‘Ring the alarm bells for adland’, a recent headline declared, the suggestion being the advertising industry is ailing. I don’t think it’s sick, however. It’s already been deemed dead. And yet despite this, millions are still attending its wake. 

You saw it on the Croisette in June, with panels about AI efficiencies delivered with the calm confidence of people discussing a minor software upgrade rather than a technology that will quietly erase one third of the room. 

With how everyone simply nodded along to talk of cost saving, optimisation, scale, too. And with how very few acknowledged the more uncomfortable truth that the thing being described, with smiles and stage lights, was – is – coming for their jobs.

Advertising didn’t collapse overnight. But it has hollowed out itself slowly, with consolidation rewarding size over judgment and lengthy procurement teaching brands to buy labour like bulk protein powder.

Performance metrics trained agencies to tie themselves up with relentless data and, somewhere along the way, the industry mistook this for relevance.

AI is not the cause of this, but it is the accelerant. And if advertising continues to define itself as a machine for producing assets more cheaply and more quickly and more predictably, then machines will win the day. 

Or perhaps they already have, and they will be better at it than us within about 12 months.

Either way, the only interesting question left is not whether advertising survives as it is – because it won’t, it’s whether it has the courage to evolve from what it was to become something else. So, here is a five-point plan for how it might.

Firstly, agencies need to stop treating AI as a junior strategist and start treating it as infrastructure to their entire business – and, as such, something alongside which they co-exist. It is a mistake to ask AI to make ads as that is the lowest value use of the technology.  Instead, use AI to clear the weeds and the undergrowth in order to get properly under the skin of research synthesis, pattern mapping, scenario modelling – the goal being not machine creativity but human time. 

The more thinking the machine absorbs, the more judgement the human can then supply – and must. If your agency is using AI to replace junior thinking rather than to create space for senior thinking, you are already in trouble.

Second, advertising must reclaim taste as its core economic value.  For 20 years, taste was treated as subjective, indulgent and even unprofessional in some places. Data was the grown-up in the room and the god particle. But in a world where everything is generate-able, taste becomes the scarcest resource of all. And knowing what to ignore and when not to speak and which references matter and which don’t is an amazing magical gift. 

Agencies should be training and positioning their smartest people not as campaign leads, but as chief taste officers: editors of cultural direction, guardians of coherence, people who can say ‘No thanks, this doesn’t really belong’. 

Machines can generate options by the truckload, but only humans can make such smart exclusions to make all the difference.

Third, agencies should stop selling campaigns and start selling their judgment. Brands are drowning in possibility with every platform, every format, every creator, every tool, every trend, all at once. What they lack is not access, but filtering. 

A future agency should be less factory, more editorial board. A focus on what we should stand behind rather than what we should make will make us a part of the fabric of culture rather than just noise.

Fourth, consolidation should be treated as a sort of cultural surgery rather than financial engineering to make the CFO happy. Merging balance sheets without merging belief systems just produces organisations with no reason to be. 

If agencies are going to shrink in number, then they must deepen in their identity. Fewer agencies, yes, but each with a sharper point of view, a clearer philosophy, and a definitive and recognisable way of seeing the world. 

Brands do not need more interchangeable partners, but they’ll definitely want stronger ones who are willing to disagree.

Finally, the industry needs to accept that advertising is no longer the centre of culture – and stop behaving like it is. 

The job now is not to interrupt culture but to tend it by investing in and building long sweeping narrative arcs, investing in formats and communities, not just thinking of brash loud moments. 

This is slower, harder, and far less compatible with quarterly CMO dashboards. But it is where trust is rebuilt between the person on the street and the clever souls behind agency walls.

To use the old Ford example, advertising will not survive by becoming faster horses in an already autonomous world but it might just survive by becoming something machines cannot be with judgment, taste, cultural literacy and moral confidence.

That is the real work, and everything else can be automated.