I interviewed AI about the industry's future

In an attempt to break the mould of future-gazing listicles about trends that are probably already on your radar, I thought I’d experiment with publishing a chat I had with an AI model - completely unedited.

For full disclosure, I briefly briefed a model on this idea beforehand - of which I was totally shocked to find out that it thought it was a good idea.

Here goes ...

Conor: Let's start off pretty simple - and please limit your answers, we don't need blocks and blocks of drivel - if we're sitting here in September 2026, what will look noticeably different about advertising 18 months from now?

AI: By early 2028, the biggest visible change will be that AI stops looking like a separate tool and starts looking like part of the advertising infrastructure.

You’ll see AI agents handling more research, planning, production and optimisation; creative teams generating and testing far more variations; and consumers increasingly discovering brands through AI interfaces rather than traditional search.

The important shift is from AI helping people do advertising to AI actively doing parts of advertising.

Conor: Is this not already happening? I feel like all you've really told me is that there will be more of everything.

AI: Yes - that's fair. Most of those things are already happening in 2026, so “more of it” isn't a useful prediction.

I'd revise the answer to focus on what changes category, rather than what gets better: 

By early 2028, the noticeable change may be that some things we currently think of as “advertising” no longer look like advertising at all. 

AI agents could increasingly sit between brands and consumers, influencing what people discover, compare and buy. Agencies may also start building AI systems as part of campaigns - not just using AI to make the campaign.

So the interesting question isn't whether agencies will use more AI. They already do. It's whether AI starts changing the basic mechanics of how a brand gets chosen.

Conor: Okay, well, you're kind of cheating there by revising your answer but I'll move on.

Where specifically are you sourcing this information on “AI changing the basic mechanics of how a brand gets chosen” in the future?

AI: The clearest source is McKinsey’s June 2026 report, “The agentic advertising economy: From attention to action.” It argues that as AI increasingly mediates discovery and purchasing, advertising success could shift towards being “surfaced, recommended and selected” by AI systems.

There’s also recent Accenture research, based on 25,590 consumers across 16 countries, finding that 74% would trust a personal AI agent more than their best friend to make a purchase on their behalf.

So that particular prediction is an inference from current evidence, rather than something I can state as a certainty about 2028.

Conor: How will this affect agencies in particular?

AI: It could push agencies upstream. 

If AI systems increasingly influence which brands consumers consider, agencies may spend less time optimising for clicks and impressions and more time making brands understood, distinctive and recommendable to AI systems. 

That could create new work around AI visibility, agentic commerce, brand data and conversational experiences - while putting pressure on some traditional media and performance-marketing models. 

The bigger risk for agencies is that the AI becomes the intermediary between the agency's work and the consumer.

Conor: Expand on this risk you mention.

AI: The risk is that AI becomes the gatekeeper between a brand and its audience. 

Today, an agency can create a campaign, buy media and put the work directly in front of consumers. If people increasingly ask an AI what to buy, where to go or which brand to choose, the AI could increasingly decide which brands get considered in the first place.

That potentially shifts power away from agencies and media owners towards the platforms operating those AI systems. Agencies would need to understand not just how to win human attention, but how to make a brand visible and persuasive inside an AI-mediated decision.

And there's a commercial question too: if the AI handles discovery, comparison and even purchase, where does the traditional agency role sit in that journey?

 Conor: This article by Sean Betts, Chief AI & Innovation Officer at OMG UK, explores a new and emerging digital category called ‘Paid Organic.’ As opposed to paying or earning the right to appear, this sees brands pay for the opportunity to influence how they appear. Stepping beyond GEO and paying AI companies to surface information, how might this new category grow in the next 18 months?

AI: I think Paid Organic could evolve into something much broader than simply paying AI platforms or doing GEO.

Over the next 18 months, I’d expect publishers, retailers and other trusted information sources to start creating machine-readable commercial inventory: sponsored product data, specifications, offers, FAQs and expert information that AI agents can discover and evaluate.

That could create a new layer of media planning: agencies buying influence over the information environment, rather than buying impressions.

Sean Betts’ TIME magazine example is a useful early signal: brands can pay to put information into pages designed for AI crawlers, while the AI remains responsible for deciding whether that information actually affects its answer.

The really interesting development would be if agencies start planning “AI consideration” budgets alongside paid media and SEO - essentially paying to make sure a brand is in the pool of information an AI considers, without paying for the final recommendation.

That is where “Paid Organic” could become a genuinely new category rather than just a new name for GEO.

Conor: Putting new categories to one side, what about the new AI tech that might emerge and impact adland? From my POV the disruption started with gen AI tools and now its agentic AI. What’s next?

AI: Yes. If we treat the progression as generative AI → agentic AI, I think the next genuinely interesting layer is AI that can perceive and act in the physical world, rather than just manipulate information.

For adland, I'd watch three areas:

1. AI-native interfaces / “computer use”

AI agents will increasingly be able to operate software like a person - opening browsers, using creative tools, moving through ad platforms and executing workflows. The significance is that the AI isn't just generating an output; it can operate the tools that make advertising happen.

2. World models + spatial AI

AI is getting better at understanding video, 3D environments and physical spaces. Over the next 18 months, that could become interesting for OOH, retail, experiential and product advertising: AI that can understand what's happening in a shop, street or event space and generate or adapt content accordingly 

3. Personal AI agents

This is potentially the biggest one for advertisers. Instead of brands having AI agents, consumers will have them. Your personal agent could search, compare products, negotiate prices, book things and eventually make purchases on your behalf.

That creates a very different advertising problem: how do you persuade someone's AI rather than the person themselves?

Of those three, I'd put personal AI agents at the top of the 18-month watchlist. If they take off, they could have a more profound effect on advertising than another generation of better image or video models.

And there's an interesting fourth possibility: AI + robotics. Probably less immediate for mainstream advertising, but once AI agents can perceive the physical world and control machines, you start getting into genuinely new territory for retail, experiential and out-of-home.

Conor: I know you’d like me to keep asking questions forever, but for the sake of brevity this will be my last question. By 2028, how much closer will AI be to creating a truly innovative and effective campaign by itself - whatever the form or brand - and how can agencies keep ahead of this? 

AI: By 2028, I expect AI to be much closer to producing a genuinely good campaign end-to-end, but I wouldn't expect it to reliably originate the kind of culturally resonant, strategically sharp idea that changes how people think or behave. 

The technology will become increasingly capable of connecting insight → strategy → idea → execution → optimisation. The harder problem is knowing which problem is worth solving, understanding culture and human behaviour, and having the judgement to recognise an idea that is genuinely new rather than merely plausible. 

For agencies, that means the answer isn't trying to outrun AI at production. It's getting better at the things AI finds hardest to replicate: distinctive thinking, cultural intelligence, taste, strategic judgement and relationships with clients. 

And I'd be wary of assuming AI will always remain weaker at those things. The agencies that stay ahead will be the ones testing that assumption constantly, rather than building their future around what AI can't do today. 

My human conclusion

Like some of the interviews I’ve conducted throughout my B2B journalism career, this chat involved quite a few safe answers with a few interesting nuggets of information sprinkled throughout.

AI “consideration budgets”, spatial AI and AI + robotics piqued my interest - not to mention the advice that agencies should avoid writing off AI’s independent creative potential.

I think we’re still a few years off brands paying media publications read by personal AI-powered robots to ultimately influence what packet of crisps I might like my robot to order before I’ve even thought about it.