How has the recent turmoil within the OpenAI offices changed your plans to use GPT in a business process or product in 2024?
Increased uncertainty means we are more likely to evaluate alternative AI chatbots and LLMs.
No change in plans, though we will keep an eye on the situation.
With Sam Altman back in charge, we are more likely to go all-in with GPT and LLMs.
What recent turmoil?

What Is AIOps — And Why You Should Care

The goal of AIops is to turn the data generated by IT systems platforms into meaningful insights. You may also notice some variations to this broad definition. This post why explain how it works and why it can be useful.
Aug 19th, 2019 12:00pm by
Featued image for: What Is AIOps — And Why You Should Care
Feature image via Pixabay.
The following is the first in a two-part series exploring the emergence of AIOps. Check back next Monday for the conclusion.

It’s hard to visit a tech website these days without reading an article about how artificial intelligence (AI) is poised to uproot entire industries and workflows. As it turns out, IT systems operations fall under the umbrella of things that are likely to change or already have due to AI. That shift created a new IT category called algorithmic IT operations — or AIOps for short.

The basic definition of AIOps is that it involves using artificial intelligence and machine learning to support all primary IT operations. The goal is to turn the data generated by IT systems platforms into meaningful insights. You may also notice some variations to this broad definition. That’s because the technology is rapidly evolving and is relatively new.

AI helps fill in the knowledge gap that frequently causes challenges for humans. This new way of doing things doesn’t take people out of the equation. Rather, it combines their intelligence with AI algorithms. Together, those two elements facilitate faster, more informed decision-making. Companies that use this technology can also detect and respond to incidents in real-time.

How has AIOps changed the typical workday for systems engineers and developers?

“AIOps brings an organization’s infrastructure monitoring data, application performance monitoring, and its IT systems management processes into shared context, enabling the application of big-data analytics to help systems engineers determine correlation and causality across all domains,” said John Gentry, chief technology officer for Virtual Instruments. Through the application of AI/ML-powered statistical analysis, heuristics and analytics, system engineers can proactively determine how to optimize application performance — and when there is an issue, quickly identify the real root cause. This also removes the need for the costly reactive firefighting that is the hallmark of traditional ‘IT war rooms.'”

AIOps could also bring benefits to the developer, Gentry pointed out. It could allow them to model and test the performance requirements of new applications and features prior to production, resulting in accelerated development cycles and improved time-to-market.

Such benefits across both the developers and operations could be especially advantageous as businesses scale up. And, AIOps could give information that helps enterprises stay up and running under pressure — such as during online traffic fluctuations.

How Is AIOps Similar to DevOps?

“Just like DevOps drove the cultural shift in the organization, AIOps is empowering a data-driven organization to uncover holistic insights from connected and disparate data to drive decision-automation,” says Bhanu Singh, senior vice president of engineering and DevOps at OpsRamp, which offers a AIOps platform for hybrid IT infrastructure monitoring and management.

Big data and machine learning are the two primary components of AIOps. But, it’s also important to note that there are three different IT disciplines within AIOps: automation, performance management and service management. The data associated with each of those areas then gets used by organizations that want to perpetually enhance their operations. Companies get continual insights, which drive constant progress.

The drive to keep making an organization better is something that AIOps shares with DevOps.

“AIOps is providing access to tools that can bring data, data analytics and machine intelligence together to make advanced decisions and perform automated actions by collecting and analyzing data,” Singh says. “It’s helping system engineers transition into Site Reliability Engineer (SRE) roles and support more scalable workflows that align with business needs.”

Similarly, teams that use the DevOps philosophy when developing software become comfortable with small-batch, fast and frequent process updates to improve their products. Statistics show that companies using DevOps deploy new software features and patches nearly 200 times more frequently than those that don’t.

Additionally, businesses relying on either DevOps or AIOps must break out of the traditional IT silos that exist at many of today’s organizations. That’s because using the big data component of the latter effectively means having information about IT operations from across a company — not just what’s going on in one department.

AIOps can make DevOps practices more effective, too. The main way it does this is by reducing the “noise” that can hinder productivity. For example, if DevOps engineers receive alerts from multiple platforms about single issues, they could waste time getting to the root cause of the problem. If AIOps streamlines all incoming notifications before DevOps teams see them, it becomes easier for them to start tackling issues.

So, keep in mind that AIOps is not a wholly separate entity from DevOps. Instead, it’s an emerging technology that can complement the goals of DevOps engineers and organizations at large.

Why Does AIOps Matter to Time-Pressed Teams?

No matter if you work on an IT team with five people or 50, you and your colleagues probably struggle with having too much to do and not enough time to accomplish everything. That’s a common issue in the IT sector and elsewhere. Fortunately, AIOps can help save time for DevOps engineers and others in a few key ways.

For starters, you can train a machine learning model to process all the data your company has. Plus, you can make that model flexible enough to readily accommodate any new information the company acquires later.

A well-trained machine learning algorithm can help keep your company’s data quality consistently high. Then, it should be easier for company leaders to trust it and act accordingly. A survey of CEOs from KPMG International revealed that 67% of those leaders ignored data analysis or insights from computer-driven models because those findings contradicted their intuition or experience. But, high-quality data helps build trust.

Also, because AIOps technology looks through such vast amounts of data so quickly, it can spot patterns that humans would likely miss without help. The conclusions drawn can then help avoid bottlenecks and other slowdowns before they happen.

Then, when IT professionals notice patterns associated with negative events, they can use automation to counteract them. By setting triggers in the system, they can make responses occur immediately after the adverse circumstances, making them less time-intensive to mitigate.

Why Do Businesses Implement AIOps?

Companies and their IT teams have much to gain by using this option to streamline their processes. But, of companies that are using AIOps now, why did most choose to do so? OpsRamp carried out a survey to determine the ways that companies most often utilize AIOps.

“It [AIOps] is helping to automate mundane tasks that do not require IT operators while providing contextual information for developers to improve MTTR [meantime to repair] and customer experience,” Singh says. “System engineers are able to proactively understand the behavior of the system and take preventative actions manually or through automation.”

One notable finding from the survey was that 87% of the businesses polled said their technology solution delivered the expected value. The top use case was intelligent alerting, mentioned by 67% of those participating. Root cause analysis/event correlation followed closely in the second-place spot with 61% of the respondents mentioning it.

The third, fourth and fifth business use cases all received similar percentages. Anomaly/threat detection was cited in 55% of cases, capacity optimization in 54% of incidents and incident auto-remediation coming up in 53% of the responses.

OpsRamp also found that businesses noticed various kinds of productivity boosters after starting to use AIOps. For example, 85% said they automated tedious tasks, while 77% said the number of open incident tickets went down after the implementation. When companies can more wisely use their IT resources, it makes sense that they could cut down the number of unresolved problems by targeting the issues and triaging them.

What’s Stimulating the Demand for AIOps?

“Most engineers dealing with infrastructure in large deployments are swamped with data, yet still lack reliable situational awareness,” points out Nigel Kersten, vice president of ecosystem engineering at Puppet. “There’s no question that we’re going to see more and more machine learning applied to solving the complexity and scale of this data, but it’s critical that organizations don’t just attempt to paper over the cracks in their unnecessarily chaotic environments and poor data sets with the promise of AIOps solutions. Resilient applications and infrastructure built with observability in mind and with a strong automation platform can solve many of these problems.”

A forecast from MarketsandMarkets assessed the market for AIOps platforms, and it expects impressive growth for the period from 2018-2023. More specifically, it predicts a 34% combined annual growth rate (CAGR). Ultimately, the market worth by 2023 could reach $11.2 billion — a substantial climb from the 2018 value of $2.55 billion.

You might be wondering, then, what’s causing the high demand? MarketsandMarkets mentions how we are in a highly digitized era, and it’s essential for companies to be flexible and responsive to business needs. AIOps helps them do that without wasting resources.

“Some of the common use cases where initial application are showing positive results,” says Singh, “include anomaly detection, noise correlation, causal analysis, prediction based on historical trend, closed-loop remediation, intelligent notification, and automation.”

Additionally, many IT teams are not well equipped to cope with the changing demands placed on them as businesses grow. In those cases, quality often suffers because engineers don’t have adequate time to assess issue notifications and get to the heart of what’s happening. Then, in the worst-case scenario, problems might go undetected for weeks or months until a disaster strikes and shuts down operations.

“The most significant advancements we’re seeing,” says Kersten, “are where organizations are using tools in this space to identify the most significant tedious and soul-crushing tasks their staff suffer under, and then both automating the tasks away and optimizing the processes around them, particularly where those processes cross organizational boundaries.”

Regardless of their industry or focus, companies with IT teams know that their ability to stay competitive depends on being in tune with marketplace needs, as well as internal matters. AIOps helps ensure businesses remain maximally informed by giving them the information they need to thrive.

How Can Your Company Get Started With AIOps?

“AIOps has reached ‘enlightenment’ in the Gartner hype cycle and is becoming mainstream,” says Singh, “but it is still early as different users have different perceptions of how AIOps could solve problems or how it will make their organization effective through automation.”

“AIOps is still a precursor to managing the complexity of today and tomorrow’s enterprise technology through automation,” he says. “Automation demands contextual and connected information to be effective, and AIOps is providing that.”

After reading this overview, you’re likely curious about how you could potentially implement AIOps into your organization. First, keep in mind that there is no universal road map to follow for success. That said, here are some general pointers you can adapt to align with your company’s present and future needs:

  • Get familiar with the basics of artificial intelligence and machine learning now. You may not plan to implement AIOps soon, but something could happen that makes it necessary to speed up your adoption.
  • Determine the most time-consuming tasks the IT team undertakes. Use those as a basis for understanding some of the problems that AIOps might solve. Pay particular attention to repetitive elements since automation could eliminate the human aspect needed for those.
  • Start small and branch out. Many companies get excited by the potential of AIOps and try to do too much at once. Instead, find your highest-priority problem and assess how this technology could solve it. Get feedback from a range of IT team members. They might have frustrations you’ve never noticed.
  • Let the system work with as many different types of data as possible. Using this tip might take more time than expected, especially if you need to get access to data from other departments in the company.
  • Consider how you’ll know whether AIOps investments are paying off. Which metrics will you measure, and how will you define what success means? Success looks different depending on your company’s challenges and business requirements.

Keep an Eye on How AIOps Develops

This overview will help you determine whether AIOps is a good fit for your company and how you might use it. Beyond that, it’s worthwhile to stay abreast of how AIOps progresses over the coming months and years. Various signs, including some mentioned here, indicate that the technology is poised for growth.

Group Created with Sketch.
THE NEW STACK UPDATE A newsletter digest of the week’s most important stories & analyses.