It's not just startups trying to capture the mood for developing natural language and semantic search in travel - Expedia is quietly working on its own platform.
The project is being run out of a new research and development department within Expedia Affiliate Network and is currently aimed at providing partner sites with the early stages of natural language search using the mounds of data collected elsewhere around the Expedia portfolio.
A sandbox site, YourVisit, is also up and running for testing (EAN admits it is a bit rough around the edges, but is in place purely for demonstration and beta purposes).
The idea behind the service is simple: provide a better, more human way of searching for travel products but using the vast myriad of data behind the scenes across the Expedia empire to return highly relevant results to users.
Expedia says the system is being trialled with a number of partner sites through the EAN and could be available for deployment across the network in the coming months.
"Replicating the human experience", according to EAN, a team fronted by chief technology officer Eachan Fletcher and director of R&D Christopher Lynch, is just one side of the project.
This allows users to pretty much type whatever they like into the search box to receive results.
But, arguably, it is what goes on in the background where the project stands out - mining (Big Data-style) around 1,300 data sources to ensure the results are not only accurate to the original search (often an issue with natural language search tools) but provide product ideas tuned specifically to the original request.
The dizzying array of data collected over the years across Expedia and its network of brands is fuelling this: material such as transaction information and trends, search queries on-site and in engines such as Google, user generated content and the reams of meta-data associated with every product page.
So, for example, if a user asks for a hotel with "free breakfast and wifi" in a destination with lots of hotels, the system will begin working on specific keywords to narrow down the search and then return results based on the data it can find about the property on its own system, rather than the content supplied by the partner.
It is all about "relevancy", EAN says, with the main aim of improving transactions on partner sites where currently web conversions often run at anything between 1% and 5% compared to call centres where they can be up to 50%, simply because the conversation between the brand and the user drills down to the exact demands of the customer.
The technology is being developed entirely in-house at EAN and is aimed at positioning Expedia's partner sites ahead of what Fletcher and Lynch say will be a significant trend in the industry very quickly.
"If you are not looking into this very soon then in five years time you are going to be behind the curve," Fletcher says.
Other players pioneering the natural language/semantic search in travel include Evature and Hopper - the former recently the recipient of $2 million in funding from Concur and the latter (fronted by Expedia alum Fred Lalonde) also attracting plenty of investment from a string of venture firms.
Travelocity-owned online travel agency Lastminute.com tested a service way back in 2008 called Pronto, allowing users to use free-form text for searches. A beta project, it was never fully integrated into the core search functionality on the site.
NB: Disclosure - Hopper CEO Lalonde is a co-founder and chairman at Tnooz.