Presenso’s technology uses sensors and machine learning to predict malfunctions in various types of equipment and prevent costly shutdowns of plants and companies
Breakdowns in services, machinery, and servers cost companies and enterprises billions of dollars a year in aggregate damages. Presenso promises to save on shutdown time by predicting malfunctions and crashes. In certain cases, it automatically takes action to prevent a catastrophe.
A short pitch: What does the company do?
It prevents stoppages of machinery with a combination of sensors and machine learning.
A slightly more thorough explanation
One painful problem today in industry is an unplanned stoppage of machinery, and sometimes even entire production lines, as a result of malfunctions. Regardless of the product being produced in the factory, such stoppages cost a great deal of money. The annual damage can total millions of dollars per plant.
Presenso has developed a system that gathers and sends to the cloud large quantities of industrial data in real time from hundreds of different machines and thousands of different sensors. This field is referred to as the Industrial Internet of Things (Industrial IoT). By using architectures of deep neural networks as infrastructure for Deep Learning, Presenso’s intelligent engine studies the behavior of machinery completely automatically.
The engine creates internal ties between events and components inside the machine, and between the machines and various systems on the industrial sites. It thereby learns to detect irregularities in the behavior patterns it studies and finds connection between these irregularities. It is able to predict malfunctions developing in the system. The company’s mission is to alert many plants before a machine breaks down, thereby reducing the cumulative stoppage time to a minimum.
Q: What stage have you reached?
A: The first version of the product was launched several months ago. At this stage, the solution is being marketed mainly to large factories that operate rotating mechanical equipment. The target market is the energy production industry – from manufacturers of equipment and control systems to power stations with fleets of large turbines producing, in addition to electricity, a very large quantity of data. Other target markets include water facilities and oil and gas companies.
Only recently launched, the product is already arousing great interest among large electricity producers in Europe and the US seeking solutions, upgrading, and advancing as part of the fourth industrial revolution, what is commonly referred to as Industry 4.0.
Q: Who are your competitors?
A: There are a number of competitors active in the sector, including large public companies like GE Smart Signal and startups like Predikto and Falconry. Our advantage is that our solution is pure software. In contrast to other solutions offered in the sector, there is no hardware, and no need to install additional sensors. The solution is unsupervised. It was designed to be agnostic to the machine and the type of data being analyzed. There is therefore no need for information from the customer’s engineers about how the machine works. This fact makes installation quick and easy. The same system is suitable for a large number of different machines, and does not require change or adjustment. It does not impose a burden on the customer, and there is no need to describe the machine structure to us.
The disadvantage is that the solution we provide is unsuitable for relatively small machines that do not have enough sensors installed on them.
Q: How do you plan to make money?
A: The business model is built as a SaaS basis where the customer buys an annual license, and pays per machine per monitoring month.
Q: Have you already received investments? How many? Who invested?
A: The company raised $2 million last June. The AfterDox fund led the round, with participation from Janvest Capital Partners, SeedIL, and other angel investors. The company also received a grant from the Office of the Chief Scientist.
Q: Who are the company founders?
A: Eitan Vesely, a mechanical engineer thoroughly acquainted with industry; Deddy Lavid, who worked at Rafael for eight years, including a period as a algorithm team leader for learning systems; and Dr. David Almagor, founder and CEO of startup Panoramic Power which sold for $65 million, founded Presenso in 2015.
Q: How did you get the idea?
A: The idea first arose when Eitan, as a support engineer at Applied Materials, traveled around the world a lot to repair machines. Deddy, who was looking for an idea for a thesis at the time, proposed the idea of using machine data to predict failures. As an algorithm and artificial intelligence specialist, Deddy was exposed to the industrial sector in the course of his work. As a mechanical engineer, Eitan was thoroughly familiar with the repair and maintenance of machinery and the need in the market. After a two-hour meeting in a coffee shop, they got started. Deddy stepped in later and brought with him a lot of experience in starting companies and familiarity with the needs of conventional industry.
Q: How many employees do you have? Where are your offices?
A: Located in the Matam industrial zone in Haifa, the company has seven employees.
Check out Presenso’s video
Presenso is taking part in the Startup Arena, Geektime’s startup competition taking place for the ninth consecutive year in the framework of the Geektime Conference. Past participants include companies such as Kaltura, Cyactive, and SalesPredict, among others. The 2016 Startup Arena competition this year is sponsored by Altshuler Shaham Benefits.