We sat down with guest speaker Rich Lane from Forrester to discuss the core issues IT Operations leaders face as they evolve towards intelligent, automated operations (AIOps). Then, we kept him for a Q&A to dive deeper into the biggest challenges facing the adoption of AIOps as a strategic driver.
Q: What are the primary use-cases Enterprises are trying to solve with AIOps?
RL: There’s a few that come to mind from the get-go. First thing that comes to mind is false positive reduction. On average, 60% of the alerts that generate tickets are false positives. Add up the amount of time wasted investigating these alerts and you realize that there must be a better way of handling all the telemetry. The second big one I see is the shift towards understanding the day in a life of a digital user. Not just from an IT perspective, but sales, marketing, and product development. So, having the ability to map out user journeys and tie it to user sentiment helps build better services and deliver better CX.
Q: How has accelerated cloud adoption impacted tools modernization efforts?
RL: In the context of the pandemic, cloud migration plans that were 18-24 months out were realized within the first few months. This exposed a lack of tooling to monitor cloud instances remotely and at scale. Add to this the increased utilization of resources to meet customer demand and remote work and legacy tooling couldn’t keep up.
Q: How does one get started with their journey to AIOps?
RL: I always tell folks to take a step back and redefine what monitoring means for their enterprise. Everything that we do now in operations should be geared towards delivering the best customer experience we can. That means we need smarter tooling with much more automation to bridge the gap between how much telemetry is coming in and how small IT teams usually are.
Q: What is the nature and role of automations within AIOps?
RL: Automation is a huge part of AIOps. IT teams spend far too much time doing low-level tasks that we should have as automated remediations. The more situations where we can use automation, the more effective teams can be in quickly clearing incidents. For instance, think about how long it takes a Level 1 support agent to do basic troubleshooting, open a ticket, update the ticket, find the right support team, etc. If the monitoring system on detecting an alert is integrated with ITSM, CMDB, CI/CD, etc. the time to when the correct support engineer is contacted is seconds versus the industry average of well over an hour.
Q: What do you think are the biggest challenges when it comes to adopting AIOps?
RL: There’s two: culture and showing ROI. If you’re doing AIOps properly you’re going to change how teams engage, areas of responsibility, and process. All this must be thought of before you even begin the rollout phase. The second area is just trying to get the contract signed. Software isn’t cheap, but the cost of doing nothing is much greater. Add up all the lost productivity from firefighting, the lost revenue, and the inability to meet customer requests and you see that an AIOps initiative will virtually pay for itself.