AI is powering smarter targeting on paid search and social

AI has changed how paid media campaigns are planned, optimised, and measured. Whether those changes are always in the advertiser’s interest is a different question.

I spoke to Owen Driver, PPC Executive here at Door4, and he explained what Smart Bidding is actually doing under the surface, straight to the point:

“Smart Bidding uses a wide range of real-time signals such as device, location, time of day and user intent to set bids at auction level. It evaluates how those signals combine to predict the likelihood of a conversion, allowing for more precise and efficient bid decisions compared to manual optimisation.”

That capability is real. For large accounts with significant auction data, Smart Bidding consistently outperforms manual bidding on efficiency. The signal volume it processes per auction – across device type, geography, time, search intent, and dozens of other inputs simultaneously – is beyond what any human analyst can replicate at speed.
Bidding efficiency is only part of the picture, though.

Where AI adds genuine value

Tom Morton, Head of Activation at Door4, is more specific about where this all actually matters:

“The audience intelligence side is where AI earns its keep. Identifying core buying audiences, mapping how they overlap across platforms, and surfacing behavioural signals that human analysts would miss are all areas where the technology adds real value. Cross-platform audience comparison in particular has become much more sophisticated.”

This is the less visible side of AI in paid media. Understanding where the same buyer appears across platforms, how audiences overlap between Google and Meta, which signals precede a conversion – work that used to trail behind the data can now keep up. For marketers deciding how to split budget across channels, that matters

Where it gets complicated

The more interesting split, though, is between AI as an analytical tool and AI as a production tool:

“Features like Google’s Performance Max asset generation and Meta’s Advantage+ creative tools frequently override original assets in favour of AI-generated alternatives, often without clear visibility into what changed or why. The performance data may show gains, but attribution is murky enough that it’s difficult to know what actually drove them.” – Tom Morton, Head of Activation

That attribution problem has a knock-on effect beyond the immediate campaign:

“These interventions often sideline the original campaign objective entirely, which erodes trust in the platform over time. For many marketers, their first encounter with these features is discovering the wrong creative has been prioritised, regardless of whatever performance uplift occurred.”

That breach of trust is the real issue. Performance Max and Advantage+ are both positioned as efficiency tools. In practice, they often mean handing creative control to platform algorithms with their own priorities – and those priorities do not always match the advertiser’s. Performance metrics may improve. The ability to know which creative ran, why it was selected, or what actually drove the result often does not.

What to hold onto

Audience intelligence and bidding automation are where AI earns its place in a paid media operation. They process signals at scale, surface insights faster, and help inform where budget goes.
Where the caution kicks in is creative automation. The platforms have a vested interest in keeping their optimisation logic opaque, and the attribution data they surface does not always give a clear account of what actually happened. These tools can be used – they just need proper oversight and a degree of scepticism about what the reported results actually mean.
As Tom puts it: “AI targeting works best as a tool that augments a marketer’s judgement, not one that operates behind closed doors while the platforms quietly redefine what success looks like.”

Contributors: Owen Driver, PPC Executive, Door4; Tom Morton, Head of Activation, Door4