Read the latest from the Web Foundation

News and Blogs

Voice Browsing: A Farmer’s View

Aman Grewal · August 30, 2011

It has been a while since we shared the latest developments on the VBAT project. We were busy preparing for the focus group with farmers in India. As part of our preparations the first task was to come up with our technology offering. Stéphane and Etienne spent considerable part of their time developing the IVR application for LifeLines India channel over past few months. The primary objective of the focus group was to investigate how farmers respond to FAQ-enabled call-flow and the same information when presented by a Text to Speech (TTS) engine.
We also aimed to gather sufficient data about the type of questions to be included in the eventual field study due later in the year.

As always OneWorld South Asia staff was enthusiastic about this phase of the project. Owing to train delays the core team of farmers and their coordinator was late by few hours, but this did not dampen their spirits. After a quick demonstration of the new service offering it was time to let them browse for agriculture information on their own, while we got down to observing their reactions and views. We gathered a wealth of information from the farmers. Some of the important outcomes we want to share here;

1. TTS: In majority of the cases the TTS was too fast and hard to understand. In cases where the TTS was relatively slow and with lesser content the farmers were able to understand the information being provided easily. But overall with the existing system they were dissatisfied with the TTS output and could not relate to it. This was the most important lesson learned for future development.

2. TTS not inherently untrustworthy: Despite the fact that they found the TTS hard to understand, the farmers were equally willing to implement the suggestions provided by the TTS system as those obtained from the recorded recommendations.

3. IVR use not an issue: Farmers were comfortable browsing the menu. There was no instance reported for wrong menu input by the respondents.

4. Voice input preferred over pressing the buttons on the phone. However we are of the view that this should be treated with caution as there is a lot of background noise in the environment where farmers work. It is impossible to create a quiet room environment for such usage.

Another key observation beyond the focus group was that farmers feel the need of accessing latest market information within LifeLines. One very interesting observation was “how LifeLines can give me information so that my eggplant crop is ready in February, one month before others and then I can get better prices.” Not only are they aware of when prices are high but also are sensitive towards the need of spot prices. For them once they are out with their produce there is no coming back without selling.

Over the next few weeks we will synthesize all the results and prepare ourselves for the next step of field study and deployment. As always we will share our experiences with you along the way.


Your comment has been sent successfully.