Dmytro Olegovich Zakharchuk/Alamy
Many of the most advanced enterprises that have adopted technologies such as cloud and AI still rely on manual work to manage and maintain their network infrastructure. These initiatives are making the job of network operations (NetOps) teams increasingly difficult, due to their networks' complexity and scale.
The irony is that this is exactly where the value of network automation shines. It reduces human error, enhances security, increases productivity and saves on costs.
More strategically, automation is the engine that powers agentic NetOps, transforming network management from a reactive human process to a proactive, intelligent and autonomous function.
When you think of agentic AI as a team of interns guided by experienced engineers, it can help accelerate NetOps tasks because it never tires. But AI needs data and intelligence to get started, and it can't be in an engineer's head. If you haven't made backup and recovery a one-touch process and you struggle to keep up with manual patches and upgrades, then agentic AI won't be able to help.
Related:AI isn't magic: It takes discipline to gain business value
Automation initiatives begin by establishing a single source of truth. This provides real-time visibility into the network infrastructure. Key details include device performance and configuration, whether it's secure and compliant, when it was backed up, and what is vulnerable.
With a trusted, reliable data profile of your network and security devices, you can leverage AI effectively. AI uses high-quality data to perform simple searches, identify anomalies and make recommendations.
To further instill trust in AI, the role of network engineers shifts to one where they provide expert oversight. Network engineers review the findings, validate and weigh the impact of the recommendations and create AI's guardrails. Training agentic NetOps to be trustworthy is a more valuable skill set than repeatedly executing mundane tasks.
The value of network automation is clear: automating simple NetOps tasks that you perform daily, or need to perform more frequently and reliably and then enabling agentic NetOps. So, why are two-thirds of enterprise networking activities still performed manually?
Based on my discussions with network infrastructure owners, there are several reasons why automation has not progressed as expected.
Market confusion. The concept of automation has been broadened and overcomplicated by including orchestration. Business leaders have a hard time differentiating between the two. Only a subset of companies, such as telcos and ISPs, use their networks as revenue generators. Those types of businesses need complex orchestration tools to support the heavy service deployments that generate business value. In contrast, the tasks that all companies must perform to maintain and manage their infrastructure are crucial in allowing it to be resilient, secure, compliant and reliable. Such tasks are simpler and can be easily automated. There's no need for complicated and expensive orchestration capabilities.
Perceived skills gap. Activities such as device discovery, backup and recovery, patching, upgrades and compliance checks are highly automatable today, and don't require network engineers to become coding experts. Instead of a blank-slate approach, low-friction solutions that include prebuilt automations for most of these everyday tasks and a no-code approach to customization make it easy to get started.
Distraction. Network infrastructure now consists of multiple hybrid clouds and campus networks, while on-prem networks and data centers still exist. The move to AI is driving additional modernization challenges related to bandwidth and availability requirements. Meanwhile, organizations acquire other businesses that use different vendors for their infrastructure.
Related:Charting the path to the autonomous enterprise
As the complexity of multivendor environments and software-defined networks compounds, the job of NetOps teams becomes increasingly difficult. The NetOps teams struggle to maintain visibility into all their network and security devices to manage, optimize, and secure them. This is the very reason why automation is needed. However, it's difficult for teams to strategically implement automation while focusing on "keep-the-lights-on" work and the evolving demands of the business.
Related:Automation Alternatives to AI
In lieu of automation, companies often throw people at the problem, hoping that their brain trust won't retire or move to another company. The situation has come to a head for network infrastructure leaders who are increasingly being asked how they can use AI to reduce headcount or to do more with existing staff. If they haven't started to automate network activities, it's hard to answer with confidence.
Common pitfalls to network automation stem from a legacy of being hard to scale and expensive. These misconceptions are holding organizations back. More recently, automation has advanced to the point that it is easy to get started with no coding required and is suited for the reality of increasingly heterogeneous environments. A modern approach democratizes the usage of automation and, by extension, agentic NetOps.
When you start with automation, the move to agentic NetOps is transformational. It's not hard. Automation can convert network engineers into managers of interns with a trusted automation framework underneath. That's how leaders can continue to scale to support their business's increasingly complex and expanding infrastructure, and do so efficiently and effectively, with a winning hand that includes AI.
BackBox
Rekha Shenoy is CEO of BackBox. With more than 25 years in B2B tech, she has held leadership roles at Tripwire, BMC Software and Belden, and was named a 2025 Cyber Leader of the Year.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。