Top 10 predictions for AI in IT operations
By
Gartner first coined the term “AIOps” few years ago to describe “artificial intelligence for IT operations,” and over the last few years, IT operations monitoring tool vendors have begun incorporating AIOps features into their products.
Now AIOps tools are commonplace, but many IT leaders remain cautious about using these relatively new capabilities. That’s likely to change next year, however, as AIOps adoption goes mainstream; use cases will crystallize for improving IT efficiencies and supporting faster decision-making.
AI-enhanced automation will become smarter and more contextual, acquisition activity will explode, and you’ll see more movement of AIOps toward the edge. Here are the top 10 predictions to track.
1. The AIOps market will continue heating up
There’s been ample expansion in this market over the past year, with new entrants as well as several acquisitions of startups. M&A activity will probably continue into 2020 as larger incumbents seek to modernize their portfolios.
There’s still a lot of ground to cover as far as customer adoption and maturity of AIOps is concerned. Just one in five organizations has implemented some form of machine-learning software somewhere in their business, according to a study by 451 Research.
The research also shows that half of respondents have either deployed or plan to deploy machine-learning software from third parties, including cloud providers such as Amazon Web Services, versus building their own AI and machine-learning algorithms. Given the dearth of in-house AI skill sets and the complexity of developing AI applications, count on third-party vendor implementation strategies to grow.
2. AIOps will change the face of IT automation
As complexity grows in IT organizations, from multi-cloud and software-defined infrastructure to expanding digital business initiatives, so will the need for automation—but not just any automation.
The next evolution of automation will be smarter, more aware, and more contextual. AI and machine-learning technologies will discover hidden resources and threats, uncover patterns, filter the noise, and aid decision making.
AI tools will incorporate self-learning algorithms so IT operators can find answers to problems faster and receive recommendations on how to optimize IT performance as conditions change.
3. AI will increasingly support contextual data ingestion and correlation
AI tools can help reduce alert noise and address routine problems automatically. Expect enterprises to move further in their adoption of machine learning for complex tasks such as correlating datasets from multiple cloud providers, hybrid environments, and edge devices.
This will not only help IT operations get to the root cause of issues faster, but will also provide insights to help the business, such as understanding how to reduce customer churn through optimizing digital experiences across customer-facing technologies. Innovation starts with the business units, so it’s critical to understand how they and customers are generating and using data.
4. AIOps will be widely used on the edge
AIOps solutions typically run from the cloud. Yet this is getting more expensive and sluggish, as data volumes and use cases grow. In response, companies will begin to deploy AI tools on the edge of the network, where it’s faster and often cheaper.
This will enable near-real-time, AI-enhanced monitoring, eliminating the travel time from the data center to cloud service and back. That time savings will bring a noticeable difference in the case of a critical incident resolution.
Best of all, implementing AI on the edge won’t require new expertise to deploy. The deployment happens seamlessly behind the scenes through the cloud. Intelligent edge technology combined with the smart cloud will solidify the benefits of AI to IT operations teams.
5. Privacy considerations will grow
As AI on the edge grows, it’s more viable for companies to monitor desktops, tablets, and other end-user devices. While security teams have been doing that for years, IT operations has typically kept its work inside the data center.
AIOps will allow IT to guide employees on maximizing the usage of the applications installed on their devices while delivering greater visibility and control around the entire IT environment.
Yet there are real privacy implications. AIOps will ostensibly be able to see everything workers are doing on their devices. Considering how the lines have blurred between work and personal time, that means potentially having access to personal banking accounts or medical appointments, for instance.
IT leaders, in partnership with legal and HR departments, will need to strike the right balance between monitoring devices for business stability and protecting individual worker privacy.
6. Vendors will finally address security
In consumer applications, there’s been much controversy in the past year with privacy, security, and ethics related to AI-enhanced devices such as Amazon’s Alexa and Google Assistant.
In IT, AI is also a risk. The same algorithms used to monitor networks for suspicious activity could also be used against companies—to aid in an attack by creating fake accounts or bypassing anomaly detection systems, for instance.
For AI to succeed in a mainstream way, the industry will need to improve the security protections in applications and find solutions to detecting AI-induced attack methods before they wreak havoc on the business. Expect to see some significant movement in these areas in 2020.
7. AIOps will get aligned with business stakeholders
IT organizations have been working for years to get closer to business stakeholders, in an attempt to understand their needs even before they do. IT operations should be following the same line of thought, and AI will help it get there. In a recent survey conducted by OpsRamp, 64% of IT operations leaders said their job is to deliver agile, responsive, and resilient infrastructure that can support fast-moving business requirements.
IT Ops will move beyond alert correlation to adopting more business-friendly metrics and mapping IT metrics to specific business services. AI will play a role by forecasting the business service impact through analyzing infrastructure metrics and tying those back to key performance indicators.
8. AI will support DevOps practices
IT operations teams are looking at DevOps tools, skills, and methods to modernize how they work in tune with business and marketplace demands. In the OpsRamp survey, DevOps skills topped the list of needed capabilities, according to 64% of the respondents.
AI can help further DevOps practices by automatically optimizing code for performance. AI can discover patterns that indicate inefficient use of infrastructure resources and even make fixes automatically. This can provide a more stable and efficient environment for continuous development and continuous integration (CI/CD) cycles in DevOps.
9. AI will affect job roles in IT operations
Just as cloud computing created an entirely new set of development and IT skills, AI and machine learning will drive a similar change in how IT teams upskill. IT operations staff will have the opportunity to pursue data science and development skills so they can manage the automation of policies and actions in their AI tools.
This also means that jobs involving data entry and ticket management will shrink. But this is good news: The algorithms will do more of the grunt work while people will focus on more strategic jobs related to managing and analyzing the data.
Data scientists will play a large role in determining the best recommendations from AI systems and understanding when to override the suggested actions. By 2025 more than 90% of enterprises will have an automation architect to manage and monitor automation so it evolves without causing an entirely new set of challenges and risks, according to Gartner.
10. Government investment in AI will further innovation
Foreign governments, such as China, are investing heavily in artificial intelligence. Offshore cyber-criminal groups are likely doing the same. These pressures will incentivize government agencies to spend more on R&D in AI and machine learning—to support their own programs for criminal and terrorist surveillance, among other data initiatives.
These efforts will trickle down to industry, to help fill gaps in security monitoring and automation.
AIOps will continue to grow in importance
As with any emerging technology, there’s no way of knowing for certain how things will shake out with AIOps. But one thing’s for sure: The need for smart intelligence in IT and business will only grow. There’s too much data, too many tools, and too much unpredictable change for humans to handle without risking significant productivity loss, customer defections, and missed market opportunities.
In IT Ops in the coming year, AI will be one of, if not the, most important innovations for positive change.
AI-enhanced tools and processes will give IT a clear line of sight into infrastructure status and service health, the ability to proactively understand and prevent issues, and the ability to find the probable root cause(s) and solutions faster to support the business.