Search

Writing for the machines that answer questions

8 July 20267 min read

A traveller asking an AI where to stay in the Cyclades will get an answer. The question is whose words it is built from.

For twenty years the object of search was a ranking: be the first blue link and the click follows. That is quietly ending. A growing share of travel research now happens in a conversation with a model that reads the web on the traveller's behalf and returns a paragraph, not a page of results.

The paragraph is assembled from sources. Which sources is not random: models favour content that is structured, specific, unambiguous about facts, and consistent with what other credible pages say. A property that publishes a genuine guide to its region — with real distances, real seasons, real opinions — becomes a citable source. A property with three hundred words of adjectives about luxury does not.

The practical work is unglamorous. Clean structured data so a machine knows what is a room, a rate and a review. Facts stated once and stated the same way everywhere. Content organised by the question a traveller actually asks rather than by your internal departments. Most of it is the same discipline that made technical SEO work, applied with the assumption that the reader is not a person.

The strange consequence is that the properties best placed for this are the ones that were already writing honestly. There is no shortcut to being the source a machine trusts, which is the first genuinely encouraging thing to happen to search in a decade.

Next

The end of the discount season

Read

Begin your journey

Let us look at your property.

Tell us about the hotel, the season ahead and what is not yet working. We reply to every enquiry within two working days.