10 Million AI Search Results: The New Rules of SEO - brightonSEO Spring 2025

Session Title：We analysed 10,000,000 AI search results: here's what we found

At BrightonSEO Spring 2025 in the UK, MediaReach, Inc. joined hundreds of global marketers to explore the future of search in an AI-driven world.
One of the most insightful sessions came from Josh Fusco, Founder of Profound (Vermont, USA), who shared the results of an extensive analysis of over 10 million AI-generated search results across platforms like ChatGPT, Google, Perplexity, and Microsoft.

This report, written by Ayaka Uchida (SEO Consultant at MediaReach), distills the key findings from the session and examines how AI answer engines are fundamentally reshaping the rules of visibility, authority, and content strategy in search.

Written by Ayaka Uchida
SEO Consultant, MediaReach, Inc.

1. Executive Summary

In this data-driven session, Josh Fusco from Vermont shared findings from analyzing over 41 million AI-generated answers across various platforms including ChatGPT, Google, Perplexity, and Microsoft. The presentation revealed that AI search is fundamentally different from traditional SEO in both architecture and ranking factors.

Key insights included the surprising findings that traditional SEO metrics like traffic and backlinks have minimal or even inverse relationships with AI search visibility. The speaker demonstrated that technical optimization, including indexing by Bing, implementing LLMs.txt protocol, and creating AI-optimized content formats like listicles and comparisons, are critical for success in this new landscape.

The session provided compelling evidence that search is undergoing a fundamental transformation, with Google losing market share while ChatGPT has experienced a 400% increase. This shift represents a new paradigm where AI answer engines sit between users and websites, fundamentally changing how users discover and engage with content.

2. Session Details

Session Title: We analysed 10,000,000 AI search results: here's what we found
Speaker: Josh Fusco (Founder, Profound)
Date / Time: Friday, April 11, 2025 — 15:20
Venue: Brighton Centre,  Auditorium 1
Event: BrightonSEO Spring 2025
Session Link: https://brightonseo.com/sessions/the-future-of-search

3. Report Details

3-1. Context and Background

The presentation addressed the growing uncertainty about how AI search engines function compared to traditional search. As Josh Fusco explained, his organization helps marketers understand, take action on, and improve their visibility within AI search environments.

The speaker positioned this analysis as urgent by highlighting significant market shifts: Google has decreased in market share for the first time in decades, while ChatGPT has increased by 400% in the same six-month period. This indicates that search is undergoing a fundamental transformation in how people discover information.

Fusco contrasted the old search paradigm (user visits website directly through links) with the new reality where AI answer engines sit between websites and users. In this new model, the AI engine arbitrates what's true vs. false, summarizes content, decides which brands or products to mention, and ultimately builds the relationship with the user.

3-2. Key Messages and Takeaways

• AI search and traditional SEO operate on fundamentally different architectures and cannot be approached with the same strategies

• Traffic shows almost no correlation with AI search visibility, accounting for only about 5% of citations

• Backlinks surprisingly have an inverse relationship with AI citations

• Being indexed by Bing is essential for visibility in most AI search engines, particularly ChatGPT

• The LLMs.txt protocol is emerging as a critical tool for controlling how AI engines access and interpret website content

• Content format matters significantly, with listicles and comparison content being the most frequently cited types in AI search results

• Social media and user-generated content (UGC) can have situational but powerful impacts on AI visibility

• The window for establishing AI search dominance is now, as patterns are still forming

3-3. Visual Materials and Slides

Figure 1:Overview of how ChatGPT fetches search results via Bing and WebGPT layers.：

Figure 2:ChatGPT and Google SERPs share only 12% of results.：

Figure 3:Traffic does not predict AI citations — only 5% correlation.：

Figure 4:Comparative listicles dominate AI citation sources across 177M results.：

Figure 5:Predictions for the future of AI search — from agent experiences to model context protocols.：

Figure 6:AI search is poised to become the most exciting opportunity in digital marketing.：

Comparison between Google’s traditional SERP layout and AI-generated answers, highlighting differences in structure, citation format, and user experience.

3-4. Practical SEO Implications

Short-term:

• Ensure your website is indexed by Bing, as this is a prerequisite for visibility in ChatGPT

• Implement LLMs.txt protocol to provide clear guidance to AI crawlers

• Use descriptive, semantic URLs that maintain meaning even when meta descriptions are unavailable

• Create more listicle and comparison content formats, which are highly preferred by AI systems

Long-term:

• Track AI search visibility separately from traditional search metrics

• Monitor where competitors are being cited in AI search results

• Consider strategic commenting on influential Reddit threads and other UGC sources

• Prepare for further divergence between traditional and AI search as technologies like AI agents continue to evolve

3-5. On-site Impressions

The Main Theatre was filled with attendees eager to learn about AI search trends, demonstrating the high level of interest in this emerging area. Josh Fusco delivered an energetic presentation that effectively conveyed complex technical concepts through accessible examples and case studies.

Despite being scheduled in the latter part of the conference day, the session maintained strong attendance, indicating the importance SEO professionals are placing on understanding AI search. Attendees were particularly responsive to the sections on LLMs.txt protocol and the finding that backlinks have an inverse relationship with AI citations.

3-6. Personal Reflection

The most compelling aspect of this presentation was how thoroughly it challenged established SEO practices with substantial data. The finding that backlinks could actually have a negative correlation with AI visibility runs counter to decades of SEO conventional wisdom, suggesting we may need to fundamentally rethink our approach to content strategy.

I found the case studies particularly valuable, especially the example of how commenting on Reddit threads dramatically increased a company's visibility in Google's AI results. This highlights how the boundaries between traditional SEO, social media, and PR are blurring in the AI search era.

The speaker's emphasis on the urgency of action resonated with me. As he noted, we are in the early days of AI search where patterns are still forming, offering a rare opportunity to establish strong positions before competitors fully adapt to this new paradigm.

4. Supplementary Materials

Speaker Deck (Josh Fusco): https://speakerdeck.com/joshbly/josh-blyskal-profound-we-analyed-10000-000-ai-search-results-dot-dot-dot

Written by Ayaka Uchida
SEO Consultant, MediaReach, Inc.