The key AIOps trends of 2020

• March 6, 2020

Large volumes of alerts, significant IT noise and signals distributed across disparate tools are holding DevOps professionals back. Meanwhile the pressure for teams to improve performance across IT infrastructure, as well as more accurately problem solve and resolve incidents faster, is growing.

Fortunately, awareness of how AI can address these challenges and help run IT operations is growing. Recent research by New Relic and Vanson Bourne found that 89 percent of 750 global senior IT decision makers surveyed believe AI and machine learning is important for how organisations run IT operations. Most (84 percent) also said AI and machine learning will ultimately make their role easier. 

This appreciation of what could be possible with AI looks set to transform into widespread usage of AIOps over the next few years. Gartner has predicted large enterprises’ use of tools like AIOps will grow from 5% in 2018 to 30% in 2023. But what will be the key AIOps trends that the industry will remember from 2020?

Leadership takes AIOps seriously

Regardless of the hype, 2020 does look set to be the year IT leaders will take AI seriously and vouch for its use in IT operations. There are more examples of organisations trialing AIOps, which will build a solid business case for it, and secure buy-in from the C-suite. 

Investment into AIOps can bring benefits from reduced downtime to a boost in profitability, creating a win-win for customers and companies alike. The more senior leadership backing there is for AIOps, and the more mature the technology gets, the easier it will be for Service Desk, DevOps and cybersecurity teams to do their jobs. 

This is because AIOps will be able to process a greater number of data types to IT teams alone, and automation and simplification of straightforward processes will free up these experts’ time to focus on more complex tasks that can’t be achieved by AI itself.

What’s key in 2020 we will be the emergence of ‘data agnostic’ or ‘open’ tools that can take in data from a wider variety of sources. IT teams that capitalize on these tools will reap great rewards, as it will give them the chance to undertake more re-engineering of the core data platform and allow the tool to draw from a greater range of data inputs and types, as well as provide comprehensive analytics. 

The richer the data the tool can go on, the greater number and quality of insights the algorithms can arrive at. The deeper the visibility afforded to IT operations, the better the understanding of problems there can be, and the easier it is to fix them. Teams can more accurately predict issues that may occur by taking certain actions and avoid implementing such changes in mission-critical parts of the business.

Automatic remediation will become a key feature of many AIOps tools in 2020, marking a clear step change for IT professionals as they work to close the gap between detecting issues, diagnosing problems and remediating incidents. This is due to most current AIOps technologies requiring human beings to use the tools for detecting, diagnosing and prioritizing, as well as carrying out the often exhausting leg work of spotting anomalies, correlating series of similar events and more.

Incident management gets augmented

A variety of incident management functions will be augmented with AIOps this year. These may include natural language processing, anomaly detection, event correlation and analysis, root cause analysis and other such IT functions to enable IT operations professionals more control. 

Alert correlations and incident intelligence will be freely accessible within the incident management platform. In addition, IT teams will be able to pre-empt the detection of anything out of the ordinary and highlight it on the systems they already use to the best effect.

Observability levels up

We are already seeing vendor point AIOps solutions and ancillary technology providers converging, however there remains a lot of fragmentation in the market. There will be an increasing expectation for AIOps to become embedded in observability platforms this year. This will allow IT management staff to leverage all the operational data and telemetry they need for machine learning and achieving faster incident resolution.

Many AIOps tools on the market today are simply not giving enterprises the right time to value they need. DevOps teams are already time-poor but are still being held back by long set-up, data input and employee education and onboarding times. However, this shouldn’t (and won’t) be the case for very much longer, as we expect 

AIOps capabilities to leverage a flexible baseline of rules that change based on each user’s input and production data, giving DevOps staff the ability to insert their own logic into the system and offering feedback on the proposed logic. There is no doubt that this year AIOps tools will become easier to realize, be trained on and capitalized upon, and that the ones that do not will become obsolete and redundant.

Source

Large volumes of alerts, significant IT noise and signals distributed across disparate tools are holding DevOps professionals back. Meanwhile the pressure for teams to improve performance across IT infrastructure, as well as more accurately problem solve and resolve incidents faster, is growing. Fortunately, awareness of how AI can address these…