AI Ads and the Shift from Inferred to Explicit Intent

Advertising playing on emotions isn’t new. It’s a tale as old as time.

Edward Bernays, the figurehead of modern advertising (and nephew of Sigmund Freud), completely redefined the way people approached wants and needs. By shifting focus away from an individual’s needs, there was room to capitalise on something far more powerful, albeit less rational: want. 

But that’s what marketing is all about at its core. Want, and the intent behind it.

Understanding what people want and finding ways to turn that into profit has always been the name of the game. That’s how it started a hundred years ago with Bernays, and that’s how it is today.

Ads are coming to AI platforms

Now, unless you’ve discovered that living under a rock is somehow more affordable than paying the £2.5K average monthly rent in London, you’ll have probably seen the announcement made by OpenAI earlier this month.

Specifically, that it’s begun testing ads within ChatGPT.

For a select group of users, the “lucky” few, depending on your view, sponsored messages are now appearing inside the app. According to Search Engine Land, while ads are labelled, the sponsored tag sits beneath the chat box, visually removed from the AI’s generated response.

A deliberate attempt to distance answers from monetised content?
Only time will tell.

A fundamental shift in how intent is captured

Either way, this marks a fundamental shift in how people are searching.

Traditionally, intent has been inferred from online behaviours: search queries, websites visited, products viewed. These are the signals that platforms like Google and other paid search ecosystems have used to fuel what is arguably the most powerful tool in modern marketing.

But ads within AI chat platforms change that.

For the first time, intent isn’t just observed and approximated, it’s explicitly spoken. And to a robot, no less.

Whether users realise it or not, through conversations with generative AI they are willingly handing over incredibly valuable datasets, the kind that are typically difficult to capture elsewhere, which can then be used to do what marketing has always done: market to them.

The trade-off 

It’s interesting to watch this technology evolve within the ecosystem, but the trade-off here feels slightly more uncomfortable.

What began as an exploration into the capabilities of machine learning now seems to be edging closer to something far more commercial, effectively a very sophisticated cash engine.

And one that will likely be successful, because it already has the data it needs.

What happens to paid search?

Another question we need to ask is whether paid search remains the dominant model in advertising.

The naive answer would be yes.

As it stands, AI-assisted answers still rely heavily on the wealth of data that Google has built over time.

But at the same time, there will always be users who want to research products the old-fashioned way. For these users, research isn’t just about functionality or price, it’s about control.

Trust will define how far this goes

AI advertising may well redefine the marketing landscape, but there’s a limit.

People don’t want answers, or ads, forced on them.

Platforms like Perplexity AI have already seen this first-hand, where introducing ads raised concerns around credibility and authenticity.

Until users are confident that the platform is giving them the best possible answers, there will always be resistance to monetisation.

Final thought

So, whilst AI might become better at predicting what you want, platforms like Google still give you the freedom to figure that out for yourself.

And for that reason alone, paid search, at least for now, is here to stay.