What We Lose When AI Takes Over

Balancing the "Black Box" of AI with the core work of marketing

Everyone knew it was coming and it was only a matter of how soon. Just in time for the holidays, Google started rolling out generated ad assets on a wider basis in Performance Max. We're using it with a couple of our clients and so far it's been - no surprise - good and cool. Chances are you’re using it too.

PMax is driven by Artifical Intelligence. It is an artificial intelligence. It has, for all intents and purposes, a brain of its own.

It takes over some of the choices and work that a media planner might have done: Channels, placements, timing, some targeting. PMax does it millions of times faster and with scientific precision, resulting in better results, faster.

It’s a black box of sorts. Brands are essentially hiring an alien intelligence - an unknown and relatively untrusted brain - to take over some of their basic decisions about media.

You’ll never fully understand the decisions PMax is making and the alternatives considered. You'll just know its working because your numbers are going up and to the right.

In the big scheme of things, adding generated images and headlines is an evolutionary step, a sustaining advantage for a platform like Google Ads. For now, the types of creative units are pretty limited. But, you know it’s only going to get better, that we’re only a couple years away from AI generated pre-roll video and ads.

For marketing teams, though, it’s an important step into their future, a milestone on the way to the real revolution that’s underway. More and more, brand teams will give away choices and decisions to the alien intelligences they’ve introduced into their mix. Today its basic media and digital ad production, but tomorrow it will be realtime, AI generated landing pages, full CSR experiences, and email creative. You know it’s coming at us.

Of course, the state of the art keeps advancing. Tools change, and that will unlock efficiencies and effectiveness in all sorts of ways. AI inside our tools will push us closer to making marketing a science.

But, be mindful of the risk: The more decisions you give up, the more your team loses.

You lose some of the benefit of learning. You won't be thinking about the creative units or the placements themselves (Because the tools are doing that). You won't have to think about setting up the test & learn experiments (because the tools are doing that).

You'll still be learning, but perhaps not as much or in the same way.

By not doing some of the test/learn efforts, you’ll have a little less understanding of what worked, what didn’t, what responses you got, how did our hypotheses work in the real world, etc.

You’ll lose some of your “touch”, your intuition about the market and the consumer. Every good marketer knows how important their intuition is in making strategic, tactical, aesthetic, creative choices. Handing over creative and production choices to AI gets you out of “Maker” mode, where you build intuition and a strong sense of what’s “good”, what is “Better” and what is “best”.

In the future, though, you won’t be able to blame your AI’s for your bad taste.

You may lose empathy for your audience or your users as more and more of your messaging and UX decisions get handed over to your GPT and AI driven tools.

Marketing teams should be realistic and thoughtful and practical about what might get lost when they hand over work to smarter tools. They have to be careful to keep what makes their work creative and, ultimately, important.

Great marketing teams (great marketers) will stay focused on the things that don’t change: Understanding the unexpressed desires and intent of our users. Communicating in compelling and interesting ways. Delivering experiences that are safe, rewarding and have integrity.

Marketing teams will have to stay committed to the expensive, hard, sometimes boring work of primary qualitative research. They’ll have to invest in active listening, deep insights and UX work. They’ll have to own the effort of finding new opportunities to build meaningful growth for our businesses.

We can let AI do the tasks, but we have to do the real work so our own expertise doesn’t get lost.