There are really two schools of thought right now on using AI for Meta ads optimization. The first camp treats the algorithm like a perfect, hands-off vending machine. The second treats it as a powerful co-pilot that still needs an expert human in the cockpit to set the destination and watch the instruments.
Camp A: Trust the Machine
This first group argues that Meta's AI is now so sophisticated, the best thing a performance marketer can do is get out of its way. The central idea is to radically simplify your account structure, collapse your audiences, and put your faith entirely in tools like Advantage+ Shopping Campaigns. You stop segmenting audiences manually and instead focus all your energy on producing high-quality creative and feeding the machine clean first-party data.
The argument, which you'll hear from Meta's own reps and successful brands, is that the algorithm can find pockets of customers you would never think to target. On The DTC Podcast, leaders from brands like True Classic and Kitsch shared how AI-driven optimization and Advantage+ were cornerstones of their growth into nine-figure businesses. Taylor Holiday of Common Thread Collective also made the case that this is the future, allowing brands to focus on the two things that actually move the needle: creative strategy and product expansion.
Camp B: Keep Your Hands on the Wheel
This second camp views AI not as a replacement for the marketer, but as an incredibly powerful assistant. They argue that while Advantage+ is effective, relying on it exclusively is like flying blind. When performance dips, you have no idea why, and you have few levers to pull. This approach advocates for using AI to augment human strategy, not abdicate it. It’s about building a more resilient system for Meta Ads Optimization.
This means using AI for specific, targeted tasks. Yahav Hartman of Madgicx explained on The eCom Ops Podcast how AI can be used for deep creative analysis to identify winning concepts faster. It also means using a more disciplined framework for creative testing, as the Pilothouse team detailed with their "Andromeda" strategy on an episode of The DTC Podcast. Instead of just throwing everything into a single massive campaign, you use AI-powered insights to run smarter, more controlled tests. As Joe Apfelbaum put it, you must have a clear objective before you deploy any AI strategy.
Personally, I believe the "co-pilot" model is the correct one for sustainable, long-term growth. The "vending machine" approach can absolutely work, sometimes spectacularly, but it’s brittle. When the algorithm changes or your creative fatigues, the whole system can break without warning. The real magic of AI in advertising is that it frees up human marketers from tedious tasks so they can focus on higher-level strategy, creative direction, and understanding the customer. The machine is only as good as the inputs, and strategy and creative are uniquely human inputs.
So, your approach should depend on where you are. If you're just starting out or have a limited budget, going all-in on an Advantage+ campaign is a fantastic way to get a quick performance baseline and let Meta’s AI find your initial customers. But if you’re trying to scale past seven figures and build a durable brand, you need to evolve. You should adopt the co-pilot mindset, using AI tools to get smarter about your creative, your offers, and your audience, and building a marketing engine that you truly understand and control.
