What is Question Box?

Question Box is an all questions answered service that attempts to democratize the world’s information for all the world’s people. Unlike many services that target only mobile or web users, Question Box takes into account the fact that some people are illiterate, some people are too poor to afford even a mobile phone and some people (often times women) are shut out from communicating with certain people or information sources. The service was started by Rose Shuman a few years ago when she got the idea to offer the internet and information found in things like Wikipedia to people who couldn’t read or who were otherwise disenfranchised. Since then the service has been piloted in many locations in India and is currently being piloted in Uganda.

It was around the time that Question Box entered Uganda that I joined as the CTO. It was a happenstance thing that resulted out of an interview request with Rose. Twelve months later and things couldn’t be better. Our pilots are still going strong and we’re extending our reach everyday. In my new role I’ve launched initiatives to support both web and mobile (after all, some people can’t hear) and our efforts have been rewarded with support from groups from all over the world. Our current focus is how to scale these services and how make them fully sustainable.

Thinking Inside the Box

In our Uganda pilot with Grameen’s Applab, we rely a great deal on Appfrica Labs’ administrative assistant Barbara Birungi. While the guys in the incubator are innovating with software and technology, Barbara is what I call a logistics innovator, managing the three Question Box operators (pictured below) as well as communication with our pilot partners and various contacts around the globe. Her job, partially, is to organize and make sense of the massive amounts of data we’re collecting.

The operators take calls from various parts of rural Uganda from people asking all types of questions. Here are some actual examples of questions asked: “What causes rotting of cassava roots?”, “Why are my tomato leaves turning white?”, “Can a mother with HIV pass it on to her baby?”, “How can we control soil erosion in our village?” These questions can come in the from any number of people speaking the 14 languages and regional dialects of Uganda. When we hired the operators we made sure that between the three of them, they could understand these languages and respond in the common tongue.

Searching Where Google Can't

One of the big innovations for this launch was an offline search engine and database developed at Appfrica Labs specifically for the pilot. This custom service performs two functions, querying an offline repository of information (collected from various authoritative sources on health and agriculture) and logging the questions and answers for easy retrieval. Why is this so crucial? Because currently bandwidth constraint in Uganda limits the ability to search the web quickly. This service relys a great deal on human interaction in real-time, so we wanted to make sure the most important feature (finding answers) was possible even when our internet connection is down or the power was out (and we're relying on our inverter).

The solution was a cross-platform java client that searches a repository of documents, government statistics and research papers that are added to the system. Multiple computers with the client connect to a central server that contains the database. These documents include vertical categories like health, agriculture, science, history etc. in addition to the ocassional source related to sports and entertainment. The real value in this search engine is that it contains a wealth of data that only pertains to the local area. For instance, a caller can call in and ask "Can I grow Matoke in Jinja?" or "I don't feel well, is it safe to drink from the water source near the market in Kibera?" While search engines like Google's can point to sources that cover the topics on the web, a lot of this information isn't even available on the internet. Until people in rural populations are online enmasse, there will never be much local knowledge available this online. Question Box not only makes it available for real-time search but it goes pretty deep allowing quick access to research on any subject. More importantly, the reports we can generate allow us to see trends in questions by district, gender and time of day.

Beyond that, the search engine is 'living', meaning all previous questions and answers logged are index back into the database. Theoretically, the longer the service is used, the better and faster the search gets.

Bottom to Bottom, Strength to Strength

Perhaps, the most exciting aspect of Question Box is that it's the perfect example of cross-cultural solutions to local problems. The Question Box was first piloted in India in a very different form. Actual boxes that were placed on walls. The box has a mobile phone in it that's pre-programmed to dial a specific number (the local QB call center). Users approach the box and push a button that triggers the call. Once connected, they ask an operator a question in their local language and get their answer back in that same language. In India the operators search the internet, they simply place the operators in a place with a good internet connection. But the software solution developed in Uganda offers the opportunity to improve the service in India as well by searching locally focused verticals and logging queries offline instead of relying upon the masssively unfocused world wide web. Likewise, the mobile-solar question box developed in India is an ideal solution for use in Uganda's semi-tropical climate. With these two bottom-of-the-pyramid countries sharing knowledge and experience, there's no limit to the concepts that can be tested and implemented.

Barbara Birungi (Administrative Assitant) and Moses Mugisha (Lead Developer) of Appfrica Labs

_{Disclaimer: Jon Gosier is the CTO of Question Box and the search product described was developed by Appfrica
Labs.}