- Insights This article was written by a TIA community member. Insights pieces undergo the same rigorous editorial process that newsroom-produced articles have.
Arnav Kumar · · 5 min readHow to get your chatbots talking
Photo credit: olegdoroshin / 123RF Stock Photo.
Nowadays, everyone is creating a chatbot. They are awesome things; they provide convenient access to customers through a variety of applications on social media, websites, or even in person. If you’re wondering what a Chatbot is, here’s a useful intro.
Their utility comes from a conversational user experience, which helps in navigating complex interactions (legal advice, personal assistance, ordering food, financial advice, teaching, etc.). The ability to translate language into software interaction offers immense potential for designing new products and services in the future.
We at SliceBread wanted to get beyond the hype and explore how to build one for ourselves.
As an experiment, we chose to develop an app for movie recommendations. We designed and implemented voice interaction and combined it with basic intent recognition for language understanding. If you’re curious, you can see it in action here.
After exploring both API.ai and Wit.ai platforms, we concluded that API.ai was easier and had a simpler, more comprehensive intent recognition framework.
Here’s how we went about building our very own chatbot.
Step 1: mapping conversations
To start, map a natural conversation that you’re likely to have with a friend who you’d usually turn to when asking for movie recommendations. Here is an example of what that conversation would look like:
User: “I would like to watch a movie with Kate tonight.”
Chatbot: “Sure, do you have any specific actor or genre in mind?”
User: “How about something funny with Tom Hanks.”
Chatbot: “Cool. Tom Hanks has been doing movies for a while – are you in the mood for something classic or recent?”
User: “Maybe something from the last decade”.
Chatbot: “Great! Let me think of some good movie titles for you!”
Step 2: identify the information required
Step 3: language understanding (identify trigger intents)
Step 4: language understanding (designing voice interface)
Step 5: language understanding (mapping user speech to intents)
Step 6: limitations and potential
Stay updated on the go with our mobile app.
Get latest insights with smoother, more personalized experience through TIA mobile app.
