Episode #458
458- Stop Turning the Light Off w/John Errington
John Errington, Director IT/IS at Pengate Handling Systems, tells his teams something they hate to hear: turning the light off every day is the thing to automate, and their time belongs on why the light keeps going off. With more than 27 years in IT, he also explains how he learns a new business from the people with their boots on the ground, and how AI took three to four days of pulling major-incident data together down to a 20-minute process while his problem managers still checked it for accuracy. John Errington fell into IT by accident. A couple of years out of high school he was working in records management at a pharmaceutical company, where knowing how to turn on a computer made him the de facto local IT guy. At GSI Commerce, a Linux shop he joined in 2010, a new leader pulled him into IT leadership "kicking and screaming," and his own light bulb moment came when he got a script put on an alert his team kept clearing by hand. He often uses the light analogy with the teams he manages, and they hate to hear it: the light you turn off every day is the thing to automate, and the value add is fixing the core of the problem. Four weeks into a new job in material handling, he explains how he learns a business he doesn't know yet. It starts with the people with their boots on the ground. Here, that meant ride-alongs with the field service techs, the people who "keep the cash register open." John also gets into how silos breed a not-it mentality on outage calls, and the SWAT team approach John's operations teams used instead: everybody jumps in and stays until they're cleared. John sees AI the way he saw automation, as a tool that needs a human at the decision-making part. In his last job, AI postmortems of recorded major-incident calls took much of three to four days of work down to a 20-minute process, and his problem managers still reviewed all of it for accuracy. Key takeaways: If your team clears the same alert every day, automate it, then spend the time you get back on why it keeps going off.; Learn a new business from the people with their boots on the ground: ride along with them, then take what slows them down to whoever owns that process.; On a major-incident call, have every team report on its own area and stay on until you clear them.; Let AI draft the postmortem from the call recording, and have your problem managers check it for accuracy.
