A revolution on the factory floor is taking place using its data-rich environment to leverage Artificial Intelligence and Machine Learning to improve productivity and reduce product defects. And it’s a large and growing market. Spending on AI in manufacturing stands at $2.3 billion in 2022 and is anticipated to reach $16.3 billion by 2027, according to Markets and Markets.
Dataprophet is one such company helping to drive innovation within this market. The Cape Town, South Africa-based company is a developer of machine learning and AI technology to serve the manufacturing industry. The company specializes in optimizing the complex manufacturing processes of key industrial verticals through machine learning and its AI-driven solutions that leverage the existing data streams from the plant’s production line equipment to identify process efficiencies.
“Our vision is to be the leading provider of impactful AI for the machines that make the world. We’re committed to helping manufacturers achieve their Industry 4.0 goals and build a solid foundation for autonomous manufacturing,” says Dataprophet co-founder and CEO Frans Conje.
The idea for Dataprophet come out of Conje’s scholarly focus on advanced statistics at the University of Cape Town, his fascination with the work of Geoffrey Hinton, known as the godfather of artificial intelligence and business experience working for Bain and Company in private equity.
“I joined Bain and Company and really enjoyed their practicality into understanding businesses problems. And so, at that point, in my career I had this grounding in statistics for business, but I felt I wanted to go a bit deeper into the technical space, thinking that you can always come back to business, and I started down a path of doing my masters in statistics around 2012. At the time Geoffrey Hinton and the lab out of University of Toronto started talking about the results they achieved with ImageNet. And so, I pointed myself in that direction. I saw this budding field with AI and wanted to explore that,” says Conje
He had the intention of pursuing the nascent AI field by either building a very good team within the field or joining a good team. Turned out that he built his own team. He connected with a couple of friends who were also beginning to see the potential of neural networks as a technology step change, and the results that Hinton was achieving at his lab at the University of Toronto. And so the founding hypothesis of Dataprophet was centred on the idea of taking academic lab results into the real world. The company was first incorporated in 2014 as a consulting firm helping companies understand machine learning and AI’s potential.
“In 2017, we started working in the manufacturing space. And it’s a fascinating environment. We’ve done a bit of work in a couple of other spaces, but what I really enjoy about manufacturing is the huge quantity of data that underlies the deep physical process. No one disputes that this information is coming off the physical world,” says Conje. The small team at Dataprophet then turned what they learned through their consulting projects to build a software product that manufacturers would use on their own.
“We actually like to call it a product because we do action all the way down from the machine, right through back to the feedback. That’s all a solution, as opposed to how I characterise platforms is that you as the customer have to bring users onto the platform, and they build value on it. And we talk about end-to-end prescriptive AI, which is to say, we ingest data all the way from your machines, all the way through, and then it starts feeding back to your operators without them having to be data scientists, without you having to have expert knowledge in that,” says Conje
Today, Dataprophet AI software is utilised in manufacturing facilities around the world. The now 50-person team is working with almost every kind of major manufacturing region spanning Japan, China, Europe, North America, South America and within its home country of South Africa. Their AI-as-a-service software currently is ingesting over 100 million unique data points on a daily basis across all of their sites. “I think the important aspect is the magnitude of the impacts we’ve had across different environments. So, across these different spaces, our customers have averaged a 40% impact upon production KPIs, whether it’s reduction in defects or production improvements,” says Conje.
Despite the company’s growth and customer success, it has only attracted a total of $6 million in venture funding to date given its location far from the major VC centers. It’s A round funding raised in 2018 was led by Knife Capital with additional investors Yellowoods Capital, the Industrial Development Corporation of South Africa (IDC) and Norican Group, a leading foundry engineering and equipment company.
Conje grew up in Cape Town, the son of a well-known and successful custom furniture designer and manufacturer. But from an early age, Conje was more interested in numbers than business. He went to the University of Cape Town where he earned a degree in actuarial science and where he would later go on to earn his masters in statistics.
“A lot of my motivation for building a business factors on wanting to see the technology through to fruition. And I guess that comes from a fortunate upbringing and observing my father follow his passion and craftsmanship. He’s going to do furniture manufacturing very well because that’s what he really likes. He’s very well recognised within the context of South Africa for the work he does. I think it’s just a bit of an underlying assumption as to my worldview because I’ve seen it happen in front of me,” says Conje.
He first learned to apply his love of numbers to the business world when he was hired by Bain and Company in their Johannesburg office. He would leave Bain after a year and then assemble the team to co-found Dataprophet in 2014.
As for the future? “The journey that we’re on is to fully realise the technology supporting much more autonomous factories. An ambition of ours is to bring it the point where the intelligence is providing instructions into the machine itself, making all the necessary adjustments to ensure that the plant is running at the highest possible efficiency,” concludes Conje.