How AI can impact long-term creative thinking

This post from Automated Creative Founder Tom Ollerton originally appeared in New Digital Age.

What makes a great advert? Spoiler: probably not what you think.

Every time we start working with a client, we ask them to play a game which involves guessing the emotional triggers that will drive performance in their ads. In the last two years, we’re yet to see one person guess correctly. 

As humans we bring bias to the table, both consciously and unconsciously. We can’t help it - we’re products of everything we’ve seen, done and experienced during the course of our lives. I like to think most of us are self aware enough to know that we’ve fucked up or made a bad decision or two in the past. Yet somehow, we’re too proud to admit that same mentality might apply to how we approach our work. 

There are many amazing, talented planners and creatives out there, but they are not infallible, or unbiased. But there is help at hand, in the form of artificial intelligence. AI can play the role of mediator between a creative idea or route, and the people that will eventually see it. 

AI in advertising is generally perceived as a tool to enable better performance. And of course, it can do this, but there’s also a bigger opportunity, one that many brands aren’t tapping into. AI can help marketers understand the semantics of their advertising - what the strategic goal of the ad is, what the visual and written triggers are and how those things are linked. It can tell you why ads work. And in doing so, create your own first party data that no one else has, that can guide your future approach to creativity.  

For example, take slipper brand Mahabis (our client). In testing the impact of machine learning on its online advertising output, its team’s ‘gut feel’ when it came to creative wasn’t just wrong; the content that the team initially disliked was the best performing. It also found that stills worked over video for certain audiences, something the team was not expecting. These results, and others, led to the brand completely changing its creative approach, reducing production costs by half in the long-term. 

The reason for this is simple. Creative ideas are guesses (or perhaps more generously, research-based opinions) that an audience will like one idea. Media agencies move them around and hope they find a match. This is fine if the ideas are great, and the audiences are spot on. But this isn’t always the case. True brilliant creative ideas are rare as hell - there are a lot of people trying to have them and not many regularly succeeding. 

So does AI mean death (or at least, dole) to all creatives? Absolutely not. The way I see it is that machines should do small tasks very quickly, allowing humans to do big tasks slowly. Strategic thought, creative thought, coming up with ideas from scratch - this takes a long time and it’s a human process. 

But creating different versions of ideas, at mind-blowing scale, refining ideas to their absolute most relevant for an audience - this, machines can do quickly. While providing insights that will reduce creative costs in the long-term, by helping spot what works and what doesn’t - from an unbiased position. 

We’ve got a saying amongst ourselves at Automated Creative - “What’s weird, works”. We’ve seen that it’s often the least expected routes that get the best results. We’ve been as surprised as everyone else. 

So marketers - it’s time to let go of the ego and embrace your infallabilities. You’re not the only one who gets it wrong. But now’s your chance to be one of those who actually gets it right.