Indian Cybersecurity Firms Use AI to Cut Vulnerability Testing to Hours

Indian cybersecurity firms including Indusface and Astra Security are adopting AI agents built on large language models to accelerate software vulnerability testing from days or weeks to hours, according to The Economic Times. The shift reflects growing attacker speed and AI tools’ emerging ability to identify exploits autonomously, prompting companies to adopt automated testing to keep pace with evolving threats.

Testing Time Acceleration

Large client security assessments that once required four to five days—or up to 20 days for larger applications—now complete within hours, according to Ashish Tandon, CEO of Indusface. This acceleration allows security teams to identify and address vulnerabilities more rapidly as the threat landscape evolves.

Attacker Speed and Vulnerability Growth

The urgency for faster testing is underscored by emerging data on attacker capabilities. CrowdStrike reported that the average attacker breakout time fell to 48 minutes in 2025. Meanwhile, Gartner projects that annual documented vulnerabilities will exceed 1 million by 2030, compared to approximately 277,000 in 2025—a near-fourfold increase.

AI Capability Expansion

Proofpoint, which expanded operations in India last year, noted that AI agents help review thousands of threat alerts daily. This automation addresses a critical challenge: companies face tightening data regulations and a shortage of qualified security analysts. The capability extends beyond alert triage; according to Anthropic, Claude Mythos Preview identified a bug in OpenBSD, an open-source operating system, that had remained undetected for 27 years. The same model achieved a 72.4% success rate in converting known vulnerabilities into working exploits, compared to 14.4% for Opus 4.6, an earlier Anthropic model.

Remediation and Organizational Disparity

While AI accelerates vulnerability discovery, remediation—the process of fixing security issues—remains a bottleneck requiring human review and approval. According to Arctic Wolf, a cybersecurity company, 76% of compromises in its incident response cases involved one or more of 10 known vulnerabilities for which patches were available before exploitation. This gap may widen across the security landscape: larger enterprises equipped with AI-driven detection and remediation capabilities may outpace smaller organizations lacking sufficient staff or budget to manage the volume of identified vulnerabilities.