Artificial Intelligence is difficult to define and even more difficult to build. You'd never know it from reading the news, though. Recently, every tech company and agency is touting how their AI is revolutionizing their industry. Alibaba has one that passed the Turing Test (no, it didn't). McCann's wrote a better ad than its human employees (well, more because of its human employees), and Google's scheduled a hair appointment over the phone without any human intervention (okay, this one merits some consideration). The articles that examine these advancements almost universally include a section about how disruptive the technology will be – ultimately rendering humans useless and putting us out of the job.
In reality, all of these projects are interesting, some of them are nonsense, but none of them are scary.
While I genuinely believe that AI will become so powerful that it will doom the human race, I think that dystopia is centuries away. I'm far more bearish on its disruptive impact this decade, particularly on the marketing front.
Nonetheless, I'll provide a few tips to bring your company's marketing efforts up to parity with some of the industry leaders in AI:
Write good creative
Build good infrastructure
Collect good data
In other words, exactly what you've always strived to do, and exactly what a broader digital transformation empowers.
Fundamentally, all AI platforms operate in the same way. They ingest data, both structured and unstructured. They organize that data and use it to respond to novel requests. They then process feedback and use that to make better responses in the future. To make the AI smarter, it is still incumbent upon us to give it good working data (creative assets, placements, messaging, brand guidelines, etc), ask it intelligent questions, assign value to its responses, and continue to provide it with novel data over time.
In the case of McCann, their AI succeeded because it was fed data from some of the most successful ad campaigns in recent history. It was able to break down the individual elements that made them successful and combine them into new assets. But the underlying data - the brilliant ideas that taught the machine how to be brilliant - were decidedly human. Marketing isn’t completely objective – at least not yet. It’s relative to our culture and our society at this moment, and it requires thoughtful curation. It was only two years ago that Microsoft thought they could automate culture, and the result was a runaway AI Twitter handle that almost immediately devolved into a racist troll. Even in this future, creative minds are as valuable and imperative as ever.
Likewise, the systems need to be taught how to learn, and this requires feedback. In digital marketing, this feedback system is our website, the analytics behind it, the way that we translate our business requirements and goals into digital values, and all of the third-party technologies that both guide and process the consumer experience. Strategists, analysts, and engineers are all critical to a successful AI.
AI isn't about replacing the value created by good ideas and skilled labor. It's about applying the same best practices that make marketing work today to new systems that can make more out of the components we create than we could otherwise. As easy as it is to break down AI relevance to marketing fundamentals, we are all still at risk of becoming complacent. If anything, this underscores the need to ensure that your company understands its own digital transformation.
More and more breakthroughs in marketing are technologically driven, and companies that lag behind on their infrastructure are being outcompeted by savvier rivals. Their ads aren’t finding the right audience. When they find that audience, they aren’t serving the right message. When they serve the right message, they’re overpaying for it. The end result isn’t just inefficiency, it’s irrelevance, and it happens faster today than ever before.
So stop worrying about AI, machine learning, or whatever term they use next month to describe smarter, faster marketing decisions. Instead, keep your creative team doing what they do best, and make sure that you have a data infrastructure in place to make the most of it.