Shawn Bice

President of Products and Technology at Splunk

The massive amounts of data that enterprises produce daily can help solve some of the most difficult business challenges. For AIOps to live up to the hype, organizations must eliminate data silos, adopting a holistic data and machine-learning strategy to create an environment that is continuously observing, learning and improving itself.

While observability creates mountains of granular data, dependency graphs and application topologies, AIOps trains on this data — resulting in identified problems turning into long-term solutions, past behavior informing improved workflows and faults and failures fueling training algorithms. Organizations armed with both observability and AIOps are enabled to take swift action on their data and can easily automate responses, which is critical for teams aiming to exceed expectations with fewer resources and shorter timelines.

I believe we will continue to see these once separate entities blend as businesses progress through this year, and it will become increasingly clear that AIOps is more than a square in buzzword bingo. There must be an inextricable link between AIOps and observability to unlock its true power.

Tweet this.

Gregg Ostrowski

Executive CTO at AppDynamics, part of Cisco

As applications have become critical to businesses, the complexity to power them has grown. Moving to a strategy with AIOps has been a growing phenomenon over the last years, and has benefits, as automation and artificial intelligence grow more important. AIOps needs to exist as a component of a wider strategy focused on full-stack observability with the appropriate business context to accurately measure the true effects of activities across the full stack and to ensure IT teams can pinpoint root causes of potential issues and address them in a targeted and swift manner. This process leads to the ultimate goal of a self-healing environment to automate resolution before it impacts the end user.

An organization must implement cultural change that removes finger pointing and encourages collaboration among teams that are closest to IT efforts. This includes IT, security teams and business leaders who should be involved so security and performance stay central to the technology and strategy decisions. By providing business context, teams can align and see how their efforts affect the business, which helps streamline resolution based on what's most valuable to the business.

For AIOps to truly fulfill its potential in helping organizations scale faster and provide context to the loads of data required to keep modern applications performing optimally, it needs to be wrapped in with DevOps, DevSecOps and BizDevSecOps. Rather than adding to the growing related vocabulary, wrapping AIOps into a full-stack observability solution with business context provides adequate information for continued digital transformation.

Tweet this.

Dave Wright

Chief Innovation Officer at ServiceNow

When used properly, AIOps enables IT teams to act more efficiently and respond to issues proactively and in real time, leading to better experiences for the IT teams, customers and employees.

For companies that are just starting out on their AIOps journey, it's often smart to start with a focused approach on a single use case before applying AIOps across the entire organization. This allows IT leaders to demonstrate the value of the implementation, showcase the power of AIOps and establish the data-driven mindset in the team that is needed to successfully implement AIOps deployments. Then, once they've demonstrated the ROI and prepared their teams for this new approach, they can scale up across the enterprise with the proper goals and objectives in place.

Tweet this.

Ram Chakravarti

Chief Technology Officer at BMC

Globally, it's estimated we'll generate a staggering 79 zettabytes of data in 2021 and more than double that at 181 zettabytes in 2025. Complex enterprise IT environments aren't immune to this data growth, which is making it difficult to scale and address the challenge. When something slows or breaks, it can take time to figure out the problem. AIOps technologies can help today — it's not hype.

An enterprise-wide AIOps strategy that includes machine learning and advanced analytics provides the monitoring, planning and automation required to cut through the noise, find root causes faster, reduce mean time to repair (MTTR) or preemptively fix issues before they become problems, and keep service-level agreements (SLAs) stringent. This translates into business savings and productivity, which in turn enables IT to focus on higher-level pursuits that support the evolution into an autonomous digital enterprise.

For AIOps to be a reality, companies must carefully take technologies, operations and people into consideration. The latter is probably the most important: It's imperative that an organization equip and empower their people to act on the insights provided by the tools and processes. It's critical that AIOps tooling has an open approach to integrate with existing tools and data sources, given the broad range of data to observe and analyze. Done right, it enables teams to support agility and collaboration across developers, operations and security, where human creativity can be capitalized on as machines handle the massive amounts of data.

Learn more.

Tweet this.

Lori MacVittie

Principal Technical Evangelist at F5

If we view AIOps as the ability to automatically adjust configuration and policies to optimize and secure applications, then we need to consider three distinct requirements.

• First: operational insights into current conditions and the ability to recognize the need for adjustment. That is ultimately driven by operational data (telemetry), so organizations need to be able to collect and analyze a broad spectrum of data from across infrastructure, systems, applications and services.
• Second: automatic adjustment implies the ability harness the power of automation. That requires being able to automate the relevant infrastructure, services and systems that deliver and secure applications.
• The third requirement is a way to connect the two. Just generating insights isn't enough, and neither is the ability to automate adjustments. The two need to be connected, with the latter driven by the former. That implies the need for an overarching system capable of taking insights and turning them into actions that result in positive outcomes. Otherwise, we must continue to rely on human intervention, which we know ultimately does not scale.

Now, even if we had all three components, there remains one requirement that cannot be developed or downloaded: trust. It took years for organizations to trust systems enough to rely on auto-scaling capabilities, and what AIOps promises to deliver has greater responsibilities — and therefore risks.

AIOps will — and does — exist in various nascent forms today. But a fully automated system capable of operating a digital business at scale will take time, technology and most of all, trust.

Tweet this.

Brian Amaro

VP of Global Solutions at ScienceLogic

The hype around AIOps isn't unwarranted; there's tremendous power in its promise. But like any modernization effort, it's a journey. Before charging ahead, executive leadership must ask: Have we established foundational elements around people, processes, and technology to pinpoint desired outcomes? Don't be misled — AIOps is no "easy button." The reality is, it requires crucial processes and specialized tools working in concert.

Automation is only one step. Building purposeful plans to coordinate tools in pursuit of AIOps' service-centric value is another, and it must come first.

What business outcomes do you want to achieve? What's your budget, timeline and current situational awareness of your data? How complex is your architected environment? What challenges do you need to break down to get to those desired outcomes — and what are the sources of those challenges?

The data element is critical. If data sets are not aligned, there's no longer a solid foundation for standardizing, baselining, integrating and evolving at speed. But humans can't align that data on their own — and this is where the AIOps methodology can fall apart. The data-gathering aspect must be automated before enabling other tools in that environment, like anomaly detection. If humans participate in the alignment process, inaccuracies get introduced, leading to more extensive projects down the road because the system's foundation is built on inaccurate data.

Once you have foundational data integrity, mature strategies and comprehensive situational awareness — once there's organizational alignment — that's when AIOps can exponentially transform resilient, digital enterprises and empower intelligent futures.

Tweet this.

Spiros Liolis

CTO Cybersecurity - Global Alliances at Micro Focus

IT organizations are increasingly pressured to optimize operational efficiency while improving service quality and dealing with more complexity and rate of change within IT environments. AIOps leverages big data and machine learning functionality in an open framework to support primary IT operations, allowing data ingestion from disparate sources and advanced analytics use cases to improve IT operations efficiency.

Which AIOps use cases are defined will determine when and how to proceed. Once this is well established with the key stakeholders, the organization must undergo several layers of transformation.

• First, they must create an architectural vision for a data lake environment, including the connectivity layer and sources with single central analytics and visualization dashboard. They can also benefit from streaming in real time and batching data ingestion for adaptable data collection and analytics.
• Second, organizations must look to transform their operational processes. This can be achieved through digitizing processes with automation and staff enablement, and empowering data-driven decisions with deeper insights into customers, products and services.
• Lastly, organizations must transform the customer experience. This starts with leveraging analytics to gain a better understanding of the customer, including behavior and segmentation. Digitized touchpoints can also improve experiences, e.g., social media for answering complaints as well as self-service capabilities.

Tweet this.

Jen Grant

CEO at Appify

AIOps is another example of a technology advancement that reduces manual work for employees - in this case, the IT and DevOps team. As with any new technology, the industry can get starry-eyed about what this will mean when it's fully scaled and in-use, but, as always, this perfect vision of the new AI world will be slower to arrive than we would like. Businesses need to prepare their employees for this reality. There are two key challenges to address up front.

First, implementing and training any AI advancements takes time, and will not produce great results as fast as one might like. Second, it's important to reinforce the continuing need for actual humans. I've seen many developers have a similar reaction to no-code advancements, worried that these new tools will somehow take away their jobs and their importance. This has never been the case as technology advancements are introduced. With every step forward in abstracting technology layers, the advancements have made engineers more valuable and necessary.

Tweet this.