For the past two years, marketing conferences, agency presentations and LinkedIn feeds have been dominated by the same question: What will AI do to marketing?

The answers are becoming painfully familiar. Faster content. Cheaper production. More efficient workflows. More output from fewer people. Useful? Certainly. Interesting? Increasingly less so.

While the industry obsesses over making marketing more efficient, something much more significant is happening elsewhere. The way people make decisions is changing.

Or, more accurately, people are making fewer decisions at all.

For years, marketers have understood the concept of the Zero Moment of Truth. Before buying anything, consumers would research, compare, evaluate and narrow their options before making a choice. Search engines, review sites and social platforms became the battleground where brands fought to be considered.

But what happens when the customer no longer conducts that research themselves? What happens when an AI agent does it for them?

We're entering what I believe is the ‘Generative Moment of Truth.’ The moment when a machine collapses the consideration set, evaluates the options and presents a recommendation, or simply makes the decision outright.

In that world, the question is no longer whether customers can find you. The question is whether you are considered at all and that is a profoundly different challenge.

Most marketers are still optimising for visibility. They want to rank on Google, generate impressions, increase awareness and drive traffic. But in an agentic future, customers may never see a search results page. They may never visit your website. They may never compare alternatives.

An AI system will do that work on their behalf.

And if that system cannot clearly understand who you are, what you do and why you matter, you simply won't make the shortlist.

So, clever clients need to ask themselves and their agencies different questions. 

For example: 

Would an AI confidently recommend your brand in a single sentence? If your value proposition cannot be distilled into a clear and consistent answer, it becomes difficult for any system to return it.

What would an AI say you are best for? Many organisations spend years broadening their offer and expanding their narrative. Yet machines reward clarity, not complexity. If your positioning is vague, you become harder to select.

Are you described consistently across the internet? This matters more than many marketers realise. Generative systems don't rely solely on your website. They synthesise information from reviews, articles, directories, forums, social conversations and countless other signals. Your brand is increasingly defined by the sum of what others say about you.

Do you appear in answers, or only in search results? The unit of competition is changing. Historically, brands competed for clicks. Increasingly, they will compete to become the answer itself.

Can your value be understood without visiting your website? Because, increasingly, nobody may need to.

And perhaps the most revealing question of all: If an AI had to justify recommending your brand, what evidence would it use? Not the claims in your latest campaign. Not the carefully crafted language in your annual report.

Evidence.

Customer reviews, independent coverage, expert endorsements, community discussions,  structured data and demonstrable proof.

The things you don't fully control are becoming the things that matter most.

None of this means brands should stop investing in creativity, communications or customer experience. Quite the opposite. Those activities become even more important because they create the signals that AI systems learn from and trust.

But it does mean that many organisations are focusing on the wrong problem.

They are optimising the factory while the point of sale is quietly being removed.

If you are going to succeed in the next few years, you need to understand that the future of marketing will not just be about persuading customers, it will also be about persuading the machines that advise them.