Welcome to episode 25 of The GEO Show, the briefing on generative engine optimization, LLM citations, and AI search visibility. Brought to you by GEOforge.
In this episode, Paris Childress, founder of Hop AI and co-founder of GEOforge, runs ten stories with slides. He covers a GEOforge sixteen-week graded-answer case study where a blended GEO score hid three contradictory engine outcomes, Yext finding four AI engines agree on the number one recommendation only 4% of the time, Semrush manufacturing data that separates mention, citation, and traffic, fashion treating AI visibility as a PR metric, Claude's B2B-skewed Similarweb audience, Similarweb data inside Perplexity Computer, AirOps repositioning AI search as bigger than SEO, Cited Index's 114-tool catalog with 97 monitoring-only products, Surfer MCP, and Surva.ai's $39 monitoring-plus-content bundle.
In this episode:
📊 GEOforge: one blended GEO score, three contradictory engines
A sixteen-week graded-answer study ran thirty buyer prompts forty times each against ChatGPT, Google AI Mode, and Google AI Overviews. Share of voice moved differently on each surface while the blended score rose. Engine, topic, and recommendation content change the verdict, so divergence belongs in product and category positioning.
🎯 Yext: four engines agree on #1 only 4% of the time
Yext's research brief finds four AI engines agree on the number one recommendation only four percent of the time. OpenAI leaned on directories; Anthropic used review platforms more often. There is no single AI search ranking reality, which is why engine consensus and divergence scores matter.
🏭 Semrush: mention, citation, and traffic are different layers
AI Overviews rose from 38% to 57% of tracked manufacturing search volume from January to July, while AI Mode and AI assistance still drove only a small share of site sessions. Visibility starts at mention and citation, moves through recommendation and visit drop-off, and ends at revenue.
👗 Fashion turns AI visibility into a PR metric
Launchmetrics tracks ChatGPT, Claude, Gemini, and Perplexity across ten markets and event windows, covered in Vogue. GEO is escaping the SEO department into PR, communications, and brand budgets because what AI says often comes from earned media.
📈 Claude's audience skews B2B (Similarweb)
Claude's worldwide mobile audience grew roughly 14x versus about 31% for ChatGPT, with users more likely to visit business-service and programming sites. Absolute ChatGPT size understates Claude for SaaS, cybersecurity, and developer tools.
🔌 Similarweb data live inside Perplexity Computer
Proprietary traffic, audience, and market-share data is now callable through Perplexity Computer. AI answers increasingly draw from connected datasets, not only public web retrieval, which further fragments where brands must be represented accurately.
📣 AirOps: AI Search is Bigger than SEO
AirOps' September 9 webinar positions AI search as a core buying channel where content visibility alone no longer explains discovery, citations, or buying decisions. The GTM move is upward: from content optimization toward AI decision influence.
🧰 Cited Index: 114 tools, 97 monitoring-only
Cited Index catalogs 114 AI visibility and GEO products; 97 are monitoring-only. Read-only dashboards without statistical rigor or an execution layer leave the hard work (citation building) to a small minority of the category.
🧩 Surfer MCP: SEO workspace inside ChatGPT and Claude
Surfer MCP lets users access and act on Surfer data from MCP-capable assistants. Large SaaS leaders that expose MCP will keep a seat; mid-tail tools risk being replaced by capabilities inside the LLMs.
🆕 Surva.ai: $39 monitoring, content, and publishing
Surva bundles thirty prompts across five platforms, content generation, citation analysis, and CMS publishing paths from $39/month, with docs updated September 8. New entrants are shipping full-funnel feature sets, not visibility alone.
If this briefing is useful, subscribe to The GEO Show so we can keep walking through the platform changes and citation mechanics that actually move AI visibility.
💬 Question: If four AI engines agree on number one only 4% of the time, which engine are you still treating as the "real" ranking, and why?