IT administrators can be a conservative bunch when it comes to adopting new technology. To some extent this makes perfect sense. When you have a large organization with numerous staff members, clients, and customers relying on your department to facilitate virtually every aspect of the business, it pays to approach major changes with care. However, this conservative mindset also means that many ITOps professionals mistrust or even fear significant new technology changes.
We recently spoke with Dennis Drogseth, VP of Research at EMA, about some of the practical steps that can be taken when deploying new technology, such as AIOps, to ensure a smooth transition for everyone involved.
The Challenge of Transformation
There is an array of concerns and challenges affecting how ITOps departments adopt new technology, beyond basic technical issues of system stability and compatibility. These include worries about staff training, skill levels, and adaptation; concerns about extending the new technology to less common use cases; heavily siloed job descriptions that don’t mesh well with modern departmental needs; and such a degree of sudden change often needed to bring about effective digital transformation that it can result, instead, in a lot of unnecessary stress and foot-dragging.
Drogseth recently hosted a webinar titled “AIOps and IT Analytics at the Crossroads: What’s Real Today and What’s Needed for Tomorrow,” where he discussed recent research into just what IT executives and professionals thought about AIOps and how they were integrating it and related automation technologies into their work. The study was based on a survey of 300 ITOps professionals regarding a range of AIOps topics, including IT analytics, big data, security-specific analytics, and digital transformation. Fully 87% of the respondents were actively using an AIOps platform, while the remaining 13% were either in the planning or testing stage. Total adoption seems imminent and inevitable.
Yet given the rumors of resistance in some quarters to the adoption and deployment of innovative tools like AIOps, we asked Drogseth for his thoughts on some of the best ways to help IT teams transition to using AIOps platforms smoothly and effectively.
Involved Leadership Is Key
Giving a brief overview of EMA’s research, Drogseth said that a big issue was that people “people want to cling to their tools and their data.” The results of EMA’s survey showed that, on average, current deployments integrated around twenty-three different sources, much of it made up of outputs from existing tools. Given that state of affairs, he said, it’s important to let your staff know that AIOps is going to help unify IT.
Another important point raised by the research data is that the most successful implementations of AIOps were driven from the top down, with executive buy-in at the VP or CIO level. Top-level executives are valued for their strategic oversight, something that will increasingly be required of all ITOps staff, as more and more low-level tasks are automated. Strategic oversight remains one of the key tasks for human minds in an increasingly AI-driven, automated IT environment.
“The kinds of people you want involved [in implementing AIOps] are strategic architects,” says Drogseth. “You want to have a strong executive pitch for sure—that’s really critical—and you want to show the values and benefits that translate across the stakeholders. Top-down leadership needs to be more than just sending money and an email and the hoping for the best. Leaders need to be educated about their need for leadership, and why these tools are worth the investment because they’re actually changing how their organization works. It’s an opportunity for new levels of efficiency.” He pauses, adding, “Trying to get everyone from the bottom up enthused about it can be like shovelling coal.”
Executive leaders need to be aware not just that change is happening, but the hows and whys of those changes. It’s not enough simply to announce a change and fund it; executives need to be strategically involved with the deployment, and be aware of the effects on how their organization operates.
Adopting AIOps requires making changes to how an organization operates at nearly every level
Lead from the Top Down, Build from the Bottom Up
When beginning an AIOps deployment, “Companies should prioritize based on use case,” notes Drogseth. “There are many things that Moogsoft can support, but start with core issues and performance and availability problems.”
Yet ITOps staff are often split between those who believe the much-fabled machine-learning “AI” won’t live up to the hype, and those who think it will take away their jobs. Many don’t think AIOps is different enough from their current automated performance monitoring (APM) tools to justify the challenge of learning how to use it.
In reality, if your old APM platform is still keeping up with your organization’s needs, then either you aren’t very far in your digital transformation journey, or you’re about to get a rude awakening when the “big” in Big Data becomes too much for your systems to handle.
“It is interesting how people are blurring APM into AIOps,” muses Drogseth. “Most APM tools are still fundamentally monitoring. They don’t give you the collective insights across all the domains; they don’t give you predictive guidance and prescriptive guidance; and they don’t really unify IT.”
Fear of losing one’s job, either due to lack of ability, or due to super-effective AI that vastly out-performs humans at the majority of day-to-day ITOps tasks, is a real concern for a lot of IT staff. Drogseth points out that the assimilation of existing staff members (and their favourite tools) will go a long way towards allaying initial fears about job loss. As in any other industry, many in IT feel that their very career is tied to their ability to use certain tools. The deep capacity for integration built in to Moogsoft’s platform “allows people to still work with their tools some,” he says, and he feels that this is key to enabling recalcitrant staff members to be part of any changes instigated from on high. “It’s human to say, ‘Look, this is how I work, I’ve got a lot of challenges, why should I have to change?’” he says. “So it takes creative and considerate leadership to get people involved.”
When it works well, fully assimilated staff will likely discover the value of AIOps and may adjust more easily to further changes ahead. Without good leadership this won’t happen, notes Drogseth, and staff run the risk of finding themselves being subjected to retraining sessions or replaced by workers who are willing and able to embrace sudden digital change.
The valuable role of an Automation Architect, for example, would ideally be filled by a senior systems administrator with a good track history in your organization, rather than by an externally hired software engineer with the right paper skills but no real idea of your current setup. Employees need to be able to see how their current skills will work within new contexts. If handled well, they might even become the strongest supporters of the digital overhaul initiated from the top.
Digital Transformation Means Actual Change
It’s important for IT staff to remember that things are going to change in this era of digital transformation, Drogseth insists. It’s simply the nature of the beast. “Digital transformation is changing IT,” he says, “and people who want to work the way they did 15 years ago are destined to lose their jobs, and should.” Ouch.
“What this does is it makes IT more exposed to the business, so it’s more vulnerable, but it also makes it more powerful in the sense that the services that are transforming the business are digital services,” he says. “IT will never get to become an effective power for digital transformation without moving in the direction that Moogsoft can provide—greater efficiency, greater cohesiveness, greater ability not just to deal with incidents but to resolve problems and help support change.”
And again, Drogseth believes that it’ll only come about most efficiently through strong, involved leadership. Nevertheless, he cautions that “one of the worst things to do is to insist on an overarching, mega transformation plan, mapping out processes that could take many months or even years to implement.” Instead, he says, “The important thing is that it’s vital to seek out tangible goals from day one that reflect the interdependencies of technology, process, organization, and culture, but which move in a laddered sequence forward so that benefits can be achieved relatively quickly.” In this way, IT stakeholders can see the benefits of the new technology sooner, making the big AI changes less threatening.
Adopting AIOps requires making changes to how an organization operates at nearly every level. Done well, however, AIOps will form the backbone of your company’s digital transformation for years to come, and is more than capable of unifying the many disparate departments and functions of IT.
About the author Dominic Wellington
Dominic Wellington is the Director of Strategic Architecture at Moogsoft. He has been involved in IT operations for a number of years, working in fields as diverse as SecOps, cloud computing, and data center automation.