Can Companies Pay AI to Recommend Their Products?
As AI starts choosing what we buy, brands may have to compete for something more valuable than our clicks: a place in the AI's shortlist.
For decades, advertising has worked in roughly the same way.
A company pays to get your attention.
You see an ad on television, Google, Instagram, YouTube, or a website. You become interested. You search for the product, compare your options, and eventually decide what to buy.
The important part is simple:
You make the choice.
But what happens when you stop choosing?
Imagine telling an AI assistant:
“Find me a comfortable pair of black running shoes under $150.”
Instead of opening ten browser tabs, reading reviews, comparing prices, and visiting several stores, the AI could do much of that work for you.
It might search thousands of products.
Then show you three.
And suddenly, a very old advertising question becomes a very new one:
Why did the AI choose those three?
Was it the price?
The reviews?
The specifications?
The reputation of the brand?
Or did someone pay to be there?
That question could become increasingly important as AI moves from answering questions to helping people shop.
AI is already becoming a shopping assistant
This isn't purely hypothetical.
ChatGPT can already help users discover and compare products. OpenAI has also been expanding the infrastructure that allows merchants to provide product information for AI-powered shopping experiences.
But there is an important distinction.
According to OpenAI, the products shown in ChatGPT shopping results are selected independently and are not advertisements. OpenAI also says product rankings are not determined by partnerships with merchants.
Advertising, where it appears, is treated separately from the organic answer or product result.
That means the current model is not simply:
Brand pays money → AI says its product is the best.
But the advertising industry is already moving toward AI.
And that's where things get interesting.
Advertising is entering AI search
Google offers a glimpse of what this could look like.
Its AI-powered search and shopping experiences can provide organic product recommendations while also displaying clearly labeled sponsored content.
That distinction matters.
An AI might recommend several products because it considers them relevant to your request, while a retailer can separately pay for a sponsored placement.
In other words:
Recommendation and advertising can exist in the same AI experience without being the same thing.
That sounds familiar.
Google Search has operated on a similar principle for years: organic results and paid advertisements appear on the same page but are supposed to remain distinguishable.
AI, however, introduces a new problem.
Traditional search gives us a long list of links.
AI often gives us an answer.
And sometimes, only a few choices.
Being one of three could matter more than being one of 3,000
There are thousands of competing products online.
In the Google era, your goal might have been to appear near the top of a search result.
Companies built entire marketing departments around this problem.
SEO emerged partly because businesses wanted an answer to one question:
How do we rank on Google?
But now imagine that a consumer doesn't search Google at all.
They simply ask an AI:
“Find the best running shoes under $150 for someone who walks a lot.”
The AI analyzes the options and recommends three products.
For the brands selling the other thousands of shoes, a new question appears:
How do we become one of those three?
That could eventually become one of the most valuable questions in digital commerce.
The first battle may not be advertising. It may be information.
Before an AI can recommend a product, it needs to understand it.
How much does it cost?
Is it in stock?
What sizes are available?
How long does shipping take?
What are its specifications?
How do customers rate it?
Is the information current?
Companies including OpenAI and Google are already building systems that allow merchants to provide structured product information to their platforms.
That might sound boring.
It isn't.
For the past two decades, companies have spent enormous amounts of money making websites understandable and attractive to humans — while also making them easy for search engines to index.
The next step may be making products understandable to AI agents.
The old question was:
How do we get Google to find us?
The new question could be:
How do we get AI to recommend us?
AI may not stop at recommendations
This becomes even more important if AI begins completing purchases.
Payment companies are already preparing for this.
Visa and Mastercard are developing infrastructure designed for what is often called agentic commerce — systems in which AI agents can help discover products and, with the user's authorization, participate in completing transactions.
In 2026, Mastercard announced the completion of live agentic transactions in South Korea, including an example in which an AI agent helped arrange and pay for transportation from Incheon International Airport to central Seoul.
The shopping journey we've known for decades looks something like this:
Search → Click → Browse → Compare → Checkout
Agentic commerce could compress that into something much shorter:
Ask AI → AI chooses → AI buys
If that happens at scale, the economics of online advertising could change with it.
What happens to brands when humans stop browsing?
Advertising has traditionally been designed to influence people.
Beautiful photography.
Celebrity endorsements.
Music.
Packaging.
Emotional stories.
Brand identity.
All of these things exist because humans don't make purchasing decisions using spreadsheets alone.
But an AI agent is different.
It doesn't admire a billboard on the highway.
It doesn't feel nostalgic when it hears a song.
And it doesn't necessarily care whether a website looks luxurious.
Instead, an AI shopping system may place greater weight on things such as price, availability, product data, reviews, delivery time, reliability, and the specific instructions given by its user.
That creates an unusual possibility.
Companies may eventually have to market to two audiences at once:
The human who wants the product.
And the AI that decides which products the human sees.
So, can companies pay AI to recommend their products?
Today, the answer is more complicated than a simple yes.
OpenAI says ChatGPT's organic shopping results are not advertisements and are not ranked according to merchant partnerships.
Google also distinguishes between organic AI recommendations and sponsored placements.
So paying for an advertisement is not the same thing as secretly buying an AI recommendation.
That distinction will matter enormously for trust.
If someone asks an AI for “the best laptop for college” or “the safest car seat for my child,” they need to know whether the answer is an independent recommendation or a paid placement.
But the larger shift is already becoming visible.
Businesses are beginning to think not only about how humans discover their products, but also about how AI systems discover, understand, and eventually choose them.
For twenty years, one of the biggest battles on the internet was getting the click.
The next one may be getting into the machine's shortlist.
BEYOND THE OBVIOUS.
Sources
1. OpenAI
OpenAI — Product Discovery in ChatGPT
2. Google
Google — AI and Commerce Advertising
3. Mastercard
Mastercard — AI Agent Completes a Real-World Purchase