No longer merely a big idea, AIOps takes on more tasks
AIOps is built on the idea that complex IT systems are managed best when humans handle fewer tasks. And, when the task is to identify glitches, AIOps offers much more information than simple problem recognition. An organization benefits from immediate notification that a system is about to fail or that a malicious intrusion is underway. But where the AIOps architecture will earn its money is in its next steps: Once an AIOps tool or platform recognizes an anomaly, can the AI evaluate possible responses? And when it selects a likely solution, can it act quickly to execute the fix?
To do all this without the assistance of an IT admin is impressive stuff, especially when the problem at hand is a complex one. Adding capacity is one thing. To determine the root cause of -- then a fix for -- a problem that the artificial intelligence has never encountered before might be too much to ask of even a sophisticated AI system.
IT consultant Clive Longbottom works through these scenarios in this handbook's first article. He explains how an AIOps tool applies simple intelligence to recognize and solve a problem it has been explicitly trained to address. When a complication is one that the tool isn't already outfitted to address, the AI must seek out potential answers and assess their applicability. Longbottom concludes that AI will gain that sophistication eventually, but the tools aren't yet ready to solve every problem by themselves.
This handbook further explores the capabilities and limits of the AIOps architecture, with a specific focus on how new and emerging tools deliver an automated remediation of certain ops problems while they also reduce the alert volume that sys admins and engineers must handle. Also, we've included a discussion of the challenges an organization is likely to face with AIOps implementation.
