ArmorCode Sponsors Inaugural Purple Book Community Conference Tackling Agentic AI Security

Two-day virtual event brings together CISOs and security leaders to confront vulnerability chaining, shadow AI, and securing autonomous AI agents

ArmorCode, the leader in Unified Exposure Management, today announced it’s sponsoring and speaking at the inaugural Purple Book Community (PBC) Agentic AI Security Conference, a free virtual event taking place October 21 and 22 from 9:00 – 11:30am PT. The conference brings together CISOs and security executives from leading frontier model and Fortune 500 organizations to examine how AI agents are changing the way vulnerabilities are found, exploited, and defended against. Co-sponsored by KPMG, Semgrep, Escape and The AI Collective, attendees will walk away with practical steps they can implement to keep software safe from modern threats, as well as insights into managing risk at the pace of AI innovation.

Also Read: Rezolve Ai to Participate in Maxim Growth Summit 2026

Nikhil Gupta, founder and CEO of ArmorCode and founding member of the Purple Book Community said: “AI agents are already operating inside enterprise environments with real access, making decisions and taking actions. Yet most security teams don’t have a playbook for using agents for security and securing the agents themselves. This conference brings together the security leaders solving these challenges today with those who will be solving them next. It’s an opportunity for peers to share firsthand experiences and lessons learned as they confront the evolving security challenges of agentic AI.”

The AI-Scale Exposure Gap

The conference focuses on the growing AI challenges facing security teams. With AI-powered vulnerability discovery and more agents gaining access to enterprise environments, security teams have more risk to manage with the same finite resources. The impact is already visible in the numbers.

Security teams today manage output from more than 45 different security tools, with findings routinely reaching into the millions. No team can realistically review that volume manually, so many organizations rely on CVSS severity scores to prioritize remediation. But CVSS measures the severity of a vulnerability, not the risk it poses to a specific organization, making it easy for teams to overlook the exposures that matter most.

Also Read: Manhattan Wins Kesko Deal for AI-Driven Warehouse Management

AI has made this problem worse in three ways:

• More code, more vulnerabilities: The volume of published CVEs grew more than 260% between 2020 and 2025, and 2026 is expected to be the first year with more than 50,000 published CVEs.
• Faster exploitation: Frontier AI models can find and validate software flaws faster than human security teams can triage them. Software vulnerabilities have now overtaken stolen credentials as the leading way attackers gain initial access to organizations.
• AI adoption is outpacing governance. Enterprises are adopting AI tools faster than security teams can track or govern them, with shadow AI increasingly appearing in breach investigations.

Against this backdrop, the PBC Agentic AI Security Conference brings together leading voices from across the technology and enterprise landscape to share perspectives and approaches for addressing these challenges.

Phil Venables, Partner at Ballistic Ventures and former CISO of Google Cloud and Goldman Sachs, who’s delivering the event’s closing keynote, said: “Agentic AI is changing the security equation for enterprises. AI systems are moving beyond generating recommendations to taking actions, accessing systems and interacting with software at a scale and speed that traditional security processes were not designed to handle. The challenge is not simply finding more vulnerabilities. It’s understanding which exposures can be combined, which create meaningful risk, and how to address them before attackers take advantage. The Purple Book Community is helping move this conversation from theory to practice by bringing experienced security leaders together to share what they’re seeing, challenge assumptions and develop approaches that work in real enterprise environments.”

Also Read: d-Matrix Announces Squadron Compute Clusters for AI Inference

Session Tracks and Topics

The conference runs 12 sessions across two parallel tracks. Each session digs into a real problem security teams are facing today as a result of agentic AI adoption, including:

• The changing role of vulnerability management as frontier AI models reshape how vulnerabilities are discovered
• Controlling shadow AI and vibe coding
• The threat of vulnerability chaining, in which attackers combine multiple low-risk vulnerabilities into a single critical threat
• Securing AI agents operating in production environments with real system access and permissions
• And managing the growing backlog of security issues created by AI-generated code at enterprise scale.