Where traditional cybersecurity frameworks struggle with lag in detection and response, agentic AI offers a different model altogether
As cyber threats grow in scale and sophistication, security operations are undergoing a profound transformation driven by the emergence of agentic AI—autonomous systems that can detect, analyse and respond to threats with minimal human intervention. Once confined to speculative discussions, these systems are now quietly becoming integral to modern cybersecurity strategies.
For Jaspreet Bindra, CEO of AI&Beyond, the dual-edged nature of such autonomy is clear. “AI has been used to execute cyber attacks for some time now,” he says, pointing to the disturbing case of a planned explosion in Las Vegas in early 2025, where ChatGPT was allegedly used to coordinate the attack involving a Tesla Cybertruck. “The 2022 Activision breach, initiated through AI-enhanced phishing attacks, was another stark example,” he adds.
These incidents, he argues, are not just footnotes in a growing list of cyber attacks but a sign that the tools designed to defend networks can also be weaponised. “Agentic AI agents can autonomously detect threats, prioritise risks, and respond faster than human teams. But their autonomy also introduces vulnerabilities,” he warns. “Misaligned objectives or exploitation by malicious actors can lead to unintended consequences. Moreover, the lack of transparency in decision-making processes raises concerns about accountability and trust.”
Despite the risks, the promise is immense—when implemented responsibly. “Proper governance, human oversight, and ethical guidelines are non-negotiables,” Bindra says. “Balancing innovation with security is imperative to leverage agentic AI effectively in the evolving cybersecurity landscape.”
From reactive to real-time
Where traditional cybersecurity frameworks struggle with lag in detection and response, agentic AI offers a different model altogether. “Agentic AI represents a significant leap,” says Ankit Sharma, Senior Director and Head – Solutions Engineering at Cyble. “It shifts us from reactive models to intelligent systems capable of autonomously detecting, analysing, and mitigating threats in real time.”
According to Sharma, the real power lies in speed and scale. “This evolution reduces response times, adapts to complex attack vectors, and scales defences without overburdening security teams,” he notes. “In high-volume, high-velocity threat environments, that agility is invaluable.”
Still, Sharma underscores the need for constraint. “Autonomy must not come at the cost of accountability,” he says. “As these systems begin to operate independently, strong governance frameworks, ethical constraints, and human oversight are crucial.”
India’s agentic AI ambitions
India is not watching from the sidelines. Startups are exploring home-grown innovations in this space, betting on language models and sovereign AI as strategic tools for national cybersecurity.
“Agentic AI is a transformative force,” says Ankush Sabharwal, Founder and CEO of CoRover. “Our autonomous AI Agents, built using insights from India’s first LLM, BharatGPT, are designed to proactively identify and neutralise threats by continuously analysing network behaviour and automating intelligence gathering.”
Sabharwal outlines a lifecycle-based approach, rooted in composite AI models. “This proactive stance allows for adaptive and evolving defence mechanisms,” he says. “These intelligent entities enhance our ability to safeguard digital assets and support the vision of Sovereign AI.”
While AI-powered agents promise enhanced security and a reduced burden on overworked security operations teams, their use must be tempered with awareness of the ethical and technical pitfalls. The question is no longer whether agentic AI will become part of our cybersecurity infrastructure—but whether we are ready to wield it wisely.
