The Rise of Agentic Advertising

Quick definition: a new approach to digital marketing where autonomous AI systems, known as AI agents, plan, execute, and optimise entire advertising workflows without constant human intervention. 

If you follow the tech industry, you’ve likely seen the disconnected fragments of a much larger story. On LinkedIn, predictions are swirling about OpenAI’s transition to a for-profit entity and what it means for monetisation. You may have also seen the news of its new, AI-native Atlas browser or noticed its high-profile ad-tech hires and just days ago, they released ChatGPT Group Chats, allowing users to invite friends and colleagues into a shared conversation with the AI.

Taken in isolation, these are just moves in a fast-changing market. But what if they aren’t isolated events?

What if they are three carefully aligned pillars of a single, unannounced strategy? A strategy to not just introduce ads, but to build an entirely new, multi-trillion-dollar advertising and commerce ecosystem from the ground up.

While OpenAI has not formally announced an ad platform, one can connect the dots. The financial pressure is the motive, the strategic hires are the intent, and the Atlas browser appears to be the mechanism.

Here’s a look at the predictive case for this new agentic ad model and what it could mean for every brand and marketer.

1. The Why: The Inevitable Financial Imperative

The idea of mission before margin is a noble one, but it’s colliding with the staggering financial reality of building artificial general intelligence. The compute costs are astronomical. Reports indicate OpenAI is generating impressive revenue (around $13 billion annually) but is also committed to a long-term infrastructure spend that could exceed $1 trillion. Current operating loss projections for 2024 and 2025 are reportedly in the billions.

The company’s popular $20/month subscription model is successful but according to some analyses, it only accounts for about 5% of its 800 million regular users. This model alone likely cannot subsidise the compute costs for 95% of free users.

This financial pressure makes mass-market monetisation seem less like a choice and more like an existential necessity. The company’s recent restructuring into a for-profit Public Benefit Corporation (PBC), a structure that allows it to raise funds like a traditional tech company, only reinforces this.

This context makes leaked internal projections, which reportedly forecast $1 billion in new revenue from free user monetisation in 2026, growing to $25 billion by 2029, seem not just plausible, but logical.

2. The Who: Actions Speak Louder Than Words

If the why is the financial pressure, the who is the strategic hiring. Companies signal their true intentions not by what they say, but by who they hire.

Recently, OpenAI has been posting job listings for roles like Growth Paid Marketing Platform Engineer. A close read of these job descriptions reveals they aren’t about buying ads. The roles involve developing campaign management tools, integrating with major ad platforms, and building real-time attribution and reporting pipelines. This is the language of a company building its own walled-garden ad platform from scratch.

But the most telling move was the hiring of Fidji Simo as CEO of Applications. Simo is a highly respected executive, but her specific expertise is what’s critical. Before her role as CEO of Instacart, she spent a decade at Meta and is widely credited with being one of the chief architects of advertising in the mobile News Feed.

She has a world-class track record of solving the exact problem OpenAI faces: how to natively integrate a massive ad business into a core product without destroying the user experience. Hiring her is an on-the-nose acknowledgement that advertising is the designated path forward.

3. The How: Atlas Browser & Group Chats

This brings us to the how. An ad platform is useless without data, and the ChatGPT interface is a notorious black box for marketers. Brands have no idea how, or even if, they appear in conversations.

This is where the new Atlas browser fits in. It could be the missing prerequisite, a purpose-built gateway engineered to solve this data deficit.

While framed as a new way to browse, its core features look like the perfect foundation for a new ad model:

• Browser Memories: This feature remembers what users read and do across sessions. For a user, it’s a convenience. For an ad platform, it’s a persistent intent-profiling layer that builds an unparalleled understanding of user habits and preferences.
• Agent Mode: This is the execution layer. The browser stops being a passive viewer and becomes an active assistant that can shop for you, make reservations, or buy tickets. It’s the mechanism that can take action on a user’s behalf.

The browser, in essence, provides the two things ChatGPT alone lacks: deep, cross-session intent data and a mechanism to execute transactions.

The other newly released ChatGPT Group Chats feature lets users invite up to 20 people into a shared conversation. This acts as a strategic Trojan horse. Instead of building a new social network, OpenAI created a link that users paste into existing WhatsApp or social threads. This lets them borrow social graphs from other platforms and insert AI into private decision-making circles. For marketers, this unlocks collective intent data which search engines miss. High-stakes choices like university courses are rarely solitary decisions. They involve debate and peer validation. In a group chat, the AI observes the hesitation and criteria behind the final choice. Because the AI only speaks when relevant, users view it as a helpful participant rather than a tool. This positions it perfectly to become a sponsored agent that actively facilitates group bookings or applications.

4. The What: A New Agentic Advertising Model

This new platform likely won’t be a clone of Google’s search ads. CEO Sam Altman has been publicly critical of the pay-to-rank model, which he feels can compromise user trust. He has spoken about his fear of a trust-destroying moment if OpenAI began modifying its answers based on who paid more.

So, what’s the alternative? The solution may be a new framework called agentic advertising.

The core innovation is simple but profound: it targets the content and context of the conversation, not the user’s personally identifiable information. This model could solve the privacy-monetisation dilemma by matching ads to conversational themes (e.g., sustainable hiking gear) rather than a user’s personal data.

We can predict that this new platform will likely offer two new types of ad formats:

• Conversational Ads: These would be passive, informational placements that feel native to the chat. Think Sponsored Answers or Brand-Powered Knowledge Cards. If you ask for a smoothie recipe, the AI might suggest a trusted blender brand.
• Action-Driven Campaigns: This is the revolutionary part, powered by Atlas’s Agent Mode. Here, the ad is not a click; it’s an executed task. A user might say, “Create a new ad campaign for my website,” and an agent does it. In this future, brands like Delta or United wouldn’t bid on the keyword flights to New York; they would bid for their booking agent to be the default one ChatGPT uses to complete the user’s request.

Analysts are already projecting it could be a $1.7 trillion market by 2030. The fact that PayPal, Visa, and Mastercard have all already built and integrated bespoke agentic commerce services signals that the financial world is betting on this future, and OpenAI is building the primary platform to host it.

What This Means for Marketers

For brands, this is both a threat and an opportunity. Right now, ChatGPT is a dark funnel. Marketers are flying blind, and recent updates have reportedly caused brand visibility within ChatGPT to drop by an average of 31%.

The greatest value of an OpenAI ad platform, therefore, won’t just be the ads; it will be the data.

It would finally solve the black box problem by providing Conversational Analytics, giving brands structured insights into what users are asking and how their products are being framed.

This is set to create an entirely new marketing discipline. An entire ecosystem of third-party tools is already emerging to measure these new, critical KPIs:

• Generative Share of Voice: How often is your brand mentioned versus competitors?
• Citation Frequency: How often is your URL cited as a source?
• Sentiment and Framing: How is the AI describing your brand?

While nothing is certain, the pieces fit together too neatly to ignore. The financial mandate, the specialist hiring, and the data-gathering browser all point to an inevitable conclusion. We aren’t just about to see ads in ChatGPT; we’re likely witnessing the construction of the entire transactional and analytical layer for the next phase of the web.

And with some analysts predicting a beta launch by mid-2026, the time for marketers to start building their AEO (Answer Engine Optimisation) strategy is now.