AIOps Evolution Podcast | Season 2
Interview with Christian Malone from ServiceNow
There is no doubt that AIOps trends are on the rise. But one of the biggest complaints from Sean and Christian was that AIOps has become a ‘buzzword’. It’s important to keep the integrity of AIOps by not letting it become a generic term. That means understanding its specific use cases, and how it can become a business game changer by freeing up IT teams for higher value pursuits.
TL; DR
AIOps Trends
AIOps = buzz wordy
AI for IT Operations, known as ‘AIOps’, has officially become a buzzword. Vendors and thought leaders in the industry crow about its potential for technological advancement, but without a clear definition, it holds as much weight as “digital transformation.” Christian pointed out that 58% of IT professionals think AIOps is a buzzword. Before we can unlock the power of AIOps, we’ve got to unlock its meaning.
“Shiny object syndrome”, “HD overkill” – whatever you want to call it – AIOps has fallen into that category. Just like getting all your personal tech in HD can be overkill, jumping onto the AIOps bandwagon can be overkill without properly identifying where to leverage it. Christian notes we’ve got to identify specific AIOps use cases that are worth our time and resources. Furthermore, we’ve got to move past AIOps as a generic term.
Using AIOps to free people for higher pursuits
While there’s a list of 90+ use cases for AIOps, primary uses include: data collection, data correlation, and anomaly detection. The AIOps industry continues to expand into natural language processing, neural nets, and more. We as humans cannot keep up with it all, but an AI can. It has the potential to streamline all of those tedious functions that keep humans in a reactive state to technology. That’s where Christian sees the greatest potential in AIOps trends – as a way to free people for higher pursuits.
He says, “I’m a big Calvin Hobbs fan and they had a great comic panel at one point that just showed Calvin raising his hand in school saying, ‘You know, with the pace of technology, maybe we should just leave the math to the machines and go outside and play.’ And, I’m one of those guys that wants to help people to go outside and play, and then I want to go out and play with them.”
Calvin’s thoughts resonate with people, especially if it means freeing up IT Operations to support greater business initiatives. That’s the promise of platforms like ServiceNow, for instance. The end goal of AIOps platforms is to solve people’s problems; their data problems and to integrate data to have more insights. It sounds simple to implement, but it’s easier said than done. This especially rings true since the market for data scientists is steep at the moment.
Where is the market for AIOps now?
According to Christian, the market for AIOps tends to encourage building a data lake and then go looking for problems. There are a large number of vendors out there who tell companies they can discover so many “insights.” Yet, this is not a meaningful use case that benefits the IT team. It’s more of a solution looking for a problem.
Also, there’s an AIOps trend towards the cloud right now. People are taking their data to the cloud, using and deploying apps there, as well as the same frameworks. Underlying logs we’re getting from infrastructure, networks, kubernetes, security tools, apps, databases are all datasets that are an opportunity for customers to utilize AIOps as opposed to hiring data scientists. AIOps is a simplified solution to all of these common problems.
Put valuable data scientists to better use
The top benefit from AIOps is freeing up data scientists to pursue more critical situations. Companies are able to put data scientists to better use creating ways to better society and lifestyle (i.e. solving world hunger or developing self-driving vehicles). Data scientists are coming out of higher education wanting to solve higher-level problems than figuring out why the database or the Q depth is higher than it should be. So, AIOps is beneficial to aiding this effort to put human minds and skills to better use.
Can companies build an AIOps team whether they want to or not?
One of the biggest problems in the industry is not whether companies want to implement AIOps. Large tech giants like Google and Facebook will have first pick of “the crop” and many enterprises and mid-sized companies can’t compete with it. In these instances, Sean and Christian predict vendors will rise as an option for mid to small size businesses to leverage AIOps.
Instead of hiring for an internal AIOps team, third party AIOps vendors provide expertise and strategic guidance on deploying AIOps within an organization. Not only does this reduce costs, but getting an objective outside opinion on your IT environment and expansion plans for AIOps can focus your core business functions and accelerate service delivery. Partnering with professionals in the industry helps companies to know what to prioritize when on the journey to AIOps.
What’s next in AIOps trends?
That’s why it’s important for vendors to offer a solid AIOps program moving forward. Christian predicts that we’re going to see where the data that resides, those sources of truth and toolsets that IT teams rely on managing that data will have to adopt AIOps. Christian says, “And so, rather than building everything into one repository, I think we’re more likely to see that customers are gonna kick out vendors who should have AIOps or ML or data insights, and are not adopting it.”
Rather than a one size fits all, we’ll find a distributed model where localized domain centric makes sense. There still might be that data lake for proprietary data; or where it makes sense to add in data from a lot of different areas and get cross correlation.
