Skip to content
AdDeck
Blog/The Next Ad Break Is a Conversation
·10 min read·#ai#search#native

The Next Ad Break Is a Conversation

What happens when the technology we turn to for advice starts selling to us? A family holiday, an AI assistant and an unexpected advert.

Retro 1950s-style illustration of a woman at a café table chatting with a glowing holographic assistant, surrounded by vintage advertising billboards in a futuristic city

A holiday, an assistant and an unexpected advert

Imagine asking an AI assistant to plan your family holiday. Actually, you don't have to imagine it. That's exactly what I did this summer.

The school holidays were approaching and we hadn't booked anything. We had a rough budget in mind, but little idea of where to go or what would keep everyone entertained. Rather than spending an evening jumping between comparison sites, hotel reviews and travel agents, I asked an AI assistant for help.

Within minutes we'd narrowed down destinations, compared hotels, worked out whether half-board was better value and found activities for the children.

After talking it through as a family, one recommendation stood out. We went to the hotel's website and started booking. Then, as so often happens with family life, something interrupted us and the booking was forgotten.

By the next day, the perfect combination of hotel, room and price had gone. Kicking myself for not booking it the night before, I went back to the assistant to look for alternatives.

And there it was. Beneath its latest response, a holiday offer remarkably similar to what we'd been discussing.

Sceptical but curious, I clicked through. The company had exactly the holiday we wanted, so we booked it. At the time I was just relieved it was sorted.

But it did leave me wondering. How much of what I'd seen was genuinely helpful advice, and how much was advertising? And does it matter, if I ended up with exactly what I wanted?

From searching to asking

For years, we've understood the basic bargain of search advertising. We type something into a search engine, sponsored results appear, and we decide whether to click.

Conversational AI changes that. Instead of a handful of keywords, we explain our circumstances, ask follow-up questions and weigh up alternatives.

Someone shopping for a laptop might search for 'best laptop under £1,000'. An AI conversation could cover their job, the software they use, how often they travel and what they dislike about their current machine.

Google reported in May 2026 that queries in its AI Mode were, on average, three times longer than traditional searches. That doesn't necessarily mean stronger buying intent, but it shows how differently we use these tools.

It isn't just the amount of information, either. A search engine retrieves information; an AI assistant seems to discuss it with us. We can challenge it or ask it to think again, and it answers patiently, without the awkwardness of disagreeing with a person. That doesn't make it human, but it can feel surprisingly personal.

Advertisers have long understood the pull of personalisation. Think of Coca-Cola's hugely successful 'Share a Coke' campaign, which started in Australia in 2011 and put people's names on bottles. Conversational advertising could take that much further. Rather than tailoring a message to a broad audience, it can respond to one person's questions as they ask them.

When technology starts feeling human

I remember a former boss enthusiastically telling us he'd named his AI assistant and spoke to it every day. We even had a live demonstration in one of those company-wide meetings.

This was a while ago, when chatting to an AI still seemed odd. Looking around the room, I noticed more than a few wry expressions, mine included.

Today, it hardly seems remarkable.

I'm still not comfortable talking out loud to an AI assistant. I prefer typing, but I appreciate being able to write a long, detailed request and get a useful answer. Five years ago, typing a paragraph of requirements into a search box would have got you nowhere.

That familiarity raises a question about trust. We tend to trust things that consistently give us useful information, but trust isn't only about accuracy. How information is presented matters too.

A 2026 analysis of 32 studies, covering more than 12,000 participants, looked at how human-like chatbot traits affect people's attitudes. It found that conversational, human-like language shapes how people feel about chatbots and whether they choose to use them.

So if we're increasingly comfortable treating an AI assistant as a knowledgeable adviser, could that make us more receptive to advertising alongside its recommendations? Or might the ads have the opposite effect, reminding us that our helpful assistant is also a commercial business?

We're trusting AI at work, too

This shift isn't limited to booking holidays. Across advertising, marketing and publishing, I regularly see AI used to structure pitches, develop concepts, summarise research, refine messaging and flag claims that need checking.

But there's a contradiction. One organisation I know has editorial guidelines stating that no published content may be generated or written by AI. That seems fair to me, but it also shows a degree of mistrust: human-written content is still the benchmark for editorial credibility.

If AI is useful enough to help us research and generate ideas, why are we, understandably, more cautious about letting it produce the finished article? I suspect advertisers will face the same question as keeping consumers' trust becomes more important.

Adoption is certainly widespread. A July 2026 survey by ISBA, the UK advertisers' trade body, found that 99% of its 200 respondents were using generative AI in some way, yet only 14% said it had a significant impact on their business results. Using it doesn't automatically mean getting value from it.

Then there's judgement. A 2025 study of 319 knowledge workers by researchers from Microsoft and Carnegie Mellon University found that the more confident people were in AI, the less critical thinking they reported doing on AI-assisted tasks.

That doesn't mean AI is making us less intelligent; the thinking can shift towards checking what it produces. But when a report looks thoroughly researched or a proposal sounds convincing, do we scrutinise it as closely as a colleague's work? The risk isn't that we stop thinking. It's that we gradually question the result less.

Advertising has entered the conversation

For a while, ads inside AI assistants were talked about as a future possibility. Not any more.

In 2026, the major AI and search platforms began putting sponsored placements into conversations, matched to the context of what people are asking. OpenAI started testing ads in ChatGPT in the US in February. Google has been showing ads in its AI Mode in the US and, in October, began rolling them out to more countries, with the UK due before the end of the year.

Some of these ads sit separately from the assistant's answers, and the platforms say commercial placements don't influence those answers.

Other formats go further. In September, OpenAI began testing 'Sponsored Agents', which let you open a conversation with an advertiser's own AI after tapping on an ad. Rather than clicking through to a website, you can ask questions, explore options and have your concerns answered within the chat.

For businesses, that's an attractive proposition: a conventional ad has limited space, while an interactive one can answer questions about your particular needs. For consumers, it could be genuinely useful.

But when an ad can hold a conversation, handle objections and adapt to your questions, where does helpful information end and commercial persuasion begin? We're used to ads trying to grab our attention. We're not used to having a discussion with one.

Haven't we seen this before?

The advertising industry has been finding ways to make commercial messages feel like part of their surroundings for decades.

Newspapers and magazines have long carried advertorials. Search engines place sponsored listings alongside organic results. Native advertising takes on the look and feel of the sites it appears on. The goal is familiar: reach the right audience with a relevant message, ideally without disrupting the experience.

For anyone in publishing, the line between editorial independence and commercial content is hardly a new debate. An advertorial may look like an article, but it's still advertising, and the reader needs to know that.

Conversational AI doesn't remove that responsibility. What it changes is the interaction. An advertorial can set out the benefits of a product. A conversational ad can answer questions, address concerns and adjust its answers as the conversation goes on.

That isn't necessarily deceptive. A clearly identified representative of a business can give genuinely helpful information. But a helpful representative is still a representative, and a system with the patience and fluency of an independent adviser could make that easier to forget.

Trust is more complicated than we think

It's tempting to assume people are trusting AI more simply because they're using it more. The research suggests otherwise.

In 2025, Pew Research Center found that only 6% of US adults who'd seen AI-generated summaries in search results trusted them a lot, while 46% said they had little or no trust in them. That isn't about AI advertising specifically, but it makes an important point: using something doesn't mean believing everything it tells us.

We know online reviews can be manipulated and sponsored search results serve commercial interests, yet convenience often wins. It doesn't make us careless, but it can make us less inclined to question what's in front of us.

We've already seen scammers exploit sponsored search results, steering people towards convincing but fraudulent websites. Similar risks could emerge in conversational AI, in a more personal package. There's no evidence it's happening yet, but it's a reason to stay alert.

As users of new technology, we need a healthy dose of curiosity and scepticism. Not distrusting everything, but not accepting everything either.

The opportunity and the responsibility

It would be easy to paint conversational advertising as yet another intrusion, but that overlooks its potential. Advertising is useful when it connects people with something genuinely relevant, and my holiday is a reasonable example: whatever questions I have about how the offer appeared, it helped me find what I wanted.

As someone working in an industry that depends on commercial success, I want advertising to grow and succeed. But not at the expense of consumer confidence, brand reputation, product standards or editorial independence.

Relevance alone isn't enough. People need to know when they're dealing with advertising, whether commercial interests are shaping the information they get, and how their data is being used.

There's an important difference between using the context of a conversation to choose an ad and giving advertisers access to someone's private conversations. The platforms say they keep the two apart, but those safeguards shouldn't simply be taken on trust. And a sponsored placement can be clearly labelled and still sway a decision.

The responsibility is shared: platforms need to be transparent, advertisers need to respect the space they're entering, and consumers need to use their own judgement.

Where does the conversation end?

Looking back at my holiday booking, the offer did exactly what good advertising is meant to do. With more thought, I'd probably have looked harder at how it appeared, but in the moment it felt like a godsend!

The transaction wasn't the interesting part, though. It was everything before it. I'd already explained our circumstances, said what we wanted and weighed up the assistant's recommendations. By the time the offer appeared, it was joining a conversation I'd already invested in.

Perhaps that's what makes conversational advertising different. We're getting used to asking technology for advice and letting it help shape our decisions, at home and at work. Now advertisers are finding ways to join in.

I don't think that's necessarily a bad thing. Advertising has always evolved alongside the media that carries it, and there are real opportunities to make commercial messages more relevant and useful.

But trust is worth protecting, especially when we're giving these systems so much personal information. Some AI assistants can already see our emails, our documents, our questions and even our innermost thoughts. It's a veritable treasure trove.

I recently finished reading The Space Merchants, Frederik Pohl and C. M. Kornbluth's 1952 satire, in which advertising agencies wield enormous power and almost every part of life revolves around consumerism. I don't think we're heading for that kind of dystopia. The challenge is keeping clear the line between technology helping us make a decision and a business trying to influence it.

Because the next ad break may not interrupt the conversation at all.

It may become part of it.

Sources and further reading

On AdDeck:

Elsewhere:

← Back to the Blog