Eva: Natural Language Search Revival

A new product called Eva (Expert Virtual Agent) aims to become your virtual travel assistant. Unveiled at the recent PhoCusWright Travel Innovation Summit, Evature is an Israeli company marketing the Eva online travel search assistant. Winner of the DEMO Award for Most Game-Changing Travel Innovation, Eva aims to please a billion travelers soon.

Natural language search. How this term brings back my own memories of “game changing” innovation. Back when Powerset instigated this term into the matrix of the so called “search engine wars” – hakia and Powerset (later Bing) we locked in deadlock in their efforts to dethrone Google. Alas, my rocket scientist friend Riza Berkan at hakia, and good ole’ Barney Pell over at Powerset never quite got over the long tail query hump back then. Barney did take home tens of millions from Powerset’s sale to MS, but Google still rules the roost, last time I looked. Enter Eva (and a score of other smarty pants search tools) in the latest high potential and high suspense “smart” technology pitch.
It’s not secret the OTA’s need some super secret weapons to combat the rising tide of service provider and customer resentment. But, is Eva or any other high tech gadgetry going to do it? Well, not unless this scientific marvel gets a lot smarter. Founder Barry Volinskey is cited by Johanna Jainchill on Travel Weekly as using this query as an example one put to Eva:

“I want to go from Moscow to New York next Monday.”

Eva, according to Volinskey, basically translates questions input by users, into language typical search engines can understand. Taking querries as nebulous as “I want to go someplace warm the first week in January,” and converting them into an XML query that Google and others can use to produce flight and hotel results – this is the “rub” Evanture is setting out to undo for travelers.

Gene Quinn, CEO of Tnooz, one of the panel of judges at the PhoCusWright event, went so far as to suggest; “This is just what the travel industry needs.”

Claiming the obvious, that natural language search has implications for search, voice recognition, and so on, seems a bit naive from here, taking in all that Google has and is accomplishing. Before we all get too excited here, maybe AI and other “smarty pants” technologies should be considered more? My first negative here is, “just how dumb do these people think travelers are?” Personally, search relevance is as much about knowing what to ask as it is some robot figuring out what the user may not know in the first place. Put simply, how many customers are so stupid they can’t look up a flight on Delta? Holy God.

Okay, maybe I am immune to claims of AI and semantic wonderment. My partner is the all time Queen guru of Google semantic technology releases. And, when I spoke with old friend Riza Berkan of hakia the other day, reliving the essence of relative search value, we touched on knowing the question being more valuable than possible answers even. Of course, this is where some Eva missile scientist will come and proclaim; “This is exactly what we are doing, reframing the question for you!” Back to my original precept then; “Do the criminally illiterate really fly that often?” Below is a result imputing the aforementioned; “I want to go from Moscow to New York next Monday.”

I: Passenger – One Person
Moscow: Moscow, Russian Federation (City), Moscow, Idaho, United States (City), Moscow, Vermont, United States (City), Moscow, Missouri, United States (City)
New York: New York City, New York, United States (City), New York, United States (State)
next Monday: November 21, 2011
November 28, 2011

The result has been reformed to something meaningful for further organizing into a travel request. Ho hum, then we get into what will no doubt be “magical” light speed permutations being permeated into possibly positive potential passenger pleasing purposeful potentialities, and by that time we could have Googled Flights from Moscow to New York and ended up with the graphic below. As it turns out, Delta flies out of there every day at prescribed times.

And for those arguing less prevalent destinations, the permutations for Eva (or any search, even human) are massive. Eva, or any search engine technology, is going to return a mix-match jigsaw puzzle of potential results. Look at it this way, if you can’t express in Google language what the hell it is you want, then how is Eva going to scour the web (maybe using Google) to find out? She or it cannot.

What the world needs is not voice recognition software, or even natural language artificial intelligence, what we need are less choices – not more. We have all become so dissatisfied with the marketed massive hunk of worthless junk out there, many have lost sight of what value really is. I remember a time when 10 candy bars dominated people’s choices. And you know what? Every, single, one, of, them, tasted better and was of higher quality than what we get today.

Does this all mean there’s no room for Eva? Of course not. Greater understanding is another need human beings have. I just think we would all be better served applying the proper technologies where they can do the most good. Hal 9000 as a way to find a flight or hotel? Well, maybe that would be better than trusting TripAdvisor reviews, but aren’t we getting carried away? Maybe we should teach more people how to ask? Aren’t we teaching the teacher to teach teaching?

For more info on Eva and other emerging technologies for travel, please visit the PhoCusWright Innovation Summit event pages. I leave you with Evature’s video of the Eva innovation.