ChatGPT Ads: Initial Findings and the Implications for Search Marketing

Digital Dynamo’s Paid Ads Specialist, Ricardo Ortega Salas, gives his thoughts on OpenAI’s integration of ads in ChatGPT – what it means for OpenAI, its users, and the future of ChatGPT as a search tool.

OpenAI – the company that brought us ChatGPT and has lately drawn criticism for its partnership with the Trump administration – continues to push how AI is used, consumed, and perceived. OpenAI’s decision to introduce ads into ChatGPT is a confounding yet foreseen development. What once was seen as a simple, user-focused platform is now being disrupted by sponsored content.

However, early indications show these ads will look and feel very different than our usual ad interactions. In this article, I discuss what the move toward paid advertising could mean for OpenAI, its users, and the future of AI-driven search.

Sam Altman, OpenAI’s CEO, adamantly stated that ads were not in ChatGPT’s best interest in terms of its business model and strategic direction. So the question arises: why the switch? Why now?

Well, first it’s important to note that not all users are opted into an ads-enabled interface. Users on the Free or ChatGPT Go versions will be opted in, while Plus, Pro, Business, Enterprise, and Education versions will not. This indicates OpenAI’s want to balance disruption with a continued focus on user experience through consultative trust.

That being said, the move to introduce ads is best explained by OpenAI’s need for additional funding for the development and research of future free AI features. Ads revenue will fund the continued expansion and accessibility of features and tools that all ChatGPT tiers will benefit from.

OpenAI has released how ads will work and the principles that govern them, but initial findings suggest that kinks are still being ironed out.

For example, Adthena found that ads are being triggered following a user’s initial prompt. This goes against OpenAI’s rhetoric of deep understanding to provide deeply precise and context-guided results even in paid media placement.

Ad placements show a brand icon to immediately differentiate and identify advertisers, as well as a “Sponsored” callout much like that of Google Ads. Content is then placed after ChatGPT’s answer to a user prompt. This provides the user an experience that highlights problem-solving and consultative solutions rather than disruptive placement or placement that promotes passive scrolling.

Another interesting finding is that most advertisers are established, widely-known household brands like Target, AT&T, and Ford. Again, this suggests a controlled and intentional move to continue pushing trust as a pillar of ChatGPT’s user experience.

Upon its introduction, ChatGPT was comparable, and perhaps even overlapped, with SEO in practice in that prompts are rooted in education and problem-solving. However, with the introduction of ads, this blurs the line between SEM and SEO in a way never before seen. While Google’s move to contextual targeting makes an attempt to create synergies between commercial intent and search queries, LLM’s core learning structure centered around nuance and context conceptually makes keywords obsolete. Queries are much more complex and varied in ChatGPT and these factors are what makes this platform rapidly adaptable and valuable in targeting commercial intent.

Additionally, initial reporting indicates that ChatGPT is positioning itself as a premium platform based on preliminary pricing data. Calculations are showing CPMs of around $60, which would explain the reason for why only household brand names are currently investing in these campaigns. The high CPM may also be an intentional strategy to keep trust front and center, as ChatGPT benefits from aligning with the credibility that established brands have spent years carefully building. The idea is that if big brands run ads on this platform, then it can be trusted.

The undertone of this new introduction is that trust is still at the forefront of the platform’s development. There is an understanding that this introduction has created a delicate balancing act between “disruptive experience” and “additive features.” The developing perception of this balancing act will shape how both users and advertisers decide to utilize the channel.

After numerous attempts to intentionally prompt these ads without signaling that I wanted product pricing, I eventually had to resort to a “what are the best deals?” prompt that would clearly trigger a buying recommendation. Even then, the ads were hidden behind a product image that displayed a single price point. Once an image was clicked, a sidebar opened on the right side of the AI answer, with additional details selling the product.

So while the AI did present answers and product options in response to the prompt, the actual ads remained tucked away until additional engagement occurred. Personally, I liked this approach. It struck a balance between being consultative and sales-driven, with the selling component only appearing once stronger intent and interest were demonstrated. While not as prominent or “in your face” as Google Shopping ads, I can see how this type of interaction could still produce clear wins for advertisers while lessening the adverse effects on users.

All in all, my personal findings and all the initial reporting mixed with skepticism & excitement makes me wonder: as powerful and useful as this introduction could be, users are weary and knowledgeable more than ever about how these corporations are leveraging deeply personal data for financial gain.

In addition, what was once a safe space without a “buy now” message on the peripheral, has now become another platform that adds more noise in an already saturated advertising space. Unfortunately, tools this powerful remain optional. Users either learn to leverage them or risk being left behind. Advertisers, on the other hand, have been presented with a channel merging concepts never seen together in the search advertising industry.

However, I think advertisers will be much less blindly accepting of the platform’s guidelines for best use. The reason for this is because we already had somewhat of a similar situation with Google’s Performance Max. Initially, many advertisers excitedly adopted what many thought was the logical next step in automation guided by AI. However, the lack of transparency within a system that requested trust to execute became challenging, and over time, led to the resignation of the campaign type.

The same could happen with ChatGPT in that the same excitement could turn into tension as advertisers balance the need for transparency with the power of contextual targeting without exposing deep user data. Ultimately, it’s still too early to reach any definitive conclusions about what this development will mean for users and advertisers. But because we all fall into one of those two groups, the field will be interesting to evolve in and watch very closely.

If you want to learn more about paid advertising either through Google Ads or Social Media and Meta Ads, please contact Digital Dynamo for a consultation.