You are here

Will AI take your IT operations job?

public://pictures/ericka_c_0.jpg
Ericka Chickowski, Freelance writer

The tide of artificial intelligence (AI) in the workplace is rising, as is speculation over how it will affect the future of the workforce across the business world.

It will take a few years before any sound conclusions can be drawn about the relationship between AI and jobs. But, for now, most observers agree on this much: Some IT operations jobs will change, a few will go away, and some new jobs—or new responsibilities to existing roles—will be added.

Inexperienced help desk and incident response professionals should start building their skills now, since most experts believe these are the jobs likely to be automated away. But a bright future remains for human-led operations. People are most likely to take on new, data science-related roles such as data curator, augmented intelligence analyst, interaction designer, AIOps architect, and automation path designer.

Here's how AI will affect your future in IT Ops.

[ Learn how to transform your IT with AIOps in TechBeacon's guide. Plus: Download the analyst paper on how AI is changing the role of IT. ]

Today's IT Ops role

IT is in an interesting position because it is not responsible solely for pushing out the AI-based innovations that will affect everyone from factory workers to clerical admins. IT is also in charge of applying the power of AI internally, to improve the efficiencies of how its own workforce delivers technology and services.

This is particularly true for IT operations, which runs IT systems that are growing more complex, and in turn creates increasingly more subtle faults within them. This is creating a huge data volume problem that can be handled only by machines, said Will Cappelli, CTO of Moogsoft and a former Gartner analyst who coined the term "AIOps" as shorthand for AI operations.

"Human IT operations teams will require the automated enhancement of all cognitive processes."
Will Cappelli

This includes winnowing "noisy" data generated by IT systems, identifying which data is most relevant to performance problems, discovering correlational patterns pertaining to performance, and even coming up with action plans for remediation, all of which he believes AIOps automation will dominate.

As AI increasingly takes over manual tasks traditionally carried out by IT Ops, it's only natural for those who work in this function to ask if they're going to lose their jobs. The simple truth is, nobody really knows.

Experts agree that AI will shape big changes in the landscape of the IT Ops workplace. Though experts disagree on the level of job displacement—some say to expect many kinds of jobs to go away, while others say it's simply a matter of rejiggered responsibilities—they all recommend that ops veterans and newbies alike be prepared to update their professional skills in advance of the evolution of AIOps.

Differing opinions

Some industry insiders are more bullish than others about the impact that AI will have on IT Ops jobs. For example, the ops world should prepare itself for whole jobs to be automated away in just about "any and every function possible" as IT races to automate away errors, increase predictability, and stabilize costs, said Brendan Caulfield. He's the co-founder of ServerCentral Turing Group, which helps move organizations to AWS clouds. 

"The order in which these functions are automated—and the level to which they are automated—will be unique to each organization."
Brendan Caulfield

He further explained that "value will be recognized or realized in different ways," based on competitive position, price pressures, and customer and market expectations. 

There's no reason to panic, since the technology has a long way to go before it can come close to the autonomous thinking power of people, said Rajesh Kalidindi, CEO of LevaData, a supply chain AI company

"Right now, AI doesn't come close to the cognitive capability of an infant. The low-hanging fruit for AI and automation will be the repetitive, time-consuming tasks required in IT roles—not entire jobs."
Rajesh Kalidindi

In the next five years, he said, the biggest benefit AI will likely provide is giving people more time to focus on creative versus routine tasks.

[ Enterprise Service Management brings innovation to the enterprise. Learn more in TechBeacon's new ESM guide. Plus: Get the 2019 Forrester Wave for ESM. ]

Humans and machines, together

Kalidindi and others believe that the more IT applies AI to operations problems, the more important it will be for humans to pull all the puppet strings. At the end of the day, the machines need domain experts to tell them how to operate in certain situations.

Software developers may be programming the machines, but operators will be the ones distilling organizational and systems knowledge into "run books" that AI will apply to various IT Ops problem sets, said Patrick Zimmer, principal at WholeStack, an IT services firm that manages and operates systems.

"We cannot deflect the responsibility of fixing our problems onto a machine. It creates unintended consequences. Machines have a hard time with abstraction." 
Patrick Zimmer

As such, don't expect AIOps to replace jobs wholesale, said Jiayi Hoffman, data science architect for OpsRamp, which sells an AIOps-based platform for data center and cloud operations management. Instead, the discipline will change responsibilities, she said. The type of human-led work will shift from handling incidents that interrupt the business to forecasting maintenance, and handling analytics and other data.

"[AIOps] will increase IT's focus on preparing environments for machine learning: Prepping training data and sending user feedback to continuously tune the machine."
Jiayi Hoffman

Which IT Ops jobs will be most affected?

So whose responsibilities are most likely to morph as AI enhances the ops function? One good rule of thumb is to assume that any role that responds to tickets will feel an impact, said Robert Reeves, co-founder and CTO of Datical, which sells tools to simulate database environments. 

"The ticket is an anti-pattern for automation and is a huge red flag for people focused on streamlining processes and offering self-service to users."
Robert Reeves

Many organizations are already aggressively applying AI to reduce ticket counts. AIOps vendor BigPanda is seeing certain organization reduce tickets by 75% to 99%, according to statistics it has gathered. As ticket-based operations models decline, expect front-line support roles across the board to see the most changes, ServerCentral Turing Group's Caulfield predicted.

Caulfield is specifically talking about the first point of engagement with a customer or an end user, he explained. 

"Intelligent automation of inquiry management, triage, and first-step escalation will continue its rapid adoption."
—Brendan Caulfield

Some organizations may be able to build intelligence around help desk levels 2, 3, and beyond, depending on how well defined the business rules and processes are. "As many steps in the business process that can be automated will be automated. It's purely a function of the breaking point for each organization—when the intelligence isn't strong enough for automatic resolution," Caulfield said.

Three of the biggest functional areas to see AI-related job effects will be availability and performance management, incident management, and problem management, said Moogsoft's Cappelli.

"In five years, users will handle many more of their own incidents directly, and the ones they cannot handle will be routed by robots to the right people or robots for handling."
—Will Cappelli

Availability and performance management will have to shift from observation to analysis, incident management will largely disappear, and problem management will have to shift from actually doing diagnostics to validating results, Cappelli said.

The ops workforce of the future

Even as first-level help desk roles start getting slashed, the ops team is still going to need hands-on operators to do the dirty, manual work for years to come during the AI transition.

"We continue to see a pretty vital role" for help desk staffers at levels 2 and 3, with domain and/or organizational expertise for the next five years or more, said Mohan Kompella of BigPanda.

As AI progresses, expect more of the ops jobs to shift from supporting the systems to maintaining the AI that maintains the systems. The way Datical's Reeves explained it is to point to movie director Tim Burton's version of Charlie and the Chocolate Factory. After Charlie's father was automated out of his job at the toothpaste factory, he was eventually hired back to fix the machine that replaced him.

AI and machine learning (ML) do not work unless someone can direct the algorithm and train it, Reeves said.

"We can’t expect AI/ML to 'just figure it out' and perform self-directed learning. We are not anywhere close to that."
—Robert Reeves

New IT Ops roles coming

This is going to lead to the rise of the data specialist instead of the systems administrator or help desk staff, said OpsRamp's Hoffman, who is among those who expect to see whole new classes of jobs emerge in the ops field. New titles that will emerge fairly soon include data curator, augmented intelligence analyst, interaction designer, and AIOps architect, they said.

The IT Ops role will evolve to defining and managing standards, rather than doing the actual build and deployment work that is done today, said Andrew Herbert, founder of Cangler Analytics, which sells AI-related data analytics services. He explained that IT Ops will be where workflows are architected, using infrastructure-as-code and automation tools. 

While this is already happening to some degree at many DevOps organizations, it will scale way beyond just the developer pipeline and move into systems such as databases that are still mostly manually maintained.

"The future will be service-oriented. The expectation will be for IT operations pros to provide their services as a consumable service via an API or some other technical function available for customers from their workstation."
Andrew Herbert

How to weather the storm

To not only weather the AIOps transition but become indispensable to it, IT Ops professionals should start thinking now about how they can add value to the automation of everything. Most importantly, don't fight progress: Automation is coming whether you like it or not.

If you're worried about something like AI or automation taking over your job, perhaps the best approach is to get in front of that by becoming an expert on those particular technologies, said Carolyn April, senior director for industry analysis at industry trade group CompTIA.

"When you become the person in the organization who helps put those technologies to use, it doesn't mean that you lose your job. It means that you're more valuable.”
Carolyn April

Observers such as LevaData's Kalindindi and OpsRamp's Hoffman believe that, to start working themselves into this position, people in IT Ops need to start learning data science techniques and understanding the basics of how machine learning works. They'll need to start applying it in their new roles, and perhaps in their current roles also.

Learning opportunities abound, via traditional educational channels (both online and in-class), self-directed learning, and online massive open online courses (MOOCs). 

"Machine learning literacy will become critical in these new roles. Candidates will need an understanding of data algorithms and data manipulation as well as quality control and engineering to orchestrate the machine learning outcomes."
—Jiayi Hoffman

Be proactive, and don't bury your head

The overall number of IT operations jobs isn't expected to take a big hit anytime soon, but the field is definitely primed for a massive shift in responsibilities. Whether you're a veteran or newbie, smart IT personnel would do well to start learning more about AI-supported IT operations. This includes data science and how machine learning works. 

[ Learn how robotic process automation (RPA) can pay off—if you first tackle underlying problems. See TechBeacon's guide. Plus: Get the white paper on enterprise requirements. ]