Last week, I tuned into a EMA Webinar, “AI(work)Ops 2021: The state of AIOps,” led by Valerie O’Connell, EMA Research Director. Valerie shared the results of an AIOps survey given to more than 30 customers of four AIOps vendors that paid EMA to conduct the surveys, view the use cases, and examine at least two customer implementations in great depth.
Summary Valerie opened by cutting right to the chase: “AIOps is in full swing across enterprises of all sizes, with more than 90% of organizations in active deployment. Although the discipline is still relatively new to IT, there are big wins to be had—both quantifiable and qualitative. In fact, AIOps has a very high success rate (95%) and almost universally pays for itself.”
Definitions. AIOps addresses day-to-day IT operations by gathering data data from within an organization’s infrastructure (servers, desktops, network, and applications), applies machine learning to it and acts upon the results through automated processes. Robotic Process Automation (RPA) focuses on business applications.
Other Findings include:
- 100% of those surveyed were successful with AIOps. 77% reported being very/extremely successful.
- An AIOps implementation at every level will at least break even.
- More than half of the respondents directly associated automation with AIOps
- The long-term goal for organizations is not to become a totally “Autonomous Digital Enterprise” but to retain a human touch. (People tend to trust human oversight.)
- Automation breeds more automation and increases with implementation experience.
- IoT and operational technology are the top use cases. DevOps and development are other leading use cases.
- AI Monitoring was not widely used by organizations in this survey.
- Those organizations with extremely successful AIOps implementations reported that digital transformation drives their AIOps strategy.
- Other top drivers include operational efficiencies, DevSecOps, and executive pressure on IT.
- Success drivers include: reduced number of outages, improved process efficiencies, and decrease in war room frequency.
- Only 18% of organizations found AIOps implementation easy or smooth.
The bottom line? “AIOps is real, early, and the near and the clear future of IT Operations.” Both BMC and Atlassian now tout full AIOps capabilities, including AI Service Management (AISM). And at BMC’s April sales kickoff we learned that BMC Helix Monitor is the only AIOps platform that does full AIOps.