• By 2027, AI will be providing more IT technical support, at least in written form, than humans.
• The CEO of cybersecurity company Palo Alto Networks, Nikesh Arora, said earlier this year that it has reduced IT support staff by nearly 50%, with an expectation that it will be reduced by 80%. 
• Technology career experts say the influence of AI on the IT desk will be similar to its impact on other careers, requiring evolution but not leading to extinction, and cost concerns may limit its replacement of workers.

Customer-facing chatbots have already become a presence in some of the most common interactions within our daily lives, from Walmart using AI to help associates on the floor to Tractor Supply sales reps' use of wearable AI tech. The IT support desk within companies that a generation of workers have come to depend on, for when laptops melt down or remote connections are lost, may be next.

It's estimated that by 2027, generative AI will create more IT support and knowledge-based articles than humans, according to Chris Matchett, senior director analyst at consulting firm Gartner. That is part of a broader business consumer self-service strategy, within what's referred to as "support case deflection," he said, and being primarily driven by mundane methods such as automated password systems, a good knowledge base and an intuitive request portal. "AI can help with these," Matchett said.

At some major corporations, the AI help with tech tickets is advancing rapidly.

Nikesh Arora, CEO of Palo Alto Networks, said during an earnings call earlier this year that the cybersecurity company had reduced its IT support staff by nearly 50%, with an expectation that it will ultimately be reduced by 80%, with AI doing most of the basic IT support for employees. Arora told Wall Street analysts after the company's August earnings call that the company's roughly 300-person team dedicated to solving employee tickets, or about 2% of its 15,000 workforce, had been reduced by 50% and replaced by an AI problem-solving technology.

"We think that can go down 80% because we've been able to automate a lot of the tasks and also use generative AI to answer employees' questions. So, you can think about 200 people, which we don't need, either our own employees or third-party contractors, because it's boring work," Arora said.

Although not a large portion of the workforce, IT workers are in a role that other employees not only interact with but often rely on in their work lives — resolving annoying problems that derail employees' ability to get their work done fast, or at all in the case of the need to address a more important systemic issue. It is also an important entry-level position for tech careers. 

The success of these AI projects remains tied to human expertise, and while the number of IT support workers will be lower in the future than the total now, the role is more likely to evolve than disappear. Matchett said that by 2027, despite Gartner's forecast that AI-produced IT support documentation will eclipse human production, 50% of AI projects at IT service desks will be abandoned due to unforeseen costs, risks or an inability to achieve the projected return on investment.

When employees reach out to the tech desk for help, at least to date, an overwhelming proportion prefer human interaction. A recent Gartner survey asked over 5,000 digital workers, "If you had an issue with the digital technology you use for work, how would you prefer to solve it?" Contacting IT support via chatbot or AI conversational agent ranked No. 6, with fewer than 10% of respondents choosing this as their most preferred option. Support channels involving human interaction (such as phone, live chat and email) remain much more popular. "Besides the technology involved, success hinges on the humans that will always be involved," Matchett said. 

Taryn Dawson, senior manager partner and ecosystems communications at Palo Alto Networks, said the IT support staff example involving 275 people involved in solving employee tickets is part of an internal AI employee experience that can reduce specific head count and the company has "more efficiencies planned in the future."

"More and more, we're leveraging gen AI techniques both in our products and operations to make it easier for our customers and employees to self-service their needs," Dawson said. Documentation is a good example, with Dawson noting it has been greatly accelerated by AI, with the company's Prisma Cloud Copilot delivering an experience 24 times faster for documentation searches. 

But at Palo Alto Networks, as with most AI adoption, utilization currently includes human plus automated elements for routine work, allowing IT teams to focus on higher-order tasks. "The overarching goal is clear: faster responses, pinpoint accuracy in threat detection, and overall, a more resilient security posture," Dawson said.

Seth Robinson, vice president of industry research at CompTIA, an organization that focuses on technology professionals and careers, said automation of IT support tasks and self-service for simple requests remain industry priorities, alongside analysis of support operations within an organization. As AI has grown, the benefits have been clear, he said, but so have the hurdles to widespread adoption, including in IT support. 

For one thing, AI delivers solutions based on probability – by definition, there is always a risk of a wrong answer. These risks are dropping with improved technology, Robinson said, but they will always exist.  Humans make mistakes too, but this process is built into the back-and-forth organizations are used to accepting when two people are attempting to solve a problem, one with greater expertise than the other. Loss of this human interaction is impossible to ignore, he said. 

"In terms of acceptance, there is a fine line between delivering improved results for the customer and causing frustration if the customer can't get an answer," Robinson said. 

Frustration with AI tends to be directed at the system as a whole, with precipitous loss of confidence and acceptance, whereas human error is understood as a step in a worthwhile process that we all more or less have faith in. Everyone has interacted with an automated call center, and if the technology can't provide a solution then the frustration spreads from a single unsolved issue to "a sense of screaming into the void," Robinson said. "Sometimes people want to talk to another human for comfort or because the problem is complex," he added.

AI is certainly causing the IT support job role to evolve, but as workflows are also evolving the demand remains in place, Robinson said, pointing to recent job market data that bears this out. CompTIA's monthly Tech Jobs Report shows that the number of job postings for IT support specialists has fluctuated within a relatively narrow range over the past year, with the lowest number of postings occurring in December 2023 and the highest number of postings occurring in September 2024. 

"Mileage may vary from company to company as firms are in different stages of adoption, but the aggregate picture suggests strong demand for this critical role for the foreseeable future," Robinson said. 

To join the CNBC Workforce Executive Council, apply at cnbccouncils.com/wec.