AI agents are already changing what cybersecurity professionals actually do each day.
As organizations scale their use of AI agents, they also need to ensure employees have the skills to manage them.
That issue is at the center of the AI Workforce Consortium’s AI Agents and the Impact on Cybersecurity report.
Founded by Cisco and other technology companies, the consortium found that cybersecurity job postings requiring AI skills doubled year over year across G7 countries.
At the same time, organizations are still struggling to find cybersecurity talent. More than 4.8 million cybersecurity positions remain unfilled globally.
The demand for cybersecurity roles across G7 countries also grew 9.5% during the six months ending in March 2026.
I’ve never viewed AI as fully replacing cybersecurity jobs. Much like the shift to cloud, it’s just changing workflows and the hard skills security professionals need.
SOC analysts are becoming AI agent managers
Security operations centers (SOCs) offer a good example. AI agents can now manage tickets and perform initial triage, work that previously consumed an analyst’s time.
Omar Santos, distinguished engineer for AI Security at Cisco, said entry-level positions could see some of the biggest changes.
“The tier 1 SOC analyst position is one of the roles that is getting reshaped the fastest,” Santos told eSecurityPlanet. He said analysts are increasingly responsible for overseeing and orchestrating agentic systems.
Cisco is already seeing this transition internally. According to Santos, a group of roughly 80 cybersecurity professionals that previously performed security tests themselves now operate AI agents that find vulnerabilities and demonstrate potential attacker impact.
Human judgment becomes more important
Giving AI agents more responsibility does not remove people from the process. In some cases, it makes experienced human oversight even more important.
Marci Paino, chief learning officer at Cisco, said people still need to validate AI-generated results and assess the risk behind them.
“AI outputs can also be unreliable, sometimes in ways you wouldn’t expect,” Paino told eSecurityPlanet. “That unpredictability is exactly why human judgment is critical.”
The consortium’s research found demand for ethical reasoning in cybersecurity roles increased 533% year over year. Systems thinking increased 251%, while stakeholder engagement rose 125%.
In my 20-plus years in the industry, I’ve seen hiring managers and recruiters place a lot of emphasis on certifications and technical skills. Those skills still matter, but as AI takes on more of the work, knowing when to question its output could become just as important.
Organizations will also need to invest in the people managing these systems.
Cisco’s research found that 36% of cybersecurity leaders are prioritizing AI-powered defense, compared with just 25% investing in people to ensure they can manage the tech stack.
Entry-level cybersecurity faces an experience problem
AI agents are changing the skill set entry-level cybersecurity professionals need, but they don’t replace the fundamentals.
New security professionals still need hands-on practice with core areas like networking and Linux. They also need experience with security tools (e.g., SIEM) so they understand how the work is actually done.
That experience gives them the context to review an AI agent’s work, spot questionable results, and know when to investigate further.
Cisco’s report also found senior-titled cybersecurity job postings increased 65% during the six months ending in March 2026, compared with only 5.9% growth for junior-titled positions.
Paino said early-career professionals need more opportunities to gain practical experience with AI agents rather than simply learning about them in theory. Cyber ranges and virtual labs can provide a safe environment for that hands-on experience.
Employers can help close that gap by giving early-career professionals opportunities to apply what they have learned before entering their first full-time cybersecurity role. Paino pointed to internships and apprenticeships as two ways to build that experience.
The fundamentals are not becoming less important because of AI.
They are what give new security professionals the knowledge they need to use AI effectively and recognize when it gets something wrong.