OpenAI executives appeared at the Cannes Lions International Festival of Creativity this week to share updates on the company’s advertising initiatives and demonstrate early results from its expanding work with marketing partners. The presentations highlighted measurable gains in ad performance and creative output, offering a window into how generative AI tools are beginning to influence campaigns across major brands.
During a session covered by The Information, OpenAI leaders including Chief Product Officer Kevin Weil and advertising lead Matt McGann described progress on several fronts. They pointed to improvements in ad generation speed, personalization at scale, and the ability to test creative variations more efficiently than traditional methods allow. One key example involved a campaign for a consumer goods company where AI-assisted tools helped produce dozens of video concepts in hours rather than weeks, leading to higher engagement rates in initial testing.
Weil emphasized that the focus remains on augmenting human creativity rather than replacing it. He described how OpenAI’s models can analyze audience data, suggest copy variations, and even generate rough visual storyboards that creative teams then refine. According to the briefing, early tests with select advertisers showed lifts in click-through rates ranging from 15 to 30 percent when AI-generated elements were incorporated into display and social campaigns. These figures come from controlled experiments conducted over the past several months with a small group of partner agencies.
The Cannes appearance marks a notable step in OpenAI’s efforts to court the advertising industry more aggressively. For much of its history, the company concentrated on research and enterprise applications, but recent shifts show a clear interest in commercial opportunities within marketing. McGann told attendees that the advertising sector represents a natural fit for the technology because of its constant demand for fresh ideas and rapid iteration. He outlined plans to expand beta programs that give agencies direct access to customized versions of GPT models trained on brand-specific guidelines and past campaign data.
One area receiving particular attention involves video ad creation. OpenAI has been refining its Sora model for generating short video clips from text prompts, and executives shared clips that demonstrated realistic product demonstrations and lifestyle scenes. While the technology still requires human oversight to ensure brand alignment and avoid inconsistencies, the samples shown suggested potential for reducing production costs on lower-budget campaigns. Agencies could theoretically use these tools to mock up concepts before committing to full-scale shoots, a process that traditionally ties up significant resources.
The executives also addressed concerns about data privacy and brand safety, two topics that frequently arise when discussing AI in advertising. OpenAI has implemented strict controls on how client data is stored and used for training, with options for brands to maintain completely separate model instances that never interact with the broader system. This approach aims to give marketers confidence that their proprietary campaign strategies and customer insights remain protected. McGann noted that several large advertisers have begun pilot programs under these conditions and reported satisfaction with the security measures.
Beyond creative generation, OpenAI is exploring ways to integrate its technology into the buying and optimization side of advertising. The company has experimented with models that can analyze real-time performance data and recommend bid adjustments or audience targeting changes. Early results indicate these systems can identify patterns that human analysts sometimes miss, particularly across multiple platforms and formats simultaneously. One partner agency reported a 12 percent improvement in return on ad spend after implementing AI-driven optimization recommendations in a regional test market.
The presentation at Cannes came at a time when many advertisers are actively seeking ways to incorporate AI without disrupting established workflows. Creative directors in the audience expressed both excitement and caution, asking questions about how AI might affect junior roles and the overall quality of work. Weil acknowledged these worries, stating that the technology performs best when paired with experienced professionals who can provide strategic direction and final approval. He cited examples where AI suggested ideas that initially seemed unconventional but proved effective after refinement by creative teams.
OpenAI has also been developing features specifically for A/B testing at scale. Rather than manually creating a handful of variations, marketers can now generate hundreds of combinations of headlines, images, and calls to action. The system then uses predictive modeling to estimate which versions are likely to perform best before any impressions are purchased. According to data shared during the session, this capability has helped some brands reduce wasted spend by identifying underperforming concepts earlier in the process.
The company is not operating in isolation. Several major holding companies have established partnerships to integrate OpenAI tools into their proprietary platforms. These collaborations allow agencies to access the models through familiar interfaces rather than learning entirely new systems. One such integration enables creative teams to stay within their existing project management software while querying the AI for suggestions or generating assets on demand.
Challenges remain, however. The executives openly discussed current limitations around consistency in long-form video and the occasional tendency of models to produce generic ideas when given vague prompts. They stressed the importance of detailed briefing materials and iterative feedback loops to achieve stronger outputs. Training programs for agency staff have become a priority, with OpenAI offering workshops on prompt engineering and output evaluation to help teams get the most from the technology.
Looking ahead, OpenAI plans to release additional advertising-specific features later this year. These include enhanced product photography tools that can place items in various environments without requiring new photoshoots, as well as audio generation capabilities for creating custom voiceovers and background music. The company is also investing in measurement solutions that can better attribute results when AI-generated creative appears in campaigns across multiple channels.
Industry reaction to the updates has been mixed but largely positive among those already experimenting with the tools. Some marketers worry about over-reliance on AI potentially leading to homogenized creative across the industry, while others see it as an opportunity to free up time for more strategic thinking. The executives at Cannes encouraged a balanced approach, suggesting that agencies treat the technology as a collaborative partner rather than an automated replacement for human insight.
The session also touched on ethical considerations. OpenAI has established guidelines for advertising use that prohibit certain types of claims and require clear disclosure when AI generates content that could be mistaken for human-created work. These policies align with broader industry conversations about transparency in an era when distinguishing between human and machine-generated material grows increasingly difficult.
As more brands begin to incorporate these tools, questions about intellectual property and originality continue to surface. OpenAI maintains that its models generate new combinations rather than directly copying existing work, but some creative professionals remain skeptical. The company has committed to ongoing dialogue with industry groups to address these concerns and refine its approach based on feedback.
The Cannes Lions appearance served as both a progress report and an invitation for wider adoption. By sharing concrete performance data and demonstrating live examples, OpenAI executives aimed to build confidence among decision-makers who may have been hesitant about integrating generative AI into their operations. The numbers presented, while preliminary, suggest that when applied thoughtfully, the technology can deliver tangible benefits in both efficiency and effectiveness.
For agencies and brands willing to invest time in proper implementation, the potential advantages appear substantial. Faster iteration cycles, more personalized messaging, and reduced production costs could reshape how campaigns are developed and deployed. Yet success seems to depend heavily on maintaining human judgment at key decision points and ensuring that strategic thinking guides the technology rather than the other way around.
OpenAI’s advertising push reflects a broader trend of AI companies seeking practical applications in creative industries. The company’s willingness to engage directly with marketers at an event like Cannes indicates confidence in its current offerings and a desire to shape the conversation around responsible use. As these tools mature and more case studies emerge, the advertising community will likely gain clearer understanding of where AI delivers the greatest value and where traditional approaches remain preferable.
The presentations underscored that generative AI has moved beyond experimental curiosity to become a practical component of modern marketing workflows. With continued refinement and close collaboration between technologists and creative professionals, the technology stands to influence everything from concept development to performance optimization. For now, the focus remains on measured adoption, careful oversight, and continuous improvement based on real-world results from campaigns running across the globe.
WebProNews is an iEntry Publication