How AI models — ChatGPT, Claude, Gemini, Perplexity — answer when prospective clients ask about venues like this. The new front page of search.
| # | Prompt | Sector | CGPT | Cld | Gem | Pplx | Avg |
|---|
This month's headline is a venue with two faces. To 412,000 readers — most of them through LondonWorld's coverage of the new rooftop terrace — 30 Euston Square is the Edwardian venue with the hidden roof and the King's Cross skyline. Sentiment is unanimous and the heritage-meets-modern story is landing exactly as briefed.
To the AI models that now answer the questions prospective bookers ask first, the picture is sharper and more useful. The venue is recognised with 83% accuracy when named directly, but only 8% of the time in unbranded discovery. That 75-point gap is the most valuable diagnostic on this report — it tells us the brand is known but not yet a category answer. The fix is structural: category-language pages, aggregator estate hygiene on Tagvenue, Square Meal and DesignMyNight, and editorial that puts 30 Euston Square inside the comparison sets the models construct.
Next month's work, in short: convert recognition into discovery.
Paste the full contents of next month's GEOCO .md report below. The dashboard will parse it, store it locally, and update the trend line automatically.