Ads in the AI world

Ads are coming to AI chat, and the argument already sounds familiar: companies need revenue, users dislike ads, and everyone promises the ads will be “relevant.”

But AI chat is not a search results page, a social feed, or a shopping grid. It is where people ask for advice, compare options, draft sensitive messages, and think out loud.

That makes ads in AI chat a UX problem before it is an ad problem.

From pages to conversations

In search, ads sit next to answers. In social, ads interrupt attention. In AI chat, ads risk sitting next to trust.

A sponsored result on Google is familiar. Some links are organic, some are paid, and the user still chooses what to click. Chat feels different. The answer arrives as a recommendation, not a list of links. The interface feels less like browsing and more like asking someone for help.

So the question is not just, “Is this ad relevant?” It is, “Did this ad change the answer?”

Stronger intent, weaker discovery

AI chat may make known-intent advertising better.

A search query like “best travel stroller” is a rough signal. A chat prompt can be much richer: “Find me the best travel stroller under $300 that folds small, works for a tall parent, and is easy to carry through an airport.” That is high-quality intent. The user has given the category, budget, context, constraints, and decision criteria.

But advertising is not only about matching existing intent. It is also about discovery.

Browsing creates discovery through the messy middle: searching, clicking, scanning reviews, comparing products, noticing unfamiliar brands, and changing your mind. AI collapses that journey into a neat answer.

That is efficient, but it narrows the user’s field of view. If the assistant gives three recommended options and the user picks one, what happens to the adjacent products, challenger brands, and unexpected alternatives that would have appeared during browsing?

AI may understand intent better than search, but it may reduce the surface area for discovery.

Where AI ads can add real value

The clearest opportunity for ads in AI chat may be comparison shopping.

If a user wants the cheapest price on a product or to compare vendors by price, shipping, warranty, returns, or availability, sponsored links can sit alongside the comparison without changing the recommendation itself.

This is where AI ads could feel useful. The assistant explains the tradeoffs; the commercial layer helps users act.

But the bigger question is what happens if AI becomes the marketplace itself.

Search engines mostly send users to the web. AI assistants could become a new discovery and transaction layer, helping users find products, compare options, and complete purchases without leaving the conversation.

That shifts the power dynamics. Will every retailer have equal access to recommendations? Will merchants need to pay commissions to participate? Will visibility depend on joining a platform’s commercial ecosystem?

The risk is that comparison shopping stops being about finding the best option and starts being about controlling which options are easiest to find. The App Store is the closest analogy: convenient and trusted, but highly influential over discovery and transactions.

The key is separation. Recommendations should be driven by relevance and quality, not commercial arrangements. Commerce can sit beside the recommendation as an action layer, but users should be able to tell where advice ends and marketplace incentives begin.

And this depends on infrastructure. Search engines spent years building ad networks, merchant relationships, product feeds, attribution systems, and retailer trust. AI companies may end up building something even larger: not just an ad system, but a new layer of product discovery and commerce.

So comparison shopping is both the clearest opportunity and the clearest constraint. Good AI ad UX is not only an interface problem. It is an ecosystem problem.

Not every chat is commercial

The harder problem is that AI chat is messy.

People do not only ask chatbots what to buy. They ask about symptoms, parenting, money, work conflict, grief, career anxiety, relationships, and big life decisions.

Sometimes they are shopping. Sometimes they are thinking. Sometimes they are vulnerable.

Treating all of these moments as monetizable intent is where the experience breaks.

A chatbot depends on the feeling that the assistant is working for the user. The moment users suspect the assistant is also working for someone else, the relationship changes.

The UI principle: Separate the layers

If AI chat collapses browsing, the interface has to recreate the useful parts of browsing without bringing back all the noise.

The design challenge is to separate three layers:

• advice
• discovery
• commerce

The recommendation should stay clean. Sponsored content should not be embedded inside the answer as if it were part of the reasoning.

Discovery should be user-led: “show cheaper alternatives,” “show lesser-known brands,” “compare premium options,” “what am I not considering?” or “show more like this.”

Commerce should be visibly separate: offer cards, vendor links, coupons, booking modules, or availability panels.

The best AI ad interfaces may look less like ads inside chat and more like action surfaces around chat.

Ads should live near decisions

The best version of AI advertising is not a product awkwardly inserted into an answer. It is a contextual offer at the edge of an action: a hotel deal after choosing a destination, a software trial after comparing tools, a coupon after deciding what to buy.

Ads should live near decisions, not inside advice.

That may sound like a small interface detail, but it is the whole product strategy. AI products are built on the feeling that the assistant is aligned with the user. Ads should not blur that alignment.

AI companies have a real business problem. Free AI is expensive to run. Subscriptions will not reach everyone. Ads could subsidize broader access. But the question is not whether AI companies can show ads. Of course they can. The question is whether they can show ads without making the answer feel bought.

Ads in AI chat may be inevitable. Trust is not.