T-Mobile T.E.R.R.A.

Inspiration

Questions we considered:

How do we expand AI beyond messaging care? What if it was possible to curtail a customer's negative experience due to real-time empathetic understanding?

Through our research, we found no solution currently offered by T-Mobile for spoken sentiment analysis.

What it does

Our goal was to create a real-time speech sentiment analysis for the Team of Experts.

How it works

Uses Microsoft Azure Speech-to-Text API to recognize speech. Takes data from recognized text and uses Azure's Text Analysis API to give a score to the text. Score is between 0 – 1, negative to positive scale respectively. Informs the user of the score. Provides appropriate feedback, including a prompt asking if the call needs to escalated.

Challenges we ran into

Initially, connecting the two API calls using Javascript as a back-end language for the browser proved harder than we anticipated.

Accomplishments that we're proud of

No sleep. One app. 24 hours.

What we learned

This was the first time we used Microsoft Azure. We became acquainted with the interface; after many failures and mishaps later...

What's next

Explore creating a mobile solution using Node.js and React-Native. A seamless experience without using buttons, i.e. make an autonomous solution. Bridging the language barrier.

Built With

• ajax
• azure
• css
• git
• html
• javascript
• jquery
• speech-to-text-api
• text-analysis-api
• visual-studio