Mobenzi - giving township SA the opportunity to work from phone

The Durban software development company Clyral adapted this model into software that gives South Africans the opportunity to earn income by completing tasks using their standard mobile phone, from anywhere in the country. This is the innovation behind Mobenzi, an initiative whose name originates from a combination of Mobile (as in phone) and the Zulu Umsebenzi, which means task, or work. The development of Mobenzi has been supported by a grant from the Shared Growth Challenge Fund, a Business Trust initiative to incentivise the private sector to innovate in pro-poor, profitable ways.

The mobile software is easily downloaded, and follows a similar set-up process to the well-known MXit application. Once the application is available on a “Mobenzi Agent's” phone, they can happily open and complete mobile tasks, which have been formulated into simple step-by-step question-and-answer interactions. An agent can open the application on their handset at any time, request and complete tasks, and receive remuneration for doing so.

The Mobenzi service was piloted with great success in December 2009 amongst a small group of trial Agents at a community centre in the township of KwaNyuswa, Valley of a Thousand Hills, KZN. The 25 Agents who took part in the pilot had an average age of 24, were equally represented through gender and all had completed matric with Zulu as a first language and English as a second language. The mobile phone is a central focus of the youth in township and rural communities, and the transition to using the phone for work was seamlessly accepted within the pilot group. The next pilot project is set to be much larger and will take place in April 2010.

Business Process Outsourcing (BPO) often involves tasks that are simple, somewhat trivial and repetitive but are not easily performed by a computer algorithm. In the commercial context, companies are often faced with a large amount of data that require human processing, such as customer SMSes or Twitter messages. To leverage the Mobenzi service a company would submit the raw data and the service would then create the tasks for the Agents to download and complete. The Agents upload their task responses, Mobenzi checks and moderates quality and the company downloads the results. A simple process with a compelling business case and a very unique additional benefit - many people who do not have access to computers, the internet or standard working time can use their phones to earn income in impoverished communities.

What tasks require human intelligence? What is sentiment analysis?
It is interesting to look at the kind of data that requires human intelligence. One example relates to the concept of ‘sentiment analysis', which primarily involves information that is being shared publicly on the internet. So much information is constantly flowing on the web, regarding every aspect of life. Opinions on brands, trends and company services within comments on Twitter, blogs and forums are important for brand, marketing, public relations and reputation managers in large companies to take note of. By processing this information and producing structured reports from it, those in marketing positions can monitor trends over time, get ideas and concepts for marketing new or additional products, identify opportunities for growth and manage reputations and any negative publicity, should it arise.

For example (taken from Twitter):

Hehehe. I like the new Debonairs Pizza ad. With the people in the lift. Now I feel like a pizza...damn you Debonairs?

ok debonairs you're awesome and all but like seriously.. how many more layers you gonna put on your pizza? you're on 4 now. stop ok

Something about #KFC always makes it seem like a good idea... It never is... NEVER

While there have been attempts to automate sentiment analysis, existing computer programmes struggle to understand the subtlety of human language. As illustrated in the examples above, people often use a combination of slang, sarcasm, Textese and in South Africa, possibly several different languages in a single comment. This is why there is value in having access to Human Intelligence to answer questions and convert free text (tweets, SMSes etc) into valuable data for corporate product and service improvement and any negative sentiment management. These are exactly the types of task that are highly appropriate for Mobenzi Agents.

Mobile Researcher Software
Clyral originally created field research software called Mobile Researcher in 2006, a web and mobile technology that allows surveys to be designed online and deployed to standard mobile phones. It was through the success of this software and the recognition that this technology could be used for business process outsourcing that the Mobenzi concept and the win-win situation it engenders between corporate and ‘township' South Africa was born.

For detailed information visit www.mobenzi.com, or contact the Mobenzi team. If you have any ideas for the type of tasks or real world data that could be used in the large pilot project taking place in April, contact the team at Clyral:

+27 31 765 3679
0861 CLYRAL(259725) Within RSA

info@clyral.com
press@clyral.com

Top Floor, Block A
Torino Court
4 Crooked Lane
Hillcrest
3650
KwaZulu-Natal


The Shared Growth Challenge Fund, a pilot program of Business Trust (www.btrust.org), has been conceived to harness and leverage the private sector's capacity for innovation, investment, risk-taking and growth. The fund aims to facilitate the inclusion of poor people historically excluded from participating in and benefiting from the formal economy. Over the 10 year period of the Business Trust a number of interventions have been pursued, with the Shared Growth Challenge Fund being the most recent to pilot market-based approaches to poverty reduction in areas of extreme poverty. The Shared Growth Challenge Fund involves the use of one-off grants to private companies to support ‘pro-poor' innovation and to stimulate private sector investment and risk-taking, to provide profitable ways of improving market access for the poor either as producers or consumers.

The SGCF is managed by Genesis Analytics (www.genesis-analytics.com). Contact Kirsten Woods (kirstenw@genesis-analytics.com or 011 994 7021) for more information on The Shared Growth Challenge Fund.