In recent times, businesses have committed unprecedented resources to the development and integration of agentic artificial intelligence (AI) as a means of bolstering cyber defense against increasingly sophisticated AI-driven cyber threats targeting enterprise networks. Despite this significant investment, a troubling trend has emerged among cybersecurity teams, who largely remain hesitant to implement these technologies due to unresolved concerns regarding their efficacy.
The latest insights come from The Enterprise Threat Intelligence Report: 2026, published by Flashpoint on October 13. This report indicates that, despite widespread discussions about the cyber risks associated with offensive agentic AI applications, only a mere 25% of cybersecurity teams are currently making use of agentic AI tools as part of their active cyber defense strategies. Such statistics hint at a disconnect between technological potential and practical application within organizational security frameworks.
The report further highlights that 75% of security teams have yet to conduct a thorough exploration into the potential benefits of AI tools. Among those surveyed, 20% are currently engaged in closed pilot programs, testing these advanced tools under controlled conditions. Additionally, 35% are assessing the feasibility of incorporating AI solutions into their operations within the next 12 to 24 months. Conversely, 20% of respondents indicated an outright reluctance to experiment with agentic AI technologies in their cyber defense strategies for the foreseeable future.
For those organizations that have ventured into utilizing or testing agentic AI, the tasks performed by these tools range from basic information retrieval and alert prioritization to more proactive responsibilities such as blocking suspicious IP addresses, updating blacklists, and monitoring suspicious network traffic. Some entities have even employed AI for more complex operations, allowing automated systems to take actions that significantly impact security measures. However, the idea of empowering AI agents with extensive decision-making authority raises considerable apprehension among various stakeholders in cybersecurity.
Concerns have surfaced about the possibility of these agents making erroneous decisions. A notable sentiment shared by cybersecurity professionals encapsulates this fear: “An agent that’s confidently wrong is worse than one that’s slow.” This highlights a fundamental apprehension regarding AI’s reliability in high-stakes security scenarios. Additional obstacles hindering the widespread adoption of AI agents include potential issues related to data leakage, the need for comprehensive audit trails, adherence to regulatory requirements, the overall cost of implementation, and the extent of control that these agents should be permitted over security workflows.
Even as organizations contemplate the integration of agentic AI into their cyber defenses, many remain cautious and are conducting thorough research to determine whether these technologies can be trusted with critical tasks. Josh Lefkowitz, founder and CEO of Flashpoint, emphasizes this delicate balance, stating, “The opportunity for agentic AI on the defensive side is significant, and our research shows security leaders are ready to invest. But they aren’t swayed by the hype.” He elaborates that these leaders are demanding substantial evidence of how AI can enhance their operations, stressing the importance of knowing where human oversight remains essential in the security process.
Amid this ongoing deliberation, the report indicates that there is a palpable eagerness among board members and executive funding providers concerning the adoption of agentic AI technologies. Over half of the respondents expressed expectations for substantial budget increases that would facilitate investigations into the viability of AI for cyber defense. Specifically, 35% of organizations anticipate budget increases exceeding 10% over the next year, while an additional 15% expect increments of up to 10%. Notably, not a single respondent predicted a decrease in their cybersecurity budgets.
This trajectory of increased investment and cautious optimism reflects a broader trend in cybersecurity, where the interplay between emerging technologies and human oversight remains paramount. As organizations look toward the future, the challenge lies in striking a balance between leveraging the innovative capabilities of agentic AI and maintaining rigorous standards for security and reliability. The road ahead may be fraught with challenges, but the potential for enhanced cyber resilience through the thoughtful integration of AI remains a tantalizing prospect on the horizon.