DataProphet mines data for profit - TechCentral

DataProphet mines data for profit

Richard Craib

Richard Craib

DataProphet, founded 18 months ago by friends Richard Craib, Frans Cronje and Daniel Schwartzkopff, uses machine learning algorithms to solve complex problems. With just six members, the start-up is already providing services to several big call centres across the country.

Craib specialises in mathematics, while Cronje has a background in actuarial science. Schwartzkopff’s interest lies in chemical engineering. The three met while students at the University of Cape Town.

After they graduated, Cronje went on to become a consultant at Bain.

Schwartzkopff ventured into entrepreneurship, developing an artificial intelligence system to play poker against humans.

He also founded a string of successful ventures, including social communications platform FSMS.

FSMS received venture capital funding and signed up several hundred thousand registered users. This provided Schwartzkopff with a vast amount of hands-on experience with processing “big data” in a real time.

Craib, meanwhile, went on to complete a course in machine learning at Stanford University in California. He also studied abstract algebra and finance at Harvard and game theory at the University of California, Berkeley.

He is a keen entrepreneur with several venture capital-funded start-ups under his belt. He trades his own money in the futures market using custom algorithms he built.

Upon Craib’s return to South Africa, the trio founded DataProphet after realising they would be a good fit with backgrounds in machine learning, consulting and entrepreneurship.

The Johannesburg office is run by Veeren Naidoo. Naidoo left his corporate career as an operating shareholder in a payments business co-owned by Absa. He sold his interest in that company to a JSE-listed business in 2008. He then started an online payments business called Triloq Payment Services, which World of Avatar invested in. Veeren also has an online and app development business called Triloq Technologies.

Together, the four men hope to convince businesses of the value of machine learning algorithms.

But what exactly is the technology?

Most of us are exposed to machine learning algorithms on a daily basis. Facebook uses the technology to fill users’ news feeds with posts it thinks they will find useful. Facebook is able to tag users in a photo by looking at other photos of what they look like. It’s also how YouTube is able to suggest which videos users might like to watch.

Machine learning used to be call artificial intelligence. Put simply, it’s a computer’s way of learning from data. It uses this data to develop an algorithm that can determine a number of possibilities, trends or patterns.

Much of DataProphet’s current work is directed at call centres. By analysing historical data, the company is able to determine who is most likely to respond to a call, as well as which agent would be best suited to speak to a particular client, says Craib.

DataProphet has just established offices in the US, in California. It hopes not only to generate new income streams from the US and other markets, but also attract machine learning specialists who are in short supply in South Africa. There are no major machine learning companies in South Africa and most of the available technical talent is based in the US, Craib says.

The team is currently meeting investors to take the business to international markets.

The company hopes to attract clients from the financial sector, where transactional data in banking and other institutions is huge. It can assist in areas such as fraud detection and customer service, he says.

The potential for delving into other areas such as healthcare also exists, with the growing use of activity trackers producing massive amounts of data.

“While few companies are presently using machine learning, we estimate that within three years it will be in standard use in large companies that handle massive volumes of data,” Craib says.  — © 2015 NewsCentral Media

Comments are closed.