The world of adtech has always been slightly confusing to navigate with transactions and signals disappearing into the black box of obfuscation. However, Lorraine Donnelly, a member of the IAB Data Council and head of data AUSEA for Yahoo, writes that AI might make it even worse.
Feeling like an AI pro because you’ve moved past ChatGPT and dropped “LLM” into a client pitch lately? You’re not alone. AI suddenly feels everywhere. The fact is, while AI is more mainstream in adtech, AI is nothing new.
What is new is the pace. The tools are smarter, the stakes are higher, and we’re officially at the “adopt or fall behind” moment.
Recently, IAB US released its State of Data 2025 report. It’s all about how agencies, brands, and publishers stateside are grappling with and getting excited about AI in the entire campaign lifecycle, including planning, activation, and analysis.
Overall, it paints a picture of a mature and central role for AI in delivering day to day operations and addressing our industry’s many challenges. But it raises plenty of questions.
The big questions
Surprisingly, the report stated that around 30 per cent of US agencies, brands, and publishers say they’ve fully scaled AI across media planning, activation, and analysis, with many more expecting to get there by 2026. But what does fully scaled even mean? And is AI helping solve our industry’s toughest challenge?
First up: measurement. Signal loss and stricter privacy laws are positively changing the way we look at ad effectiveness. Introducing smarter and more sophisticated models that lean into AI to measure success, as seen in the meteoric rise of marketing mix modelling solutions, is allowing marketers to understand how each touch point influences a sale without needing to process individual exposures. That’s a win.
Additionally, according to the IAB report, AI’s great at tasks like audience segmentation and modelling (hello, data clean rooms), helping us move past cookies while still offering safer, more reliable ways to build audiences and tailor messaging.
But AI is falling short in a few key areas, like ad fraud prevention and tasks outside the pipes that require strategic decision-making – some of our industry’s most common pain points.
The interoperability of AI across platforms is also non-negotiable as we put more onus on the algorithms. Yet the tools don’t always play nicely together, and that impacts critical things like data quality, fraud detection and campaign effectiveness, not to mention trust, control and transparency.
So, while we can identify plenty of promising use cases, shoehorning AI into the workflow unnecessarily can add more opacity than value, causing distrust between the agencies, brands, publishers and ultimately our audience.
And let’s not forget the hidden cost of AI. Research shows that the environmental impact of generative AI alone is significant, with every 100 prompts consuming at least a litre of water. I’m all for progress, but with stats like that, we need to be intentional and strategic with how and when we choose to activate AI solutions.
The need for some ground rules is impossible to ignore.
In the US, there’s still a lack of consistent standards around data privacy, accuracy, and ethical AI use – things we need in place if we want to scale responsibly. Closer to home, Australia’s privacy reforms are on the horizon, and thankfully, we’ve dodged the chaos of a state-by-state AI framework.
However, the global patchwork of policies means that alignment feels more out of reach than ever. This will ultimately slow progress and put businesses on shaky legal ground as they try to balance the differing definitions of consent and still deliver quality ad experiences at scale.
So, while AI will absolutely solve some of our biggest industry headaches, let’s not kid ourselves, it’ll create just as many new ones that we’ll need to tackle.
One of those is that brands are getting a bit antsy about how their agency and publisher partners are using AI on their behalf, demanding more transparency. We’ve recently seen major movements from the big holding companies to bring data management in-house, such as WPP’s recent acquisition of Infosum and Publicis’s acquisition of Lotame. The question will be how agencies educate – and charge – clients on their bespoke offerings and the level of transparency clients can expect.
While the pace of adoption and the specific challenges differ in Australia compared to the US experience, the goals are the same. Save time. Do more. Make money. Let’s just hope we can do that while also maintaining trust, control and transparency.
But perhaps keep your head torch handy to help navigate through the new look AI-fuelled adtech black box.
