Jakubowski-Field Test- Airport Bot
By Max Jakubowski
May 12, 2017
For my field test, I created a chat bot that acts as a “FAQ bot” for the Syracuse Airport. I had worked on this idea during an early assignment during our class but thought the idea could be expanded. The goal of this messenger bot is to quickly inform patrons about the Syracuse FAQ, especially when sitting on planes and waiting to leave the plane and go to baggage terminal.
Airports can be overwhelming to people especially those who travel infrequently. Some of the main things people want to know about a new airport they are traveling to is gain information, restraint locations, and baggage claim and transportation. In doing some preliminary research, I found that many airports, including some big ones on the east coast, don’t have a chat bots that lets users quickly access information, instead of waiting to ask an airport or airlines employee. I decided to make a test bot for the Syracuse airport and provided all the relevant information.
I used chatfuel.com to create the Chabot. This website has a simple interface and allows for me to input answers, links and images, all necessary items for the airport bot. I wanted to use the base of the bot I already used for a prior NEW 300 assignment, but I instead had to make a private Facebook group page in order to share the link more effectively. (Since a private Facebook group link can link directly to the messenger, as opposed to the group itself)
Another big feature I added in included “blocks” that are made to give the users of good questions to ask, based on the information available to the bot. I set the bot up into three blocks to pick from, with ‘Gate information”, “About the airport”, “Transportation” as the main headers where people could file their questions.
I initiated the bot back in mid-April because I knew I would need multiple users at the Syracuse airport.
I sent links to friends and family who I knew would be flying through or to and from the Syracuse airport. In total, I got 26 different users over the span of 1 month. I was always texting people I knew were flying to/from Syracuse to try my idea! A majority of the questions (asked by the users) were focused on gate number and plane departure time, but there were a few general questions as well.
One noticeable trend I found via early user feedback was they did not like the “menu main “function of the 3 main blocks to pick from. They preferred to ask the bot a question and then receive an answer based on the word association. After dissembling the menu and instead just having a typical “welcome” message, users generally liked this opening interaction where they could ask anything to the bot.
One item that was very instrumental was the ability to embed images in the bot answers. One user asked for a terminal map of all restaurants, and the sent a message contain the terminal map for the user to browse. An idea for bigger airports would be to use direction to a certain airport store or restaurants, especially if the user can specify the gate. It saves the traveler time of trying to find a certain spot to eat or meetup with fellow travelers.
Another item I noticed that was asked commonly is if they could pay for parking via their smartphone. That is a great feature addition to add, where the parking payment method could be easily linked to the user so they could pay for parking when arriving home (right from the landing plane) and be on their way without having to wait to pay.
Below is the overall data from user responses.
26 overall users
94 total messages sent from users
91% of time Bot responded with answer. Others sent default “I don’t understand..” answer
4 times obscenities were used by users
45% of answers involved gate information
28% mentioned either “transportation, parking, taxi or bus”
26% asked about restaurants
Only 34% interacted with the block menu before I made the change to go to open menu message.
9% of user messages included the words “dumb, stupid, not what I said.” LOL
I do believe this a great idea that airports and transportation companies should look into. A load of data is necessary to do this and make sure customers are given accurate and quick information, emphasis on accurate.
The key thing is to run the bot through Facebook, an establish platform that many users already have access to. Users should be able to quickly find the bot on FB messenger and then type in their question. With nearly 50 different answers prepared for this bot, imagine how advanced and in-depth a real bot would be for a major airport with LaGuardia. With the proper resources, a bot could answer thousands of customer’s questions, everything imaginable a traveler could ask. One potential negative thing is the bot will to have to be updated every day due to activity surrounding the airport or updates that can not use automation.
Overall, bots help bot could help save plenty of time and resources for airports, while helping out passengers tremendously by supplying a plethora of on the go information, right in the fingertips of every traveler.