Welcome to Tignis TV—Physics, Machines, and Data. In today’s episode, Harel Kodesh, a pioneer in digital transformation, shares unique insights and advice based on his time as VP of the information appliances division at Microsoft; CTO at GE Digital and General Manager of Predix, its operating system for the Industrial Internet; and now Operating Partner at Silver Lake, a global technology investment firm. Your host, Jon Herlocker, is an industry-leading technologist, former DARPA researcher, and Founder of Tignis.
First, Kodesh was invited to summarize his roles and accomplishments at the three companies. He explained how he arrived at Microsoft at the beginning of its foray into the mobile world, when the vision for personal digital assistants (PDAs) – the predecessor to smartphones – was forming. A meeting with Bill Gates and the CTO led to Kodesh running the Object Linking and Embedding (OLE) project, which not only produced a platform enabling applications to communicate with each other, but also became transformative in the Windows ecosystem, and the foundation for OLE for Process Control (OPC) and the OPC Unified Architecture (OPC UA). He also managed the development of Microsoft Windows CE, one of the first embedded mobile operating systems, during his 11-year tenure. Kodesh added it was “just pure magic that you can dream big” at Microsoft.
He then served as an executive at some software companies before becoming CTO at GE Digital, an opportunity he likened to being asked to coach at the Super Bowl. He described a meeting with CEO Jeff Immelt where the excitement about going digital was palpable. The vision was a new public cloud for securely processing and analyzing industrial machine data to improve equipment maintenance, and Kodesh led the development of the solution: Predix Platform.
Today, as an operating partner at Silver Lake, he is helping technology companies improve their approach. Because technology had become so central to our lives, Silver Lake was founded to bring the private equity approach used with big industrial infrastructure companies to technology companies, he explained.
Herlocker then asked for more color on GE Digital and its grand vision. Kodesh noted how the biggest bane of industry is unplanned downtime, and conventional time-based maintenance protocols are inadequate because machines operating under different conditions have different maintenance requirements. Predix Platform provided the data structure to collect machine-generated data in the cloud and apply machine learning (ML) and analytics to detect trends and formulate insights for timely maintenance. It is also a digital twin management system where simulations can be performed on a digital twin (replica) of a machine’s structure in the cloud to further improve maintenance predictions and outcomes.
Asked about learnings from the process, he described the importance of being collaborative and realizing “the best minds” are not necessarily in the company; the advantages of focusing the effort vs. setting overly broad goals; why to fly under the radar until ready to launch; and why a small team of good people can be far more efficient and effective than a large team that is less astute.
Herlocker observed how Kodesh was at the nexus of the information technology (IT) and operational technology (OT) convergence at that time. Asked to address the cultural differences, Kodesh acknowledged that while IT allows you to put together software systems, connectivity in OT is between software and machines that are not built to be connected and have unique security and data protection requirements. He also described how IT and OT have different audiences (executives vs. technicians).
The conversation then turned to the role of the cloud, AI, and ML in industrial transformation. Using jet engines and HVAC systems as examples, he explained how these capabilities enable modern software to monitor, compare, simulate, and continuously learn from environmental and operating conditions and their effects on the assets, allowing for smarter recommendations and improved asset management. The benefits are threefold: optimizing maintenance service efficiency and pricing, improving predictive maintenance, and improving machine design.
Finally, when asked what advice he would give to those interested in digital transformation, Kodesh observed it is one of the misunderstood concepts for industry. He elaborated on why digital transformation should be led by someone who has done it before; timeboxed with a set start and end and a means to measure performance through the phases; and inclusive, with each organization contributing in some way. He closed by saying good digital transformation exists, so if you fail, try to understand why, wait a little bit, regroup, and do it again.