Bluefish finds ChatGPT and Google favor different brands in 35% of topics

The AI visibility gap

New research from Bluefish reveals a significant discrepancy in how different AI models perceive market leaders. The study found that ChatGPT and Google favor a different brand at the peak of their AI-generated responses in 35% of 5,810 buying topics.

This means that for roughly one out of every three shopping topics, the two platforms do not agree on which brand is the primary choice. This discrepancy suggests that a brand's visibility is highly dependent on the specific AI architecture and data sources used by the provider.

The findings come as the marketing industry begins to grapple with the "agentic funnel." This concept refers to a shift where AI agents, rather than just search engines, guide consumer behavior by synthesizing information from multiple sources at once.

Bridging organic and paid

To address these inconsistencies, Bluefish launched a new product called Paid Media Briefs. The tool is designed to help media teams create ready-made ad recommendations for both Google and ChatGPT based on where a brand is missing from or misdescribed in AI answers.

The platform, which Bluefish calls the Agentic Marketing Platform (AMP), expands its previous work in organic brand monitoring into the paid advertising space. It identifies specific gaps where a brand is absent or losing ground to competitors and directs paid media to those opportunities.

Bluefish argues that the context of an AI answer changes how an ad lands. If an AI answer already recommends several rival brands, a brand's paid ad must be strategically positioned to stand out within that specific conversational or summary context.

Platform-specific mechanics

The new tool provides different types of instructions depending on the platform. For Google, the briefs cover AI Overviews and AI Mode, providing up to 50 search themes, 15 headlines, and four descriptions tailored for Search campaigns using AI Max or Performance Max.

ChatGPT ads operate differently because they do not rely on exact-match keywords. Instead, the briefs provide up to 50 context hints, which are phrases that guide the AI's matching system without guaranteeing delivery in every specific conversation.

The briefs are pre-generated and can be filtered by product line or topic. They are designed to be moved into bulk upload templates, which ensures that the suggestions are still processed by the advertiser's own systems and the platforms' internal tools before going live.

The evolving ad market

The growth of this space is significant, with US spending on advertising adjacent to AI-powered search results projected to reach $26.4 billion in 2026. Additionally, ChatGPT Ads has reached an annualized revenue run rate of approximately $1 billion as of late 2026.

OpenAI has already expanded its ad reach to 89 countries, with more than 100,000 advertisers running campaigns. However, the effectiveness of these ads can vary; some tests suggest that AI-driven targeting can sometimes result in a lower return on ad spend compared to traditional methods.

Because AI answers vary between runs of the same prompt, the industry is still learning how to measure success. Bluefish frames AI visibility as a critical input for budgeting and prioritizing campaigns rather than just a simple report card of a brand's performance.

What comes next for brands

For large companies, the goal is to move toward a unified way of orchestrating marketing. Bluefish cofounder Alex Sherman noted that the agentic funnel does not distinguish between organic and paid marketing, and neither should the tools used to manage it.

Marketers will likely need to monitor how different AI models describe their products in real-time. This data will become a foundational element for deciding where to invest and how to position products against competitors who may be favored by certain AI models.

As platforms like Google and OpenAI continue to update their ad surfaces, the ability to identify and fill visibility gaps will be essential. Brands that can successfully navigate these different AI ecosystems will have a significant advantage in reaching consumers who rely on AI for shopping advice.

Technical nuances

It is important to note that Bluefish has not released the specific categories tested in their research or the exact dates the answers were collected. They also did not provide a specific scoring method for how they determine if a brand is "leading" an answer.

Furthermore, some Google ad surfaces are restricted in sensitive verticals like finance and healthcare. Bluefish serves many of these industries, but the company has not clarified how their briefs handle categories where Google does not serve ads within AI answers.

Additionally, OpenAI's creative enhancements may automatically rewrite headlines and descriptions for some campaigns. This means that while a brief might recommend a specific title, the final version seen by the user may be modified by the platform's own system unless the advertiser opts out.