To start with I recently joined the Glific Team and started working with the NGOs in my second week. What I learned about these NGOs is that they all are trying to solve different problems and to be able to help them with the product which we have gives me immense pleasure to solve their problems. 

During the sprint in Garudmachi we got to show the features we are building for SunoSunao. To give you a context, the SunoSunao team wanted a feature where users can send an audio clip and based on that audio we will convert the audio to text and give them a suggestion based on that text. Here we are trying to show what we have done and tested so far.

Google ASR

Let’s talk about Google Automatic Speech Recognition. So what we are trying to do is here we are giving users a feature where users can send us the audio format. 

The reason we are doing this:-

1. AI power voice-added chatbot  to empower women and vulnerable groups. 
2. Create awareness minus the privilege of literacy, i.e. needing to type in a request or reading the response. 
3. The bot aims to collect audio messages from the women who are the target audience and send them appropriate video responses based on their query. 

Feature Walkthrough

1.  Google ASR which takes the audio and gives us the confidence level of the audio.
2.   It can support multiple languages and regional languages. Which you can read about here language support

3. After the conversion of the text which we get we will train them through our machine learning algorithms which will check the intent of the text and based on that we will send the video.

The flow will be something like this

This is the sample flow which we have created in the glific

To show you what this does I’m posting a link here which will give you the demo for the google ASR and how we’re using it with the flow.

Demo Link

What next:- 

1. We will continue to explore more. How we can make this better and easy to use for our users.
2. We need to do more testing with different regional languages.
   

As we are done with the first phase of Google ASR. We are looking forward to people using it and giving us feedback.