by Kimeko McCoy * 18 August 2025 *
Ivy Liu.
The promise made by generative AI was that it would make marketing cheaper and faster, optimizing everything from creative to media operations. Silicon Valley has been pushing AI to give marketers an all-in-one solution. But marketers claim that AI hasn’t given them brand building solutions.
The advertising industry may still be some way off from Mark Zuckerberg’s vision for fully automated ads next year.
AI tools and automated tools like Google PMax or Meta Advantage+ are largely relegated for performance marketing campaigns. This is according to six marketers Digiday interviewed for this article. This is partly because there are enough hard data and attributions in a campaign to train automated tools. The Trade Desk’s Kokai platform, which offers media buying and marketing automation tools for CTV, digital videos and other brand marketing campaigns hasn’t exactly won over marketers.
“In general, AI seems to be useful when you tell it what you want and then feed it a bunch of information to make it happen,” Chris Rigas said, vp media at performance media agency Markacy. Rigas stated that AI excels in data-driven tasks, and scaling content creation. The problem with brand marketing metrics is that they have longer feedback loops and fuzzier results. “With upper funnel channels, telling AI that you want a click, view or impression is not helpful as those metrics don’t really drive business value in any manner,” Rigas added.
Despite AI’s black-box nature, marketers claim that automated performance marketing tools are delivering on their promise to target audiences who are likely to click and purchase. It’s harder to target audiences for brand recognition in upper funnel campaigns.
Brand marketing has been difficult to quantify in some cases, say marketers. Measuring things like brand affinity and lift, or awareness, aren’t as conclusive as performance results such as return on ad spending and sales. To bridge the gap, agencies had to develop proprietary tools on top of existing tech tools.
Brandtech Group’s Jellyfish agency launched this year its Share-of Model platform, an AI powered market research tool which scrapes insights out of large language models such as OpenAI’s ChatGPT and Meta’s Llama. Karen Bennett, U.S. Managing Director for Jellyfish, says that these insights can be used to improve media buying and audience targeting.
Jellyfish’s approach is not unique. Alex Beddoe is the head of biddable advertising at Transmission, a B2B marketing company. Transmission uses lookalike audiences, digital twinning, and brand simulation to improve campaign results and simulate brand impact. It’s an effort to create digital customer models to better predict the performance of brand campaigns. Beddoe stated that a large portion of client spending is now going to predictive audiences in order to contextualize campaigns. He did not give specific figures on client spending.
Rigas, at Markacy, says that clients have also withdrawn from Meta Advantage+ campaign. He said that over the past year the agency has recommended clients to carve out a video purchase in order to increase brand awareness for Meta Advantage and Google campaigns.
Platforms such as Google, Meta, and Amazon promise automation and scale with their AI tools. But brand marketers haven’t fully embraced AI. Concerns about the black-box nature of these tools have been around for a long time. These concerns have been intensified as tech innovation continues at a pace faster than marketers can keep. In one instance, a 20 year growth marketing veteran advised clients to walk away PMax.
Despite this, marketers still believe that it is only a matter time before automated marketing tools become one-stop shops. For now, however, their current capabilities and measurement limitations keep them at the bottom.
It’s still early, but clients are already seeing some benefits by leveraging AI tools. Anthony Costanzo is the chief analytics officer of Mile Marker, an independent media agency.
