In an interview with CIO Applications, Eric Carey, CEO and Tarka L’Herpiniere, CTO of Predictive Technologies, share their insights on how the company has been able to harness predictive analytics and artificial intelligence to help businesses.
Could you shed some light on the current predictive analytics landscape?
Carey: Predictive analytics is still largely a new topic to a lot of people across different markets and one of the biggest challenges we face concerns the education process. What I mean by that is, most people do not know or understand predictive analytics and how products developed based on the technology help predict data anomalies, downtime, maintenance issues, and more. Companies spend significant money to ensure that they offer uninterrupted services to their customers. With predictive analytics, businesses will be able to reduce downtime by proactively monitoring equipment health in real time. This will help them avoid costly breakdowns that can interrupt production and ensure that the equipment always operates at optimum performance.
Can you give us some insights into your core solutions and how they work?
Carey: Having established Predictive Technologies in 2016, we initially focused on providing wireless monitoring services in the telecommunications industry and, today, have evolved to offer comprehensive network monitoring solutions across different applications and industries. Our suite of next-generation cost-effective solutions consists of two core products— Predictive Suite of applications and Gateway, both of which allow enterprise-level access to quick predictions that help identify and solve problems before they surface. Our SaaS applications allow users to aggregate data across disparate systems, provide predictive capabilities, and facilitate back-end data crunching. Meanwhile, Gateway, is a remote security hardware solution which lets users maintain a secure connection. If the network is ever down, the wireless failover replaces the connection and the gateway continues to gather data from the devices it is monitoring.
However, one of the strongest features is our workflow engine that provides a great deal of flexibility in managing workflow against existing business processes
L'Herpiniere: Our aim is to make AI accessible to businesses to solve real-world problems. We realized very early on that our users were better served by allowing them to define their own problems and giving them the toolkit to train and optimize our AI solutions specifically in relation to the challenges they faced. One of the difficulties we tackled was how to take one of the industry’s most complex and advanced topics, and without diminishing its capabilities, make it so that companies could use it with absolutely no background or understanding in AI. What we ended up with was a form of AI autopilot; around 95 percent of our existing clients use cases can be solved using the autopilot. However, for the remaining 5 percent we give them full manual control.
What are some of the aspects that make your solutions standout?
Carey: Our products have several different innovative features. In addition to predictive analytics, we also support IoT monitoring, issue tracking, and an in-depth reporting application. However, one of the strongest features is our workflow engine that provides a great deal of flexibility in managing workflow against existing business processes. Another strength of the product is our ability to integrate new devices easily. We work closely with OEMs to learn and understand what their product offers and then we write Micro Services for the product’s integration. So, in a very short window of time we can move from request to development to integration of the new devices.
Would you like to share success stories where Predictive Technologies helped mitigate the challenges clients faced?
Carey: We had a customer who was struggling to aggregate different types of devices under a single plane of glass from a monitoring perspective. Using our monitoring platform, we provided them the ability to easily bring disparate devices together. We also offered them the capability to monitor all the different devices without having to continuously log in to them to check if the devices are functioning properly.
L'Herpiniere: In another instance, a client in the telecommunication industry came to us with an issue that involved excessive false positive alarms. The client used our anomaly detection algorithm to reduce the number of false positive alarms from several thousand down to less than 30 real alarms. They then used a series of other algorithms to actually make intelligent decisions about proactively resolving the alarms. All in all, it made a huge difference in our client’s ability to serve their customers while cutting down their operating expenses considerably.
What are some of Predictive Technologies’ landmark achievements and where is the company heading in the coming years?
L'Herpiniere: Specifically, with regard to AI, our team is actively working on interfacing AI solutions with quantum chipsets. Quantum computers are just in their infancy but by their very nature they are perfectly equipped to solve 3D stochastic gradient calculations (the basis of quite a few machine learning algorithms) instantaneously.
Carey: Around April last year, we released our first application, an innovative platform for monitoring, which has now been upgraded with several other different features to handle millions of devices without performance degradation. Moving forward, we are hoping to expand our product suite to offer additional capabilities likes scheduling, inventory tracking, and more. We believe our product is one of the premier AI solutions that will see a roll-out across different industries. Using our innovative, next-generation applications, we want to help companies solve problems that they otherwise wouldn’t be able to solve. So, problem-solving is the key to how we define what the goals of our company will be going forward.