Nicolas Sitter
AGENTIC|

What AI searches when asked for hotels in Melbourne

Before answering a hotel question with web search, an AI assistant rewrites it into one or more searches. These are the searches it actually ran for 11 traveller prompts about Melbourne, captured week after week, not autocomplete suggestions.

ChatGPT
193
answers, 1.3 searches each, May 2026 to Nov 2026

Explore by prompt

Percentages are shares of answers. Grok and Gemini were captured on an earlier prompt set, so most current prompts only have ChatGPT.

Best hotels
Luxury
Budget
Families
Couples
Business
Solo
Amenity
Neighbourhood or landmark
Prompt
“best hotels in Melbourne”
Answers captured
19
Searches per answer
1.6
Searched the prompt as written
5%
Captured
4 May 2026 to 2 Nov 2026

The searches ChatGPT ran

Most frequent first. “×3” means the same search appeared in three separate answers.

  1. best luxury hotels×3
  2. best hotels Melbourne Australia×3
  3. best boutique hotels×2
  4. best hotels Melbourne Australia luxury 2026×2
  5. best hotels in Melbourne Australia 2026 luxury boutique hotel reviews location×1
  6. best hotels in Melbourne CBD×1
  7. best family hotels×1
  8. best hotels in Melbourne Australia luxury boutique family 2026 reviews guide×1
  9. best rated hotels downtown×1
  10. best hotels in Melbourne Australia luxury boutique central 2026×1
  11. best luxury hotels Melbourne Australia 2026×1
  12. best hotels Melbourne×1
  13. best hotels Melbourne Australia 2026×1
  14. best hotels Melbourne Australia luxury hotels 2026×1
  15. best hotels Melbourne Australia luxury×1

Showing 15 of 24 distinct searches.

Words it added to the prompt

australia79%luxury68%202642%boutique21%cbd11%reviews11%address5%carlton5%central5%collins5%downtown5%family5%guide5%hyatt5%location5%official5%park5%rankings5%rated5%ritz5%

Places and names it added

CBD11%

Hotel topics in what it added

All Melbourne prompts together

Share of answers whose searches added each topic, counting only words that were not already in the prompt.

What a Melbourne hotel page can cover

The topics ChatGPT adds most often here, with what a page can do about each. These are editorial suggestions drawn from the searches, not a measured ranking effect.

  1. Country name40% of answers

    Write the city and country together at least once. Engines disambiguate with the country.

  2. A year19% of answers

    Say what is current: this year’s renovations, prices and openings, with dates on the page.

  3. Hotel class or style19% of answers

    State your class and style plainly (five-star, boutique, historic) in text and in schema.

  4. Best / top16% of answers

    The engine is looking for lists. Being on the editorial and booking-site lists matters as much as your own page.

  5. Who it is for12% of answers

    Say who the hotel suits (families, couples, business) and why, with specifics.

Which hotels get named? The answers behind these searches are tracked in the AI Hotel Landscape for Melbourne.

How the searches were captured and cleaned, and what is left out: method and limits. ChatGPT data runs to Nov 2026.