As content consumption gets ever more personal with several platforms providing access to a plethora of digital content, content discovery has become a daunting task for the average consumer. We identified that a seemingly simple task, that is to select a movie title to watch at one’s leisure, can take up several minutes.
The process to search, discover, explore and arrive at a suitable movie title is tedious and requires several web searches. Many people turn to asking for recommendations from their friends. Our hypothesis was to establish whether we could recreate the interaction around movie recommendations.
A convenient solution was to develop a virtual assistant that could interpret the user’s preferences for the type of movie s/he might be interested in watching at that time by engaging in a conversation to understand the user’s mood, prefe-rence of a genre, actor or time period of the movie. This was implemented by designing conversational graphs that could gather the requisite parameters, while maintaining the context of the conversation to enable a ‘life-like’ con-versation with the machine. Once the conversation came to its natural conclusion, the program would search through a database of movie titles and make recommendations.
With this intelligent virtual assistant a user can converse in the following way:
User: “I would like to watch a movie with Kate tonight.”
M.A.: “Sure, do you have any specific actor or genre in mind?”
User: “How about something funny with Tom Hanks”
M.A.: “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”.
M.A.: “Great! Let me think of some great movie titles for you!”
The program then goes on to search and recommend movies that fill the following criteria, “Tom Hanks, last decade, comedy” to come up with a list of appropriate recommendations.
With this mobile application, users can interact with a NLP-powered virtual assistant to significantly reduce the time taken to discover a movie title, and spend more time getting entertained instead. We wanted to showcase its potential in information access that could extend to applications from patent and news to medical database searches.
The potential for natural language interfaces is immense with solutions coming up to complement customers service experiences, create virtual assistants for scheduling appointments and responding to emails, or automating HR screenings while interviewing new candidates. And the ability to integrate these interfaces into software workflows by understanding user intents and mapping them to actions is creating a huge opportunity to develop new solutions that can help companies reduce long term operational costs.
You can check it out here: