According to Precedence Research, the global market for agentic AI in cybersecurity is projected to grow from about $30 billion this year to $147 billion by 2034. Marty McDonald, principal security advisor at Optiv, explains why there's so much interest in agentic AI among organizations. "Agentic AI enables automation of tasks that humans would typically perform, such as gathering data, analyzing it, and summarizing findings. In advanced environments, agents can monitor threat intelligence feeds, run queries, and provide real-time context and summaries for threat hunters," McDonald explains.
As organizations move to embrace AI in their security operations programs, the most successful in its adoption will be those that implement AI in phases, enabling gradual integration, testing, and a smooth transition. This will help organizations better navigate the complexity of layering AI on top of modern security operations.
The core challenge isn't just adopting best practices, but fundamentally asking why and where AI will bring the most value. This requires a thorough assessment of their existing program to understand where it excels and where it could use assistance. Then, AI can be applied initially in high-impact use cases, such as where it can measurably accelerate detection, reduce false positives, and free analyst time for higher-order work.
Decision-makers should ask themselves:
What are the real "pain points" that current tools or processes fail to address meaningfully?
Where do human analysts get overwhelmed by data volume, complexity, or alert fatigue?
Which functions have previously proven resistant to automation, and why?
How do we ensure AI deployments don't create new blind spots or risks?
How will AI initiatives directly tie to critical business outcomes beyond cost savings?
When the best AI use cases are understood, along with well-defined objectives, the deployments need to be properly governed, secured, and designed to improve:
Develop Strong Governance Structures: Begin by establishing clear policies and procedures that define the intended roles and acceptable applications of agentic AI within your security operations. These should detail decision-making protocols and mandate compliance with industry regulations, ensuring all team members are aware of their responsibilities and operational boundaries. Continuously update these governance models to reflect evolving threats and new legal requirements, fostering organizational transparency and accountability.
Strengthen Data Oversight and Quality: Make data governance a priority by implementing comprehensive validation processes and safeguarding against bias in the information supplied to AI tools. Rigorous source verification and quality assurance measures preserve the integrity and confidentiality of your data, enhancing the dependability of AI-driven actions. Emphasizing data privacy also helps your organization avoid breaches and meet international standards. "The effectiveness of agentic AI is highly dependent on organizational maturity. Less mature organizations struggle to gain value because they lack formalized processes and quality data, leading to faster but not necessarily better," says McDonald.
Boost Transparency and Monitoring: Enhance the clarity of AI decision-making by leveraging explainability solutions and maintaining thorough audit logs. This approach strengthens stakeholder trust and facilitates the quick detection and resolution of anomalies or errors. Increased transparency also streamlines compliance reviews and demonstrates the ethical application of AI to regulators and business partners.
Prepare the Organization and Upskill Teams: Assess your company's infrastructure, workforce capabilities, and organizational culture before integrating AI, identifying any potential shortcomings. Offer targeted training to enable staff to effectively work alongside automated tools, allowing them to focus on more strategic activities. Such preparation reduces pushback and maximizes efficiency gains from AI adoption. David Marcus, federal senior security technologist and principal engineer at Intel, says that such training is essential. "You want to take those analysts who are level one and level two analysts and upskill them to start performing AI work, such as performing the care and feeding of AI agents and models. These models aren't going to run themselves, at least not for a long time," Marcus says.
Adopt Modular and Scalable AI Deployment: Begin integration in low-risk segments of your security operations by utilizing adaptable frameworks that can be expanded as requirements evolve. This incremental approach enables real-world testing while minimizing disruption to essential workflows. Scalable designs ensure your systems can adapt to future demands and business growth.
Address AI-specific Risks Proactively: Do this by recognizing potential vulnerabilities unique to AI, such as system manipulation or inadvertent errors, through dedicated risk assessments and targeted mitigation strategies. Addressing these factors early prevents smaller issues from escalating and strengthens your overall security posture, ensuring resilience in autonomous operations.
As organizations continue to deploy and manage their Agentic AI systems, they will encounter continuous challenges related to capability, data silos, and disconnected systems, hindering seamless integration. To minimize these issues, organizations should prioritize interoperability when selecting security tools and maximize the use of integration frameworks and APIs to enable data exchange between systems.
Also, for long-term success, one of the most critical factors will be training and workforce development. Security professionals must develop proficiency in AI-powered security platforms while maintaining their core expertise in threat analysis and incident response. This dual competency—technical AI skills combined with deep security knowledge—will define the most valuable cybersecurity professionals in the AI era.
Agentic AI will change security operations; it's just a question of how much. While the level of AI hype is high, it's a present reality, and it's moving fast. Organizations that approach this transition strategically, with proper governance, phased implementation, and workforce development, will likely emerge more efficient, stronger, and more secure; those who resist and delay risk falling behind in a competitive disadvantage due to outdated security operations.
"The cybersecurity workforce of the future will eventually be smaller in some traditional roles but more skilled, more strategic, and more effective in others when it comes to protecting against ever-changing threats. However, for security professionals, the message is clear: keep their AI skills polished," says Marcus.
