Based on a survey of 14,000 households, the American Society of Travel Agents reports that it is currently seeing the highest numbers of consumers booking through travel agents in three years. The most common reason provided for doing so? Travelers said it saved them time.Though, travel agencies’ business isn’t expected to return to pre-internet levels, online booking has shown travelers the frustrations that can come with booking travel themselves—and that outsourcing the stress of arranging travel can be worthwhile . You can also find more statistics that paint a bright future for travel agents here  .
But the current process of interacting with multiple travel agents and choosing between them is not streamlined, and usually involves filling of clunky forms multiple times. This bot is a concept to show how to get around this problem, by interactively collecting travel preferences over chat and sending them to nearby travel agents for them to respond with customized, creative travel plans. Also, there is no need for the travelers to download any new app or navigate a complex website in a mobile screen. All that they need to do is start chatting with the Travel Agent Connector bot from FB Messenger.
What it does
The chatbot basically collects your travel preferences in a user-friendly manner. It then sends those to a database of local travel agents, who can design perfect travel packages to suit the traveler's requirements and contact the travelers.
How I built it
Messenger platform components were used to build the frontend. For the backend, a RESTful Web Service was built using spring boot. The travel preferences from the users are sent as POST requests to this service. Since Spring supports embedded Tomcat servlet container as the HTTP runtime, the packaged bootable 'fat jar' was deployed in openshift easily using the openJDK S2I image of JBoss Middleware.
What I learned
I learnt to use JBoss Middleware like openJDK to deploy the backend services on OpenShift Online. I also learnt to design user-friendly interface for the frontend using the messenger platform components.