Companies accelerate adoption of AI agents for digital defense

The proliferation of artificial intelligence agents in the corporate environment is no longer a futuristic hypothesis, but a tangible reality that is transforming companies digital defenses. A May 2025 Gartner study of 147 CIOs and CTOs revealed that 24% of companies had already implemented some AI agents, with more than half of these systems operating in internal areas such as technology, human resources, and accounting, compared to only 23% in external customer-facing functions.

This trend, according to the firm's principal analyst, Avivah Litan, is perceived as "moderately beneficial" in the field of cybersecurity, although questions remain about these agents' ability to scale beyond simple tasks.

The advancement of generative AI and large-scale language models has radically altered the cybersecurity landscape. Attackers now have accessible tools that allow them to create convincing voice and video deepfakes, personalized phishing campaigns, and malicious code with unprecedented ease. This new scenario has forced companies to incorporate AI into their defensive strategies, especially through autonomous agents capable of detecting, analyzing, and alerting on threats in real time.

According to Brian Murphy, CEO of cybersecurity technology firm ReliaQuest, the magnitude of the challenge is considerable: "It's a huge challenge to detect, contain, investigate, and respond in large organizations." Murphy argues that AI can filter out noise and eliminate low-value tasks, traditionally assigned to the first levels of security teams and rarely providing relevant information about real threats.

The automation of repetitive and laborious tasks has been one of the main arguments for the adoption of AI agents in the business environment.

Amazon CEO Andy Jassy conveyed to his employees in June his conviction that “AI agents will change the way we all work and live,” anticipating a future with “billions of these agents, across every company and in every imaginable field,” allowing workers to “focus less on mechanical tasks and more on strategic thinking,” as well as making “our jobs even more exciting and fun than they are today.”

In the field of cybersecurity, Murphy sees an industry saturated with routine work that shouldn't consume so many human resources, contributing to burnout and exacerbating the shortage of specialized talent. He also highlights how AI has raised the level of sophistication of attacks: “Those phishing emails used to be almost laughable due to spelling and formatting errors. Now, AI can turn a mediocre attacker into a much more effective one, so if you're on the defensive end, you have to use AI because of what this technology is capable of.”

ReliaQuest recently launched GreyMatter Agentic Teammates, autonomous AI agents with defined roles that take on tasks previously assigned to detection engineers or threat intelligence analysts within security operations teams.

Murphy describes these agents as “teammates” who collaborate with humans, who direct and supervise them, thus multiplying incident response capabilities.

A common example in global companies is managing international executive travel: every time a laptop or phone connects to a network in countries like China, the security team receives an alert and must verify daily that the device is being used securely. With an AI agent, this process can be automated, and the same is true for other common situations like board meetings or corporate events. “There are hundreds of cases like that,” Murphy says.

Automation isn't limited to reviewing logs or generating alerts. Justin Dellaportas, chief information and security officer at communications technology company Syniverse, notes that AI agents can already perform actions such as isolating suspicious emails and removing them from inboxes, or restricting access to compromised accounts across multiple systems.

Dellaportas warns that cybercriminals are using AI to identify vulnerabilities and exploit systems on a massive scale, allowing them to gain earlier access and move laterally within organizations faster than ever. “Digital defenders must embrace this technology more than ever to stay ahead of an ever-evolving threat landscape and the pace of cybercriminals.”

The integration of AI agents into cybersecurity is not uniform. Dellaportas describes adoption as a gradual process, a “crawl, walk, run methodology.” Initially, agents reason and act, but they must learn from their previous actions. “I trust but I verify, and as we gain confidence in their effectiveness, we tackle new problems.”

Regarding the impact on cybersecurity professionals, Dellaportas believes that AI agents can take on certain tasks, but their primary role will be to enhance workers effectiveness, not replace them. Murphy agrees, emphasizing that automation helps fill the skills gap faced by many organizations. “There may be a lack of trained and skilled cybersecurity professionals, but there is no shortage of people interested in training. Knowledge transfer is slow because entry-level positions are equivalent to working in a help desk.”

Murphy acknowledges that much remains to be done in terms of training and understanding how AI agents make decisions. Dellaportas adds that familiarity with these systems has grown because they are used in various business areas, which facilitates discussions about their usefulness in achieving business objectives.

In the cybersecurity sector, AI has established itself as one of the most immediate applications. Litan recalls that the technology began to be used in fraud detection and that "it's certain that in the future we will have digital security assistants that free up staff to confront new attacks; the key will be for these systems to evolve at the pace of innovation to cover the entire attack surface."

Murphy predicts that the adoption and evolution of AI agents in corporate cybersecurity will advance even faster than in sectors like finance or law. “Companies fully understand that AI is being used against them, and the only way to defend themselves is to use it in their own defense.”