GEA: How agentic AI is transforming online advertising

As a Partner at Converteo and a specialist in AI, Agentic & Data strategies, Guilhem Bodin helps Marketing and Digital departments transform their operations by defining agentic strategies, optimizing customer journeys in the conversational era, and measuring the business impact of their AI investments.

Key Takeaways

• The illusion of a neutral AI is fading in favor of a model where recommendations become a sales channel, giving rise to GEA (Generative Engine Advertising).
• Google holds a major strategic advantage, as its existing advertising ecosystem (Performance Max, Shopping feeds) is already poised for integration into Gemini’s responses.
• To survive, brands must enrich their product feeds with contextual data (benefits, use cases), making them “readable” and relevant to conversational engines.

The illusion of a neutral artificial intelligence, guided solely by relevance and capable of recommending “the best” product, is dissolving. Not through a groundbreaking announcement, but through a series of subtle technical and economic signals that reveal a clear trajectory: conversational AI will not just advise; it will become a full-fledged sales channel.

The question is no longer if advertising will come to AI assistants, but how and who will define the rules. In this arena, the strategic gap between Google and OpenAI is already visible.

Google shopping in Gemini: the dawn of a revolution

Recent tests have shown Google Shopping listings appearing directly within Gemini conversations during explicitly transactional queries. Officially, Google has downplayed some interpretations and denied any formal announcement of a widespread advertising rollout in the short term. But the key point lies elsewhere: the infrastructure exists, the building blocks are in place, and the experiments are very real.

We are witnessing the emergence of a new discipline: after SEO, the era of GEA, or Generative Engine Advertising, is now upon us. This is a logic where visibility is no longer just about content quality, but about the ability to be selected and recommended by a conversational system integrated into the purchasing process.

GEA has entered the chat : Google’s strategic sdvantage

In this field, Google has a decisive advantage. Not because it has reinvented advertising, but precisely because it hasn’t needed to. Performance Max, AI Max, Shopping feeds: these systems, which advertisers already use, were not just automation tools. They form a foundation ready to be injected into Gemini’s responses.

No radical interface change is necessary; the continuity between intent, recommendation, and conversion is seamless. However, this continuity implies a quiet but crucial evolution for brands: product feeds can no longer be limited to technical or promotional attributes. To exist in a response generated by an LLM, they will need to incorporate more usage context, consumer benefits, and intent signals—in other words, be designed to make a recommendation immediately actionable.

Google’s shopping assistant vs. OpenAI’s universal assistant: two visions

It is precisely this ability to turn intent into action that marks the difference in approach. Where OpenAI aims to become a universal assistant, Google is leaning towards a shopping assistant. By integrating advertising logic directly into the core of transactional responses, AI no longer just informs a decision: it closes it. Friction disappears, the recommendation becomes actionable, and value is captured at the most critical moment of the journey.

The only unknowns are the timing and the competitive dynamics: will Google activate this lever first, or in reaction to ChatGPT’s monetization?

The end of neutrality: toward an AI guided by business models

Whatever happens, one thing is certain: the era of neutral AI is coming to an end. Tomorrow, AI will no longer recommend “the best product,” but the one that has successfully integrated itself into its economic model. This battle will not be limited to bidding levers or ad formats; it begins now, with brands’ ability to make their products readable, contextualized, and interpretable by conversational engines.

The first experiments already show this. Whether through direct ad integration in responses or more indirect methods—like Amazon monetizing questions asked to the consumer in its Rufus response engine—recommendation is no longer a neutral space. It’s already another, more subtle way of guiding the decision.