James Hutchison, Head of Performance Media & Growth, Ryvalmedia.
James Hutchison, Head of Performance Media & Growth, Ryvalmedia
There’s an overwhelming wave of businesses across all industries leaning into AI, the basis of the idea is that efficiencies can be found that outweigh the risks. Simply acquire AI tools and see profits rise.
It's a compelling story. It's also dangerously incomplete and I'd argue most of the people that have actively engaged with AI understand the trade offs involved.
AI is the most powerful tool the media industry has ever seen, but a tool without a skilled operator isn't a solution, it's a liability. In an industry where a misplaced dollar is a client's lost opportunity, that distinction matters enormously. More than most platforms would like you to believe.
Meta has been pretty transparent about its ambitions: fully AI automated campaigns by 2026. Set a goal, let the machine handle the rest. Google's Performance Max already bundles search, display, YouTube and Shopping into one AI-driven campaign type. Both promise efficiency at scale. These advancements have numerous advantages, however, the tradeoffs are not clearly articulated or responded to.
Performance Max gets called a "black box" so often that Google was actually forced to add negative keyword support to provide enough confidence for mass adoption of the new campaign type. Meta removed detailed targeting exclusions in early 2025, citing improved performance metrics – never providing the ability for agencies to truly test comparisons. While many AI-driven platforms still limit full visibility into optimisation logic, advertisers today have more tools than ever to interrogate performance patterns, test outcomes and apply strategic oversight. If a Google or Meta campaign underperforms, or a brand ends up in an unsafe context, agencies and marketers are limited in their ability to investigate, pivot or respond in ways that were previously available.
A recent survey found that 70% of senior marketers have encountered at least one AI-related incident in their advertising.
Here's what the platform pitch glosses over. AI gets you part of the way there. There is still an absolute requirement for direction, creativity and context. AI in isolation lacks any of the details needed to provide a complete outcome. The difference between a straightforward message and a sarcastic tone. The cultural nuance that makes creative land in one market and fall completely flat in another. The fact the campaign that’s being created is only 1 of 6 in the total campaign strategy.
Programmatic buying is the clearest example. AI-powered DSPs now evaluate hundreds of campaign inputs in real time, optimising bids, filtering fraud and adjusting creative dynamically. Layer that with Meta's Advantage+, Google's AI Max for Search and first-party data strategies across walled and open ecosystems, and the efficiency gains are genuinely impressive. Over 91% of the programmatic display purchased in the USA now flows through private marketplaces and direct channels. That's not despite human involvement. That's because of it, because skilled buyers bring trust and curation to supply-path decisions that no platform can replicate.
The best programmatic teams understand that AI flags potential risks. It doesn't decide what crosses the line. That still rests with trained professionals. Brand safety in 2025 isn't a technical detail, it's a strategic risk. 53% of marketers name social media as the top threat to their brand's reputation. One misplaced placement in a volatile content environment can erode years of brand equity faster than any campaign can rebuild it.
The competitive advantage now isn't knowing how to manually execute a media buy. It's knowing how to interrogate, direct and quality check what AI produces. Knowing when Performance Max is driving a massive uptick in conversions because it’s targeting generic, unrelated terms. Knowing when Advantage+ is optimising for the metric it was given, not the outcome the client actually needs. Those are different things.
Without that human layer, every brand in a category starts to look the same, because they're all running the same backend tools, with the same approach and the same tone. They’re also at the mercy of optimisation without strategic context that doesn’t connect to required business outcomes.
As AI governance matures, agencies are increasingly combining automation with measurement frameworks, human review processes and explainability tools that make outcomes more interpretable and accountable.
The future of performance marketing isn’t human versus machine. It’s knowing how to get the best from both.
Have something to say on this? Share your views in the comments section below. Or if you have a news story or tip-off, drop us a line at adnews@yaffa.com.au
Sign up to the AdNews newsletter, like us on Facebook or follow us on Twitter for breaking stories and campaigns throughout the day.