In the past few weeks I have been asked many times by partners, customer and friends what I think about the travel booking bots that seem to be popping up like mushrooms after the rain.
NB: This is a viewpoint by Gadi Bashvitz, founder and CEO of OLSET.
Having created the first iteration of automated booking bots and solutions more than three years ago I feel we have a good understanding of what works and what doesn’t work when it comes to bots.
Here's some thoughts for anyone building, or considering building a travel booking bot.
[caption id="attachment_158920" align="aligncenter" width="550"]
OLSET’s text-based booking bot with Any.do (2013)[/caption]
Many companies in the travel industry are jumping into the world of Facebook Messenger bots. Some examples from major players include Skyscanner’s flight booking bot, Expedia’s hotel booking bot and additional solutions from smaller providers such as Alterra.
Unfortunately, the solutions they are building leave much to be desired. These bots feel as if they are created by engineers who do not travel frequently, resulting in significant gaps in the areas of trust and traveler preference learning.
Having a functional bot that can understand the request a traveler texts is necessary, but not enough.
The interaction with the bot needs to elicit trust from users in order to offer a superior solution to what is already offered on OTA and meta sites, or their mobile apps.
Facebook created the carousel structure for these bots about three months ago. Now companies are releasing the bots and while they do work, questions remain. Do they offer a superior way to book travel? Will they be adopted en-masse?
Travel booking is a complex and emotional process that requires more engagement than simple transactions such as buying a shirt, an example Facebook uses in its documentation.
When we released the first two integrations of text-based booking solutions with Any.do and Evernote more than two years ago, the technology worked. But when we spoke with users they gave us a lot of feedback which made it clear that having the right technology only got us to the start line. Two themes which emerged then are still not being addressed by bots today.
Building trust
How do I as a traveler know that you are showing me the best flights or hotels? We have trained travelers to sift for many different options but the Facebook Messenger carousel limits the number of options that can be shown to a traveler. For travelers to trust the limited inventory shown, suppliers can start by presenting them with more options while giving an explanation of why each recommendation is presented.
Over time and after travelers are shown a longer list of options and they select one of the top ones, you can start narrowing down the number of options while maintaining trust with them. But this can only be achieved if the bot can learn what travelers are likely to be interested in.
Personalized results & learning traveler preferences
For bots to be effective, the booking process needs to be simplified. To do this they need to learn what a specific traveler wants in order to remove steps from the booking process. Going back and forth and asking all the questions that a traveler would complete in a website or app is not a viable solution.
With the bots that are available today you have to go back to the mobile site in order to apply filters and narrow down results. Even after you do that, the bots do not have the ability to retain the filters you applied and learn from them for the future.
If bot providers want to offer an effective solution they have to know more about travelers want and enable them to quickly make changes when needed.
So in order for travelers to adopt text booking bots, travel providers will have to take a leap forward in the services they provide and not just emulate what they do on their site or app.
NB: This is a viewpoint by Gadi Bashvitz, founder and CEO of OLSET.
Related reading from Tnooz:
Cheapflights puts flight and hotel search into Facebook Messenger (June 2016)
KLM on Messenger and being a speck in a galaxy of digital giants (June 2016)
KLM tries WhatsApp chat, Hyatt sees Messenger results (Jan 2016)