For years, digital advertising worked on a simple principle: people searched, saw results, clicked, and decided. That journey still exists, but it is being compressed. Increasingly, the question goes straight to an AI assistant, and the decision starts being shaped by an answer that is ready, summarised, and persuasive.

This is where GEA comes in. GEA stands for Generative Engine Advertising, advertising designed to appear within AI driven search and answer experiences. Instead of competing only for the click, you compete for a place at the exact moment the AI is guiding the choice.

And this has moved beyond theory. In January 2026, OpenAI announced it plans to begin testing ads in ChatGPT, with clear rules and principles on how this will work.

What OpenAI has already confirmed about ads in ChatGPT

The most important information, because it is official, comes from OpenAI’s own channels: ads are not “switched on” for everyone yet, but OpenAI says it intends to test ads in the coming weeks, early 2026, in the United States, for adult users on the Free and ChatGPT Go plans.

According to OpenAI, the initial format will appear at the end of responses when there is a sponsored product or service that is relevant to the conversation. The separation between the ad and the organic answer will be clear, with explicit labelling.

There are three more points that matter to any brand because they define the playing field.

OpenAI states that ads do not influence ChatGPT’s answers. The answer is optimised to be useful, and the ad appears separately.
OpenAI says conversations remain private from advertisers and that it does not sell user data to advertisers. If a user interacts directly with an advertiser through an ad, the advertiser only sees what the user chooses to send.
OpenAI says it will give users control, including the ability to disable personalisation, clear data used for ads, and dismiss ads with feedback. It also states it will not show ads to minors and will block ads near sensitive topics such as health, mental health, and politics.

For brands, this means something very concrete: ad inventory is born right next to the context of the conversation and the moment of intent. That is a major mental shift away from “keywords” and “results pages.”

What Sam Altman has said about advertising in AI and why it matters

The topic gained even more attention because Sam Altman had previously expressed discomfort with combining ads and AI, describing it as “unsettling” and framing ads as a “last resort” in public remarks linked to Harvard.

The practical point for the market is not the drama. It is understanding the direction: OpenAI is building a model that attempts to fund access through advertising while keeping a set of principles around separation between the answer and the ad, and user control.

It also means recognising the public and regulatory pressure that will land on these implementations. A clear example was Senator Ed Markey’s reaction, warning about risks such as “camouflaged” advertising, emotional manipulation, and privacy concerns in chatbots, even when early safeguards exist.

This will influence formats, restrictions, transparency, and what will be allowed. Brands that prepare seriously will have an advantage.

What GEA looks like in practice

GEA is not “running Google Ads inside ChatGPT.” The mechanism is different, because the unit of intent is the conversation, not a single isolated query. Advertising stops being just a box next to a list of results and starts living alongside recommendations and decisions.

You will see this concept described in several ways: ads in generative engines, sponsored answers, ads in AI responses, conversational search advertising. “GEA” is a useful label for organising the discipline, and it is already being adopted by parts of the digital ecosystem.

What changes for a brand when advertising moves “inside the answer”

The main shift is this: the message stops competing only for attention and starts competing for trust. In an AI environment, the user wants a choice that is quick and defensible. That changes what works in advertising.

Generic creatives and vague promises tend to fall away. What tends to rise is immediate usefulness: specificity, proof, clear comparisons, simple policies, public credentials, verifiable reviews, real guarantees, and language that resolves doubts.

In its own advertising post, OpenAI says it wants to create experiences that are more useful than traditional ads, and it even mentions the possibility that a user may see an ad and then ask follow up questions to make a purchase decision. That is a major detail, because it pushes the ad into a conversational format.

The bridge between GEO and GEA: brands that “enter the answer” win twice

Here is the strategic advantage many people will overlook early on: GEO (Generative Engine Optimization) and GEA (Generative Engine Advertising) reinforce each other.

GEO increases your likelihood of being cited organically, as a source the AI chooses. GEA strengthens your sponsored presence when there is ad inventory and when the context fits.

When a brand has enough clarity and proof to show up organically and, at the same time, has campaigns and assets ready for sponsored inventory, the user’s perception tends to be authority. And that has a real impact on selection rates.

The typical mistake is waiting for the “official launch” in your country. When the product arrives, the early winners have already tested messaging, built assets, and refined measurement. Preparing for GEA starts with foundational work that applies across platforms.

1. Define the transactional questions that precede purchase

GEA lives in the moment the user is deciding. That is why the raw material is real questions with real intent, such as “what is the best option for…”, “what offers the best value”, “what is the difference between…”, “is it worth it”, “how much does it cost”, “how long does it take”, “what do I need for…”.

A brand that can answer these questions in a direct, concrete, verifiable way is more likely to convert, with or without an ad.

2. Build assets that support a short, defensible answer

AI works well with information that can be summarised without losing accuracy. That favours pages and content with clean structure, objective definitions, examples, public proof, and unembellished language.

In an advertising context, this matters because the user reads the ad with a “I want to solve this now” mindset. If the ad promises and the click does not deliver, costs rise and trust drops.

3. Enrich feeds and data with context, not only technical attributes

In ecommerce and service catalogues, this will be a major competitive advantage. One interesting market perspective, from performance consultancies and specialists, is that feeds need to move beyond purely technical attributes and include usage context, benefits, and intent signals, so they are “readable” by conversational engines.

In practice, that means descriptions that explain who it is for, in which scenarios it makes sense, what the trade offs are, and what proof supports the choice. This applies to products and services.

4. Adapt copy to conversational logic

Many ads are still written to grab attention in a scroll. In conversational environments, the job is different: the text has to fit a decision and survive follow up questions.

A good rule is to write as if the user will respond with “okay, but why?” and “does it fit my case?” The ad and the landing page should anticipate that with clarity.

5. Prepare measurement for AI driven traffic and conversions

A meaningful part of the advantage will be analytics. You need to separate visits coming from AI experiences, measure quality, and understand whether the user is closer to purchase than in other channels.

That starts with basics done well: consistent UTMs, clearly defined conversion events, analytics validation, and a reading focused on intent, not only volume.

6. Brand, privacy, and trust stop being “soft” and become performance drivers

When a channel behaves like an assistant, users project trust onto it. That raises expectations around transparency, labelling, and data use. OpenAI is positioning advertising around principles of privacy, control, and separation from the answer, and there is already public scrutiny around manipulation and targeting risks in chatbots.

Brands that play it clean and build real proof will run more sustainable campaigns. Brands that look for shortcuts will face more backlash, tighter restrictions, and higher costs.

What comes next

The most likely scenario is gradual evolution: starting with simple formats, clearly separated, with many restrictions and limited tests, then expanding as there are signs of acceptance and usefulness.

OpenAI is already framing ads as a way to support access and pricing, keeping ad free tiers and reinforcing the argument of “useful and relevant.”

From the market side, there are voices arguing that advertising will become an inevitable part of the AI “recommendation channel,” and that players with mature ads ecosystems will have an integration advantage.

For a brand, the relevant question is operational: when the decision happens inside the answer, is your business ready to be the easy choice?

How ONTAG is already approaching GEA

At ONTAG, the first AI management agency in Portugal since 2024, we see GEA as a natural extension of the transition already happening in search: less dependence on the click, more dependence on presence, proof, and clarity at the moment of decision.

Our work starts by mapping the questions that truly drive revenue in your sector, turning them into assets the AI can use and the user can validate, and preparing campaigns and measurement for when sponsored inventory becomes available at scale. When the channel opens, the brand already has the language, proof, and structure to show up consistently.

If you want to take this seriously, start with a simple question: what is the sentence your customer types into an AI assistant right before buying what you sell?

The day you answer that well, you have already started winning the Generative Engine Advertising game.