Inspiration From being on the telephone with robots and getting useless information from automated systems, we decided to create specific responses to allow customers to feel more engaged with the company when they are upset. Hopefully, users are able to have a more satisfying response to their tweets to our company.
What it does Our program takes in specific tweets the user inputs and generates a specific response to their query or complaint.
How we built it Using python and several different modules and functions, we developed a program to first test out the overall sentiment to analyze the overall sentiment users are feeling towards a company. It will either be positive or negative to show its overall standing. We then created a driver file to input a test case to generate the responses with added personalization to allow the user to feel comfortable with our company
Challenges we ran into
1) Creating a dataset from Twitter and making sure the outputted response makes sense with the given content 2) developing a loop to handle tweets that do not have keywords in them
Accomplishments that we're proud of
1) Ideation 2) Creation of data set
What we learned
1) Learned about how to enhance customer experience and how much business and technology can intersect one another
What's next for MetriCX
1) Creating a timestamp to ensure the response is given out within a specific time frame. (5 minutes) 2) Use it in conjunction with Twitter auto-responder to ensure the solution is applicable to real businesses