Recently, there has been a mania for chatbots in the ecommerce world and, as you would expect, the travel sector was heavily involved.
NB: This is a viewpoint by Valentin Dombrovsky, chief alchemist for Travelabs.
Over the past few weeks we’ve seen chatbots from Kayak, Musement and Yatra – and many more have been launched over the three months that have passed since Facebook announced its Messenger Platform launch.
Actually, it all started a bit earlier with Telegram Messenger arguably the one that announced “bot revolution” about a year ago. Before that, Mary Meeker had written about messaging apps becoming universal platforms in her yearly internet trends report, although she didn’t predict bots developed by third parties becoming the core element for such platforms.
Part of me thinks that many firms are developing bots just in order to get coverage in the media. But if you want your travel bot to be more than a PR exercise, there are some big problems ahead, one in particular.
For context, we have had semantic search enabled interfaces in the travel industry for a while. Take Evature – a company that has been developing its Expert Virtual Agent since 2009 and has even launched its own “botkit” recently.
But we haven't seen a rapid rise of travel semantic search, unfortunately. In 2012 there were suggestions that semantic search would become standard in the industry by 2020. Four years on, we don't see much progress in its usage. And here we come to the question “why?”
I think the answer is quite simple and lies in one word – “consistency”.
Let’s take a look at few search boxes on some of popular travel websites.
What you see here follows a basic UX principle: present a form with a few simple questions for starters and ask everything else later. You’ll see all sorts of filters on each of these websites on the following pages but initially the main things that these sites need to know in order to answer to your basic needs is who, where and when.
It might seem that such interfaces create unnecessary constraints for the traveler. But in fact, they’re a way to get the information needed to begin the search. They’re turning chaos into structure and these interfaces serve the same function as a “smart assistant”.
Users know what should be typed into these forms and understand what they will get as a result - and that’s what consistency means. It's what we have come to expect when we engage with machine search.
The question is whether bots (or semantic search) can provide a similar user experience. I am really not sure what the answer might be and how it will be presented when I ask this or that bot to find me certain flights.
As things stand, the bot asks me: “What do you want?” and expects me to respond, without me knowing what it can do.
So, I am staying with the tried and tested forms for a while. I believe that there might be some situations when bots might be convenient (suggestions welcome in the comments below!), but these seem to be niche use cases at the moment.
When we have more information on how travelers are actually using the bots we can start to make conclusions about their long-term impact on travel.
NB: This is a viewpoint by Valentin Dombrovsky, chief alchemist for Travelabs.
NB2: Image by Matrioshka/BigStock
Related reading from Tnooz:
What CMOs need to hear about chatbots (July 2016)
Are bots hot or not? (June2016)
Allo, can Google have your attention please, travel execs…