The Assistant Becomes the Ad Platform

OpenAI did not go to Cannes Lions only to introduce itself to the advertising industry. It went there to make a commercial claim:

ChatGPT is not merely a tool used by hundreds of millions of people. It can become a new advertising platform.

The Financial Times reports that OpenAI made its debut at Cannes Lions this week, pitching its nascent ChatGPT advertising business and its Codex programming tool to marketers and agencies. The company occupied a villa along the harbor, placing itself physically and symbolically among the advertising and ad-tech companies that have long shaped the digital media market. Dave Dugan, the former Meta executive who now leads OpenAI’s global advertising team, told the FT that OpenAI is bringing “something brand new to the market.”

That claim is ambitious, but not empty. OpenAI is trying to build a large advertising business before a planned public listing, and it is doing so at a moment when the economics of AI are becoming harder to ignore. The FT reports that OpenAI spent $34 billion last year, with costs rising faster than revenue. Advertising offers a way to monetize the enormous free and low-cost usage of ChatGPT without forcing every user into a higher subscription tier.

The advertising industry should read this as more than a monetization update. OpenAI is not simply adding sponsored messages to a popular product. It is trying to move advertising into the assistant layer, where users ask questions, compare options, make plans, and signal intent before they reach search results, social feeds, retail marketplaces, or brand websites.

If that works, the next major ad platform may not look like a search box or a feed. It may look like a conversation.

The Search Budget Is the Target

OpenAI’s ad pitch is aimed directly at one of the most valuable businesses in technology: search advertising.

According to the FT, Dugan said OpenAI is attempting to challenge Google’s dominance in search ads. That is the correct competitive frame. ChatGPT ads are not merely competing with display ads, social placements, or generic sponsored content. The larger prize is commercial intent.

Search advertising became powerful because users tell Google what they want at the moment they want it. A person searches for flights, insurance, software, skin care, sneakers, restaurants, or tax help. That query becomes a commercial signal, and Google turns the signal into an auction.

OpenAI’s argument is that ChatGPT can see intent earlier and in richer form.

Dugan told the FT that about one-fifth of ChatGPT queries have direct commercial intent. He said ads for travel, retail, health and beauty, and financial services products are performing best.

That statistic is the heart of the story. If one-fifth of ChatGPT usage has direct commercial intent, ChatGPT is not only an information product. It is a pre-search commerce environment. Users are not always asking for facts. They are asking for help choosing, planning, comparing, narrowing, and acting.

That creates a different kind of inventory. A search engine usually captures the query after the user has turned need into keywords. A chatbot can participate in the process before the need is fully organized. The user may begin with uncertainty, ask follow-up questions, refine constraints, and only later become ready to click, buy, book, or contact a provider.

For advertisers, that is attractive. For Google, it is strategically uncomfortable. For agencies, it raises a practical question: where does media planning begin if the assistant becomes the first place where commercial intent is shaped?

Ads Inside an Assistant Are Not Ordinary Ads

OpenAI says ads are labeled as sponsored, appear below regular answers, and do not affect ChatGPT’s responses. Ads are currently shown only to Free and Go users, with Go priced at $8 per month. OpenAI has also said that ads are intended to support broader access to ChatGPT and that higher-priced tiers remain ad-free.

Those details are important, but they do not fully resolve the trust issue.

An ad inside or near a chatbot conversation carries a different meaning from an ad on a search results page. Search has trained users to expect a mixture of organic and sponsored results. Social platforms have trained users to expect commercial interruption inside a feed. ChatGPT still sits in a more ambiguous category. Many users treat it as a helper, adviser, tutor, analyst, planner, or problem-solving interface.

That means the environment itself gives ads a different kind of proximity. Even if a sponsored unit is separated from the answer, it appears inside a system designed to assist the user. The commercial message is placed near a tool that has borrowed authority from usefulness.

This is why conversational advertising cannot be governed only by old assumptions about ad placement. The issue is not only whether the label is visible. The issue is whether users understand when the assistant is helping, when the platform is selling, and how those two functions are kept apart.

OpenAI’s public position is that advertising does not influence answers. That commitment will become central to the company’s credibility. If users start to suspect that sponsored content changes the assistant’s recommendations, the trust cost could be severe. If advertisers believe the separation is too strict to create performance, the revenue opportunity becomes harder to justify.

The commercial opportunity and the trust constraint are locked together.

Cannes Was Also a Platform Pitch to Agencies

OpenAI’s Cannes presence was aimed at marketers, but agencies were just as important to the room.

The FT reports that OpenAI used the event to pitch both ChatGPT ads and Codex, its programming tool. The Codex demonstration showed brands how they could create apps that help users build their own advertising campaigns. OpenAI executives played down the idea that this would replace traditional agency work, but the signal was still clear. OpenAI wants to be relevant not only to media buying, but also to how marketing work is produced, automated, and distributed.

That creates a complicated agency story.

On one side, agencies get a new platform to advise on. Clients will need help understanding where ChatGPT ads fit, which categories should test early, what creative belongs in a conversational environment, and how brand safety should be managed. The major holdcos will want early access and influence because new platform rules often become difficult to change once they harden.

On the other side, OpenAI is building direct infrastructure. The company has launched a self-serve platform in the United States and other English-speaking markets for small and medium-sized businesses. OpenAI has also announced a beta Ads Manager, CPC bidding, expanded measurement tools, and buying partnerships with major agency and ad-tech players.

That creates the familiar digital-platform pattern. Agencies help clients enter a new channel. The platform controls the user environment, the ad delivery logic, the measurement system, and the product roadmap. Over time, the operational buying layer becomes easier, while the platform’s strategic power grows.

The agency opportunity is therefore not simply to buy ChatGPT inventory. The stronger opportunity is to help clients understand AI-mediated customer journeys. If agencies reduce this to another dashboard, they will miss the structural shift.

Brand Safety Moves Into the Conversation

Brand safety has usually focused on where an ad appears. Conversational AI changes the question. Brands now have to ask what the user was asking, what the assistant had just said, how sensitive the context was, and whether the sponsored placement could appear to benefit from advice-adjacent authority.

The FT reports that travel, retail, health and beauty, and financial services are among the stronger-performing categories. That mix shows both the promise and the difficulty of the format. Travel and retail are relatively intuitive. Health and beauty can become more sensitive depending on the user’s concern. Financial services can quickly move from ordinary product discovery into regulated or vulnerable decision-making.

A credit card ad near a casual travel-planning conversation is one kind of risk. A financial product appearing during a conversation about debt stress is another. A beauty product appearing during a routine shopping query is one kind of placement. A similar product appearing near a conversation about body image or health anxiety is very different.

OpenAI can define sensitive exclusions, but the operational problem is harder than a static category list.

Conversations move. A user can begin with a simple commercial question and end up in a high-stakes context within a few turns. The ad system has to understand not only the current prompt, but the surrounding exchange.

That means brand safety in ChatGPT will require more than keyword controls. It will require policies, context classification, auditability, and clear escalation paths when placements go wrong. Brands should not treat this as a normal adjacency problem. They are buying media inside an interface that may be helping users make decisions.

That is a different standard.

The Google Comparison Is Getting Sharper

The timing of OpenAI’s Cannes pitch is important because Google is also moving advertising deeper into AI-shaped search.

The FT notes that Google recently said it would introduce more ads in its AI Mode search results, including highlighted answers to queries, while Gemini itself does not show ads. Google’s own marketing materials describe new AI-era search ad formats that use Gemini capabilities, conversational discovery, and highlighted answers.

This creates a revealing divide. Google is integrating ads into AI-enhanced search. OpenAI is trying to build ads into an assistant that competes with search before the user reaches Google.

Both companies are moving toward the same commercial territory: AI-mediated intent. The difference is the starting point. Google starts with the search business and adds AI features around it. OpenAI starts with the assistant and tries to attach advertising to the commercial intent already appearing inside conversations.

That is why the ChatGPT ad story is strategically larger than a new format. It is part of a broader fight over who controls the interface between intent and action.

For two decades, Google owned that position at scale. Users had a question, typed it into Google, and often began the commercial journey there. ChatGPT changes that pattern because users may begin with the assistant instead. They may never perform a traditional search, or they may arrive at search only after the assistant has already shaped the criteria.

If marketers follow that shift, budgets will follow. They may not move all at once, and they may not move quickly enough to justify the most aggressive revenue projections. But the direction is clear. Advertising money follows behavior. If commercial behavior moves into assistants, media infrastructure will move with it.

The $100 Billion Question

The FT reports that OpenAI has told investors it could grow advertising into a $100 billion business by 2030, according to advertising executives, although OpenAI declined to comment on the figure. Reuters separately reported the same $100 billion projection, citing Axios, along with expectations for major revenue growth over the next several years.

Advertisers are right to be skeptical. A $100 billion ad business would require not only massive user scale, but also trusted formats, reliable measurement, strong conversion performance, broad advertiser adoption, and enough user tolerance to avoid damaging the product.

The ad market has seen ambitious platform stories before. Some scaled into dominant businesses. Others struggled because attention did not convert into high-quality intent, because advertisers could not measure outcomes, or because users resisted the commercial experience.

OpenAI’s advantage is that ChatGPT usage contains commercially valuable intent. Its challenge is that the assistant experience depends on trust. The more aggressive the ad load becomes, the more fragile that trust may become. The more conservative the ad experience remains, the harder it may be to reach the largest revenue targets.

That tension will define the business.

The company is loss-making, facing enormous infrastructure costs, and preparing for a public-market future. A public investor base will want evidence that OpenAI can turn usage into durable revenue. Advertising offers that story. It also changes the story investors, regulators, agencies, and users will tell about ChatGPT.

The product is no longer only an AI assistant. It is becoming a monetized commercial environment.

Codex Shows the Second Layer of the Pitch

The Codex part of OpenAI’s Cannes presence should not be treated as a side note.

OpenAI showed brands how Codex could be used to code apps that help users create advertising campaigns. Company executives played down any ambition to replace traditional agency work, but the demonstration points to the second layer of OpenAI’s marketing strategy.

The first layer is media. ChatGPT becomes a place where ads can appear in front of users with commercial intent.

The second layer is production. Codex becomes a tool that helps brands build applications, automate workflows, and reduce the cost of campaign creation.

Taken together, OpenAI is not only asking for ad budgets. It is positioning itself inside the operating system of marketing.

It wants to influence how campaigns are built, how customers are reached, and how commercial intent is converted.

That is why agencies should pay close attention. The near-term question is whether ChatGPT ads perform. The longer-term question is how much of the marketing stack moves into AI platforms that provide both creation tools and distribution channels.

When the same company offers the assistant, the ad inventory, the campaign-building tools, and the measurement layer, the balance of power shifts.

What Executives Should Watch

The first issue to watch is user tolerance. OpenAI says ads are labeled and separated from answers, but users will judge the experience by feel. If ads are useful, limited, and clearly distinguished from assistance, they may become accepted. If they feel intrusive or too closely connected to recommendations, the backlash could become sharper than it would be on a conventional media platform.

The second issue is performance. OpenAI’s claim that about one-fifth of ChatGPT queries have direct commercial intent gives advertisers a reason to test. But advertisers will need evidence that conversational intent produces measurable business outcomes. Strong early categories such as travel, retail, health and beauty, and financial services provide direction, but they also expose different regulatory and reputational risks.

The third issue is agency positioning. Agencies can treat ChatGPT ads as another platform buy, or they can treat them as evidence that the customer journey is moving into AI-mediated environments. The latter approach is more valuable. Clients will need help deciding not only where to spend, but how to maintain visibility, trust, and brand safety when assistants become intermediaries.

The fourth issue is Google’s response. Google is not standing still. Its AI Mode and AI-era ad products show that the search giant is defending its commercial-intent territory. The fight will not be between traditional search and pure chatbot advertising. It will be between different models of AI-mediated discovery.

The fifth issue is governance. Ads inside assistants require standards for disclosure, answer independence, sensitive-context exclusion, privacy, measurement, and advertiser accountability. Those standards cannot be improvised after scale arrives. The commercial category is forming now.

The Real Shift

OpenAI’s Cannes debut shows that the advertising industry is entering a new phase.

The first phase of AI in marketing was about content generation. The second phase was about workflow automation. The next phase is about distribution and intent capture. OpenAI is not only offering marketers a tool to make things faster. It is offering them access to the moment when users ask for help deciding.

That is a powerful place to sell media. It is also a risky place to commercialize.

If ChatGPT becomes a serious ad platform, marketing infrastructure changes. Search budgets become contestable. Agencies need new advisory models. Brands need new safety standards. Regulators will eventually examine how commercial messages function inside assistant interfaces. Users will decide whether sponsored content inside a chatbot feels useful or invasive.

The assistant is becoming the ad platform. The companies that understand that shift early will not treat it as another ad format. They will treat it as a new control point in the customer journey.