Cisco’s security team recently scanned 1.8 billion lines of code across 25 programming languages in eight weeks. Before artificial intelligence, that task would have taken eight years. The speed gain is impressive, but the real lesson goes deeper: the operating model for cyber defense needs a fundamental overhaul.

The old approach—finding vulnerabilities, patching them, and reacting after an attack—was built for a world where threats moved at human speed. That world is gone. Frontier AI models can now discover flaws, generate exploits, and chain them into attack paths in hours. Meanwhile, enterprises average 43 days to patch a critical vulnerability. In 2025, 40% of the most targeted weaknesses affected systems with limited or no patch options. That gap is now a major risk, and it demands a continuous, AI-accelerated defense posture rather than periodic projects.

Cisco’s executive brief outlines three practical shifts for teams making progress. First, make remediation a predictable rhythm, not an emergency scramble. Second, retire unsupported and end-of-life assets to eliminate exposure that security teams shouldn’t have to carry. Third, raise the overall security posture by segmenting networks, closing identity gaps, and equipping SOC teams with faster tools. No organization needs a perfect program to start—each step reduces risk and makes the next one easier. The goal is to move from point-in-time security to a continuous discipline that builds resilience and trust.

Source: Cisco