A young South African company is bringing artificial intelligence onto the factory floors of some of the world’s biggest manufacturers.
Cape Town-based DataProphet mines data from manufacturers, and the machines they use, to predict where mistakes will be made. The firm then suggest pre-emptive actions to avoid the mistakes being made at the outset.
For example, take any large machine that manufactures car components. Over months and years of steady use, a machine will typically start misaligning and quality problems will start to slip in.
By learning from historic data, DataProphet can build up models to predict where and when these problems will happen, and pre-emptively recommend corrective action. This saves money, says Frans Cronjé, managing director of DataProphet.
Its expertise has been recognised across the world, and venture capital firm Knife Capital, which also manages money on behalf of Mark Shuttleworth, has just invested in DataProphet.
DataProphet was started almost five years ago by UCT friends Cronjé and Daniel Schwartzkopff. After his actuarial studies, Cronjé worked at Bain & Company as a consultant, while Schwartzkopff built artificial intelligence systems, including a poker AI program and a spam filter for a new social network. They decided that machine learning offers major opportunities, particularly in manufacturing, and founded DataProphet.
Shortly after that the company won a contract with Atlantis Foundries, which produces automotive castings for the commercial vehicle industry. This laid the foundation of its expansion, eventually also overseas.
But the company first had to overcome a hurdle that is common among SA companies.
“Initially, we were skeptical about approaching the international market, due to the perceived sophistication (of overseas companies)," says Cronjé. "How could we expect to outperform in a field as advanced as AI against the likes of billion-dollar giants like IBM?”
This came to a head a couple of years ago, when a Japanese company invited DataProphet to bid for a contract. “A number of Silicon Valley and international companies also tendered and we were very careful in evaluating our chance of success when deciding to pursue the bid.”
“We won the contract. It taught me that we have the skills not only to match, but to outperform major global competitors,” says Cronjé.
Being based in South Africa has benefited the group, particularly given its access to a good pool of local talent in its field, he says.
DataProphet’s 30-strong team predominantly consists of data and computer scientists as well as engineers and mathematicians.
“The local universities offer a quality education in the data sciences and we have the advantage of not having to compete with many other companies in the same space,” he adds. “We can build a team of the best talent in South Africa as a startup without going head-to-head in a bidding war against monolithic companies like Apple or Google.”
“There is also a massive cost advantage, salaries and living costs are about a quarter of Silicon Valley - but the talent is just as good.”
The company now does business across the world, including Germany and the US. They spend a lot of time on factory floors to work with managers to implement the machine learning. Cronjé says the company plans to further develop its presence in both markets.