Practitioners of technology in the developing world have long accepted that the "normal" ways of interacting with the Internet, the ways that we Westerners have become accustomed to, don't always apply in the developing world. Mobile is a central and critical aspect to any well thought out technological offering for the developing context. At Google, we focus on it relentlessly. The numbers are clear: mobile penetration outstrips traditional Internet penetration by at least an order of magnitude in Africa. This picture becomes even more revealing when we look at the types of phones that people have. The majority of African mobile users don't own smart phones or even feature phones; the majority rely on phones with only basic calling and messaging services like SMS and MMS. So if we're going to reach the vast majority of Africans, we need to focus there. To this effect, a lot of work has been done around SMS. Google has launched its Health and Agriculture Tips, for instance, but SMS has a number of serious contextual limitations: illiteracy, lack of functional literacy, and local languages not supported by phone keyboards. But there's also another route that hasn't been so explored: What about voice? What about voice as a tool for interacting with the Internet?