'Industrial AI’ is a synonym for advanced analytics in Industrial Internet of Things (IIoT). The economic impact of the IIoT has already achieved $1 trillion and is rapidly growing. The Industrial AI applications process IIoT data to impact business processes and provide added value.
Internet Revolution that happened over last three decades created Internet of People (IoP) where most of the data is generated by people. The IoP impacts 10-15% of the economy. The on-going Fourth Industrial Revolution (a.k.a. Digital Transformation) is impacting the rest of the economy, a much bigger part. It creates Internet of Things (IoT), where the data is generated by Internet -connected devices equipped with physical sensors.
The term IoT is often used in narrow sense, to describe connected low-cost end-point sensors and consumer-related devices. For Digital Transformation of industrial enterprises, a term Industrial IoT (IIoT) is adopted, to describe connected industrial high-cost assets with embedded sensors.
The IIoT is predicted to impact over a half of the economy in the next decade or two. The Industrial AI market provides computational applications that are at the heart of IIoT. So far, comparatively more effort is dedicated to the IIoT computing platforms including sensor endpoints, computing hardware, software technologies, and data management. To be sustainable, the industry investment into the IIoT platforms requires demonstrated returns. The value is returned by the Industrial AI applications. These AI analytical applications process IIoT data to impact business processes and provide value added economic or social outcomes.
What does the industry need from the Industrial AI analytics applications for the IIoT, to get the value? First, the analytics need to be automated and scalable, data driven. The scalability means new application development is possible with little engineering labor. The AI provides automation of special skills making broader workforce available. Second, the IIoT analytics must support industrial systems that are mission critical, have high cost or failure. Such analytics require rigorous engineering processes to develop, verify, validate, and operate.
Current landscape of AI/ML technologies and Data Science skills has been established by non-critical IoP applications that focused on generality of the tools and fast data-driven development foremost. A representative example is on-line advertising, where clicks-through rates are on the order of 0.3%. In academia, such technologies and applications are usually purview of CS departments.
The landscape of mission-critical analytics in the IIoT world is defined by control systems, and critical Industrial AI apps that require high reliability and accuracy. Representative examples include aviation with its 99.9999% reliability requirement and energy that must be 99.97% available.
The IIoT computing started convergence of IT and OT (operational technology). Traditionally, OT included embedded and control systems that were developed to support highly critical applications such as control of aerospace or ground vehicles. In the IIoT, the OT data is transferred into IT systems (e.g., cloud) enabling the Industrial AI.
The emerging Industrial AI apps combine the requirements and features of the two domains and are somewhere in the middle. A representative example is Predictive Maintenance, where decision support accuracy of 95% is expected.
Capital-intensive industries rely on their vital assets. Failures and unavailability of assets mean missed schedules and lost revenue. As such, industries spend billions of dollars annually and 60-80% of annual budget on Operations and Support (O&S). Mitek Analytics provides AI/ML apps for IIoT that allow achieving optimized, sustainable MRO strategies and run fleet of assets at peak performance and availability.
To summarize, Industrial AI provides significant values for the New Industrial Revolution. At scale deployment of the IIoT apps is supported by Industrial AI/ML use of data-driven modeling and optimization. Mission critical apps of the IIoT require rigorous analysis and provable algorithms in Industrial AI. These are the technologies where our team excels.