putilov_denis - stock.adobe.com

AI in network management poses challenges for network pros

Research shows that, while AI helps increase business success, network pros struggle to use it more than their peers. Find out how organizations can encourage network pros to use AI.

The rise of generative AI has led to AI taking over the business world. But don't take the adoption rates as a sign of endorsement. It was only in March 2022 when Pew Research reported that most people tend to err on the side of caution when it comes to AI.

Approximately 37% of 10,260 U.S. adults surveyed said they are more concerned than excited about AI, while 45% said they are equally as concerned as they are excited. They reported fear of job loss and the possibility of AI surpassing human skills as some of their top worries regarding AI. Network professionals, too, have these misgivings about AI and are hesitant to use it despite its use cases in network management.

Most IT professionals believe AI-driven network management will increase business success. But network teams struggle to use the technology, and enthusiasm has waned in recent years, said Shamus McGillicuddy, vice president of research at Enterprise Management Associates, during a webinar based on an EMA research report.

The study surveyed 250 IT professionals who engage with AI to some extent about AI-driven networks. It found that network professionals are among the most cynical users of AI in network management. McGillicuddy offered best practices to eliminate some of this skepticism.

Strategies for AI in network management

Respondents told EMA they use several strategies to implement AI in network management. Most IT pros follow a two-pronged approach, McGillicuddy said. They typically first choose AI, machine learning and AIOps products from network management tool vendors and then combine them with AI and ML products from network infrastructure vendors.

Regardless of the approach taken, McGillicuddy said respondents who reported using AI in IT management as a high priority experienced greater success with AI-driven networks. Those who placed less attention on AI tended to struggle more.

"That tells you when the CIO steers the ship toward AI-driven operations, the people in the networking silo -- trying to transform network management with AI – are having a better time with it," McGillicuddy said.

Use cases, benefits for AI-driven network management

Almost 57% of organizations told EMA network optimization is a use case of AI, and those from successful organizations were more likely to say so. Vendor management, network digital twins and anomaly detection were other use cases for successful organizations.

AI-driven network management is enabling network teams to help enterprises modernize how they build, maintain and deliver digital services to their users, end users, employees and customers.
Shamus McGillicuddyVice president of research, EMA

Approximately 47% of respondents also said network optimization is a benefit of AI-driven network management. Among other benefits, many respondents told EMA that AI helps network teams modernize application infrastructure, McGillicuddy said.

"AI-driven network management is enabling network teams to help enterprises modernize how they build, maintain and deliver digital services to their users, end users, employees and customers," he said.

Most IT professionals believe AI-driven networking creates benefits, but their opinions on how effective AI is vary among members of different IT teams. Over two-thirds of respondents said AI-driven networking has improved end-user experience, and almost a quarter said the increase was significant. However, IT professionals with the most success with AI were likelier to report a greater improvement, McGillicuddy said.

Additionally, 92% of respondents said AI-driven networking will improve business outcomes, which is a slight increase from 90% in 2021. However, the number of respondents who said they strongly agreed with this sentiment dropped from 56% in 2021 to 50% in 2023.

Network professionals struggle with AI in network management

The study found most network professionals struggle to evaluate the success of AI in network management. This challenge has less to do with the kind of strategies network teams use and more to do with how teams assess the strategies, McGillicuddy said.

Tool evaluation

Approximately 60% of respondents told EMA their ability to evaluate their success with AI in network management was only slightly effective or worse. McGillicuddy said IT professionals had different perspectives about how to evaluate tools.

Network engineers and NetOps staff, in particular, were more pessimistic than other IT teams and were less likely to rank their ability to evaluate tools highly. The percentage of successful AI users increased from 30% to 36% compared to 2021, but network professionals were still less likely to say they were successful compared to members of other IT teams.

AI-driven networking issues

Respondents reported issues on both the business and technical sides of AI in networking. Some business issues respondents said they experienced included security and compliance risks, skills gaps and budget issues.

Approximately 20% of respondents also said they lack trust in AI, and network engineers were twice as likely to say this, McGillicuddy said. This resistance comes from apprehension that AI can replicate human skills, he said.

Network complexity ranked as the top technical issue of AI in network management, with a quarter of respondents saying so -- and NetOps pros were twice as likely to cite this issue. Respondents told EMA they find that AI struggles to understand complex networks and the various parts of a system.

Increase AI optimism for network pros

Much of the cynicism about AI stems from a place of pessimism and mistrust. Network professionals have long feared the prospect of automation replacing human staff, but research has predicted that AI will assist rather than overtake human roles.

To mitigate apprehension, organizations should encourage network professionals to develop skills and engage in conversations with peers, analysts and vendors to understand how to use the technology, McGillicuddy said.

He added that network pros should consider using chatbots, such as ChatGPT, for network management. Some have already started to take advantage of the technology. For example, McGillicuddy said a network engineer told him they used ChatGPT to complete a full day of work within a mere 10 minutes.

Most network pros who told EMA their AI tools were successful committed to use the technology across their network wherever possible. AI in network management can help organizations create and manage nimble, modernized networks. Enterprises will continue to implement AI, so network professionals must embrace the technology as it becomes a vital tool in business.

Next Steps

NetworkGPT, ChatGPT plugins show promise for NetOps

The role of generative AI in networking

Dig Deeper on Network management and monitoring

Unified Communications
Mobile Computing
Data Center