Explore CiteRank studies, industry benchmarks and multi-LLM experiments covering ChatGPT, Gemini, Claude, Perplexity, Google AI Overviews, Copilot, Grok and Meta AI.
Latest publication: 21 Jul 2026 · 10 entries · 8 engines tracked
The prompt-set design, sampling cadence, parsing rules and scoring model behind the AI Visibility Score — including what the score deliberately does not claim.
5 Jan 202614 min readRead the report →
Everything published in the Research Centre, newest first.
0 filters active · Showing 9 of 10 entries
Sorting by citation or download volume is not offered because CiteRank does not publish verified citation or download counts for these entries.
Google AI Overviews expands citation panels, Perplexity ships shopping-intent scoring, and what both changes ask of your source layer.
CiteRank AI Research5 min readRead research →
A default model shift inside ChatGPT changes citation behaviour on long-tail prompts. Short-term volatility, not a strategy change.
CiteRank AI Research5 min readRead research →
A new Claude release ships grounded citations by default, with early effects on B2B recommendation share.
CiteRank AI Research6 min readRead research →
Grok expands enterprise API access and Meta AI begins surfacing citations inside chat surfaces — two very different distribution shifts.
CiteRank AI Research4 min readRead research →
How often to re-sample prompts, engines and competitors so you catch real movement without drowning your team in alert fatigue.
CiteRank AI Research6 min readRead research →
Structural patterns that raise citation probability across ChatGPT, Perplexity, AI Overviews and the rest — and the popular patterns that do nothing.
CiteRank AI Research9 min readRead research →
The canonical entity fields AI engines look for before recommending a brand, and the reconciliation process that keeps them consistent.
CiteRank AI Research7 min readRead research →
How ranked, cited and mentioned events across eight engines are weighted into a single share-of-answer metric, and how to read its confidence interval.
CiteRank AI Research12 min readRead research →
How an ideal-customer profile becomes a statistically usable prompt corpus — stratification, phrasing discipline, localisation and versioning.
CiteRank AI Research11 min readRead research →
Ten standing lines of enquiry. Collections without published work are marked upcoming — we do not fill them with placeholder studies.
Population-level baselines for how often brands appear in AI answers.
Who gets named when an engine is asked to recommend, and in what proportion.
The domains AI engines lean on, ranked and explained.
How much a single prompt's answer moves between replays and over time.
Entries marked “illustrative data” contain modelled figures used to demonstrate structure and method. They are not client results, and no client workspace data is ever published. Every entry states its author, sampling approach and limitations — see the methodology.
