The Evolution of the SOC: AI Agents Redefining Roles Rather Than Replacing Analysts

rifanmuazin - Thursday, November 19, 2026 | 04:46
Finance Poetry
The Evolution of the SOC: AI Agents Redefining Roles Rather Than Replacing Analysts

The landscape of modern cybersecurity operations is undergoing a profound transformation as artificial intelligence agents shift from experimental novelties to mainstream operational staples. According to the latest research from security startup Strike48, the vast majority of security leaders—a staggering 87%—are currently in the process of redesigning their security operations centers (SOCs) to accommodate the integration of AI agents. However, contrary to fears that this technological pivot would lead to the widespread displacement of human personnel, the industry is seeing a concerted effort to adapt and evolve existing roles rather than eliminate them.

The findings, detailed in the report The State of Agentic Security H2 2026, provide a comprehensive look at how organizations are grappling with the rapid influx of autonomous systems. Strike48 gathered data by polling 534 security and IT leaders across the United Kingdom, Canada, and the United States. To qualify for the survey, participants were required to be from organizations with 200 or more employees that maintain a dedicated security operations function.

The Rapid Rise of Agentic Security

The data indicates that AI agents have reached a critical tipping point in adoption. Approximately 74% of respondents reported that they have already deployed agentic technology within their SOC environments. This figure represents a dramatic acceleration in adoption, rising from just 36% in May. Perhaps even more telling of the technology’s growing utility is the fact that nearly half of the organizations surveyed (49%) are now utilizing these agents across multiple, distinct use cases, moving beyond single-task automation toward more complex, multi-layered workflows.

This rapid integration has historically fueled intense anxiety among industry professionals, particularly regarding the potential for AI to render entry-level security positions obsolete. The “Level 1” (L1) analyst role has long served as a vital career gateway, providing junior professionals with the hands-on experience necessary to climb the ladder of cybersecurity expertise. If AI were to fully automate the triage and identification of alerts, it could theoretically close this door, leading to a talent crisis down the line as the pipeline for senior security talent dries up.

Redefining the Human Element

However, the reality emerging from the Strike48 report suggests that organizations are choosing a path of professional evolution over workforce reduction. The data shows that rather than being laid off, L1 analysts are being strategically redeployed.

The shift is manifest in several ways. Half of the organizations surveyed (49%) are currently prioritizing hiring for emerging skill sets, such as AI detection engineering and specialized automation management. Simultaneously, 42% of respondents are focusing on retraining their existing L1 analysts to take on new, higher-value responsibilities—most notably the supervision and oversight of AI agents.

Furthermore, 38% of organizations are transitioning L1 personnel into deeper investigation roles, allowing them to focus on the complex, nuanced threats that require human intuition and critical thinking. Perhaps most surprisingly, 30% of companies are even continuing to grow their teams by adding more entry-level analysts. This indicates that while the nature of the entry-level role is changing, the necessity for human entry points remains robust.

These findings serve to allay some of the most persistent fears in the cybersecurity industry. There has been a growing narrative that the automation of triage would result in a “hollowing out” of the SOC. Earlier reports, such as one published by Swimlane in September, highlighted that 24% of SOC analysts already feel that current technology has begun to limit their ability to build foundational skills, with one in ten predicting fewer opportunities for juniors to gain experience. The Strike48 report offers a counter-narrative, suggesting that proactive management and a focus on “agentic supervision” can preserve these career paths while leveraging the efficiency of automation.

Most SecOps Leaders Are Changing the L1 Role as Agent Use Grows

Building Trust for Autonomy

Despite the clear enthusiasm for adopting AI agents, the transition is not without its hurdles. The report highlights a significant “trust gap” that currently constrains how effectively organizations can leverage these new tools. While the technology is being widely adopted, it is not being granted the autonomy that its architects intended.

Currently, only 30% of security teams permit their agents to operate with high or full autonomy. The vast majority of organizations remain cautious: 47% of teams require a human to manually execute actions or pre-approve every step taken by an agent, while 23% have implemented strict, limited boundaries that constrain the agents to specific, low-risk actions.

This hesitation appears to stem from a disconnect between expectation and implementation. Tim Leehealy, co-founder and president of Strike48, suggests that the problem is not with the maturity of the AI models themselves, but rather with the methodology used to integrate them into the SOC environment.

“Most teams start the same way: one agent, one long prompt telling it to act like a Level 1 analyst, and a pile of tools bolted on,” Leehealy told Infosecurity. “Then it meets a real alert queue, and you find out a single prompt can’t carry an entire triage process. In my view, this is why distrust of agent outputs is highest for teams with just one use case.”

Leehealy argues that the industry is currently going through a learning curve regarding the effective deployment of agentic architectures. The models, in his assessment, are sufficiently mature to handle a wide array of SOC tasks. The immaturity lies in the application layer—specifically in how teams structure their workflows.

“I’d argue the models are mature enough,” Leehealy noted. “What’s immature is how most teams use them. The SOC teams that get to agent autonomy break the work into narrow jobs and fixed, deterministic workflow steps.”

This philosophy—breaking down monolithic processes into granular, manageable tasks—appears to be the key to building the necessary confidence to increase autonomy. As teams move away from the “all-in-one-prompt” approach and toward a modular architecture, trust in the system grows, governance becomes more effective, and the agents become more reliable partners in the security operation.

As organizations continue to refine their approach to AI agents, the role of the human analyst is shifting from one of manual triage to one of oversight, strategy, and orchestration. While the tools of the trade are evolving at a breakneck pace, the fundamental requirement for skilled, human-led security operations remains as critical as ever. The challenge for leaders in the coming years will be to ensure that as they automate their SOCs, they also cultivate the human expertise necessary to manage the machines that have become their newest, and most active, employees.

Source : Finance Poetry
Author : rifanmuazin
Editor : Finance Poetry Editorial Team