AIOps early adopters have been identified. Nine percent of the 846 respondents to a Turbonomic survey have rolled out artificial intelligence or machine learning (AI/ML) IT management into production. Another 28% are “experimenting” with AI/ML for IT management, but we don’t know if that means an internal expert is creating algorithms or if a vendor product that touts AI/ML under the hood is being piloted.
Does this accurately represent the extent of AIOps adoption?
Adoption figures for AI, ML, and intelligent automation are all over the map depending on how the terms are defined and who is asked about it. For example, 74% of IT managers surveyed by EasyVista say their organization implements machine learning, while only 50% of that same group implements artificial intelligence. These figures are higher than in many other surveys because the respondents were overwhelmingly users of IT Service Management (ITSM) tools, which have been at the forefront of using chatbots and other automation techniques.
AIOps gets slammed as a buzzword because it is an awkward attempt to append the term “artificial intelligence” to describe functionality that is being deployed in many IT monitoring and management tools. In the real world, that means that any product in the category that utilizes machine learning or predictive analytics for “intelligent automation” can describe itself as being in AIOps.
The number of vendors trying to provide AIOps is hard to keep track of. There are over 15 mentioned in these two TNS articles by Mary Branscombe: “Machine Learning for Operations” and “AIOps: Is DevOps Ready for an Infusion of Artificial Intelligence?” While some are pure-play start-ups, the majority of AIOps vendors appear to be just re-branding themselves. That being said, if they are using AI or ML to make your life easier, does it really matter how they market themselves?
Instead of defining AIOps narrowly, OpsRamp surveyed 200 people in IT operations, DevOps or site reliability engineering roles that had already implemented an AIOps solution in their organization. They found that AIOps tools are used most often for intelligent alerting and root cause analysis. That is believable, but I wonder about the 53% that use the tool for incident auto-remediation. Is this just reverting back to a previous environment? The way OpsRamp phrased its questions, it appears that AIOps automates tedious tasks, reduces the volume of alerts and reduces the number of incident tickets. Yet, there is an overwhelming worry about the accuracy of these tools, which could significantly inhibit future growth.
Feature image via Pixabay.