap what the technology can deliver. Mainardo de Nardis, global CEO of media agency OMD, characterised the industry as being at 1% of its potential use of AI, and contrasted how developed even this low level of use felt when compared to industries such as air travel that have yet to begin adoption.
Dmexco 2017 saw AI being talked of in the same way as programmatic was five years ago. The discussions show the marketing industry sees it as the way of delivering on the internet’s early promise of enabling one-to-one marketing at scale. And while adoption is currently limited, the pace of change will never be this slow again.