Artificial intelligence in marketing

Artificial Intelligence has been around since the 50s, but the beginnings go back even further. In 1912, Spanish engineer Leonardo Torres y Quevedo built a functioning electric chess computer – who only mastered the endgame with 3 pieces but always won. People were already enthusiastic about this artificial intelligence in 1912, for the standards of the time. – while some had tremendous misgivings. Today it is quite the same: artificial intelligence is a big topic that is euphorically acclaimed from the one half and eyed critically to apocalyptic from the other.


Artificial intelligence in job market.

Today we all get in touch with artificial intelligence as it is being used for many marketing activities. Often it’s things that don’t even impress us any more, such as simple recommendations for online shopping. The shopping spree on the Internet stores data about the preferences and needs of the customers. AI’s assemble these puzzle pieces into attractive offers and personalize the customer’s selection. Current figures suggest that Amazon generates 35% of its current sales through automated product proposals.

One topic that is being discussed a lot these days is unemployment. Many people see their job at risk through automatization, and figures confirm it. Could it be a thread in creative industries as well?

Authors, editors, artists – all these professions thrive on their creativity. It may not be a big thing for artificial intelligence to take on rational messages, but when it comes to writing whole articles and stories, there might be a reason to worry.  But if you look at the subject of AI from a different perspective, it can just as easily represent an opportunity instead of a loss. It will become possible to do data analysis at a speed and a mass that is simply not feasible for humans. Furthermore, the combination of artificial intelligence and people could optimally distribute tasks. If an AI software takes over a routine task, that saving of time could be invested in personal and human creativity. Especially in marketing, that’s an enormous benefit.

Though, when it comes to automatization there will always be another issue: data protection. Marketing activities like retargeting are common and very useful today, but as the AI ​​progresses, more and more private data is being stored. Artificial intelligence relentlessly deals with data. AI-based algorithms create a filter bubble which lead to a kind of a restriction. Once the user was looking for something particular, he keeps seeing the same or similar content over and over again. In this way, targeted influence can take place not only for marketing purposes but also in the areas of politics and elections, including price manipulation. Sure, each person has his own responsibility to inform himself comprehensively and outside of the internet as well, but still, limits must be set. The legal position of individuals and their control over their data should not lose importance. When using AI applications, therefore, concepts for data protection and IT security are required.

Although, there are areas where artificial intelligence has asserted itself already, for example in customer service. By launching a chatbot platform for Facebook in April 2016, the hype around so-called intelligent dialog systems began. These are cloud-based robots that simulate human communication. Meanwhile, more and more companies use these services and integrate chatbots into their client communication, especially when it comes to social media. As artificial intelligence is still in its infancy at the moment, they certainly do not work as a human responder and should not be expected to be flawless – still, they are constantly being optimized. The software of Stanford NLP, for example, has already reached the point where it recognizes emotions and tries to figure out content. It understands which words carry meaning and how they are linked to one another. Therefore positive and negative moods or ratings can be detected and responded to accordingly. Stanford NLP offers new opportunities to customer services – employees could be replaced completely.

That takes us back to talking about unemployment. But on the other hand it is also clearly more economical. Also, surveys show that people who work in customer service are more dissatisfied than other employees. They have to deal with human dissatisfaction everyday – which artificial intelligence would take away from them.

Taking a look at AI in customer services leads to the question of the international usability of this kind of Intelligence. Big companies mainly advertise internationally and should be able to answer customer questions in any language and in the same quality. The main point proves the inability of AI translating humor or the tone of an article as the editor meant it. It knows nothing about the author’s personal background and cannot find out about the author’s intention behind a sentence. Being profound, using humor and touching peoples individuality are three of the most important tools used by marketers nowadays. “Deep Text”, a technology used by Facebook, is a translating technology that could be able to act like that one day. It already understands the content of over a thousand posts in a second. In the future the technology will be able to translate advertisement campaigns as well, recognizing the different tones and senses of humor used by marketers.

Mentioning all these opportunities offered by artificial intelligence may give the impression of AI taking over whole marketing campaigns – from planning through creation up to settlement. AI is still far from perfect. Furthermore, the AI has to be developed by the marketer as he has to feed it with knowledge. The marketer cannot be replaced, not today nor in the future. Artificial intelligence is a powerful technology for simplifying and accelerating marketing campaigns. It makes it more efficient, but will not completely replace it. Future marketers must find a balance between automation and individualization – the AI’s job is to support and simplify that process.


A text by Lisa Fritz, Frauke Lippert & Helen Gleixner