Can AI Actually Make You a Better Advertiser?

Question

Hey Jon. I've been hearing a lot about Claude Cowork connecting directly to Meta Ads recently and was considering exploring it. Given how important creatives are these days, where do you view AI's role in creating and evaluating creatives? And have you explored specifically Claude's Cowork connectivity directly to Meta?

Answer

This is an excellent question, Kyle, and certainly relevant right now.

Meta recently made AI connectors available, which allow advertisers to connect their business portfolios to their favorite AI tools, like Claude. I’ve personally completed that connection and messed around with it, though I also should be clear I’m not using it with CoWork.

I consider myself to be cautiously pro-AI, but not fanatical about it.

There’s a whole lot of noise right now in this space, and the problem is that everyone wants to seem like the next AI expert. They’re coming up with all kinds of complex systems to automate their processes and iterate at scale. Or at least, they claim to.

I’m naturally skeptical of much of what I see about AI. I doubt it’s much of a stretch to assume that much of the content raving about advanced applications of AI was actually written by AI.

But anyway, I’m not anti-AI, as I was saying, but I’m cautious about how it can and should be used.

The mistake many advertisers make, and we certainly don’t need to limit this to advertisers, is that we assume AI has magical powers as a subject matter expert. We assume that strategies generated by AI are better. Or that if AI reviews our ads, the recommendations are correct and should be followed.

There absolutely is enormous potential for AI applications when it comes to Meta advertising. But it’s critical to understand that AI is not a replacement for knowledge.

Those who get the most out of it already have a strong foundation of knowledge about Meta ads. That way, they can spot bad strategies and bad advice.

The biggest risk is for those who either don’t know what they’re doing or have a foundation of bad strategies and information. They can use AI to automate and iterate at scale, but that just means they’ll be executing bad strategies faster than before.

They’ll compound their problems. Or they’ll trust AI’s direction and get frustrated when they realize that AI automation wasn’t the magic pill they had hoped for.

I think you’re right to focus on potential uses of AI for creative development because it’s the biggest bottleneck for agencies. Much of what happens in the campaign and ad set now doesn’t need constant tweaking, and using AI to make automated adjustments there would likely be counterproductive.

But let’s say that you are an expert on ad copy and creative and what makes a good or bad ad. You could feed your favorite AI tool this information to be used as a foundation for evaluating your ads. It could also be the foundation for generating copy variations based on different personas and pain points.

The key is that you have a provable process, and you can check against the output.

The creative side is getting better, and I’m personally using AI to generate more creative variations. But even that is far from automated since it needs constant tweaks and corrections to keep it in line with what I want.

Eventually, of course, these things will get better and better, and making these adjustments at scale would be more plausible. But right now, it’s extremely risky if you don’t know what you’re doing.

It's a shiny object that a lot of people are gonna be making some big, expensive mistakes around.

Thanks for your question, Kyle!