Can You Trust One AI Engine to Measure Visibility?

Business websites accounted for nearly 60% of Gemini's citations in local AI search, according to a new study. Steady Demand's analysis of 14,472 citations from 1,487 local search queries across 50 U.S. metro areas also found that identical searches often produced different AI recommendations.

What Happened

Steady Demand analyzed 14,472 Gemini citations across 1,487 local queries in 50 U.S. metro areas. Business websites captured nearly 60% of citations, beating directories, review sites, and forums combined. Reddit came second at 13.7%. Co-founder Ben Fisher flagged what he calls Grounding Drift: repeat Gemini queries returned overlapping sources only 40% of the time, and named the same top business just 7% of the time. ChatGPT and Gemini agreed on domains only 8% of the time.

The Numbers in Context

Google's traditional local pack returns the same top listing roughly 90% of the time. Gemini returns the same top business only 7% of the time on identical queries. That is a 13x collapse in result stability between classic search and generative answers, and Gemini and ChatGPT overlap on top recommendations just 4.2% of the time.

Why This Matters for B2B Marketing Leaders

If you run demand generation at an HR Tech or FinTech brand, you are likely already tracking AI citations as a proxy for share of voice. This study says that proxy is noisier than you think. A single Gemini snapshot showing your brand cited alongside Workday, Rippling, or Plaid may not repeat an hour later. Worse, ChatGPT is probably citing entirely different sources for the same question your buyers ask. You need multi-platform monitoring, larger query samples, and time-series tracking before you make budget decisions based on AI visibility data. One good screenshot is not a trend.

The Starr Conspiracy's Take

Grounding Drift is the AEO story of the year, and most brands are not ready. If you optimized your site for one engine's citation pattern, you optimized for a moving target. The winners will treat AI visibility the way media buyers treat reach and frequency: sampled, weighted, and measured across platforms over time. That reframing is exactly what we cover in our guide to answer engine optimization for B2B and why we push clients toward owned-content depth over directory dependence. Your business website is still your most citable asset. Make it worth citing.

What to Watch Next

Expect AI visibility platforms to add stability scores and cross-engine overlap metrics within the next two quarters. Watch whether Google exposes any Gemini grounding controls at its next Search Central event, and whether Perplexity and Claude publish comparable citation data. Buyers will likely demand standardized reporting before year-end.

Related Questions

How should HR Tech marketers measure AI search visibility?

Run the same query set weekly across Gemini, ChatGPT, Perplexity, and Claude. Track citation frequency, source type, and stability over time rather than single snapshots. Weight results by buyer-relevant queries, not vanity terms, and benchmark against two or three named competitors.

Why does Gemini cite business websites more than directories?

Gemini's grounding favors first-party content when the site demonstrates topical depth and structured answers. Directories often lack the specific, question-shaped content AI engines prefer. Our pillar on content strategy for AI answer engines breaks down what actually earns citations.

Does ranking in Google still predict AI citations?

Partly. Classic ranking signals feed grounding, but AI engines reweigh sources by answer quality and entity clarity. A page can rank well and still be ignored by Gemini if it buries the answer. Optimize for extractability, not just position.