9:50 AM – 10:30 AM | Breakout Sessions – Round 1

CrowdStrike Session:

Securing the AI Era: Introducing the Agentic SOC

In the AI era, adversaries are increasingly weaponizing artificial intelligence models to execute attacks at machine speed, frequently overwhelming traditional Security Operations Centers (SOCs). Security teams are often hindered by operational inefficiencies, including fragmented security toolsets, critical resource shortages, and the persistent challenge of analyst burnout driven by high volumes of alert noise. To defend against these sophisticated, AI-driven threats, organizations must move beyond reactive measures toward a cohesive, intelligent, and proactive defense strategy.

This presentation introduces the "Agentic SOC," a transformative paradigm that bridges the gap between human intuition and the unmatched speed and precision of artificial intelligence. We will discuss the three steps critical to enabling this transformation — Onboarding, Operationalizing, and Orchestration, each of which is meticulously designed to establish a robust, AI-ready, data-driven security foundation. Through this strategic transformation, security teams can fundamentally improve incident response times, enhance triage accuracy, and dramatically reduce burnout, even when operating at massive scale. Attendees will walk away with actionable strategies to modernize their security infrastructure, ensuring they not only adapt to the current threat climate but actively outpace it. By embracing this evolution, organizations can secure improved operational health and foster a sustainable security culture in an era defined by continuous adversarial innovation and relentless automation.

• Presented by: [Presenter TBD]
• Location: [Room TBD]

ExtraHop Session:

The Specification is Not the Mission: Authoring Trust in Agentic AI

Federal guidance on trusted AI focuses on the model: evaluate it, red-team it, monitor it. Necessary, but incomplete. The most consequential AI failures of the past year were not model failures but specification failures: systems faithfully optimized objectives humans wrote, without the constraints they assumed. Machines can optimize a specification. They cannot want the mission. This session argues that trust in AI begins with the objective, not the model. In federal mission settings, where success is ongoing rather than final, AI must support a division of labor it cannot invert: machines answer, humans ask. Drawing on recent agentic incidents and two decades of building machine learning for cyber defense, this talk offers a practical framework for authoring objectives, constraints, and accountability so AI advances the mission, not just the task.

• Presented by: [Presenter TBD]
• Location: [Room TBD]

10:35 AM – 11:15 AM | Breakout Sessions – Round 2

F5 Session:

Securing the AI Era: Extending Zero Trust with Guardrails, Gateways, and API Protections

Agencies are deploying AI and agentic systems faster than their security architectures were designed to accommodate. Autonomous agents call APIs, move data, and act on behalf of users at speeds that can render traditional controls irrelevant. Zero Trust still offers the right foundation, but it does not yet account for a subject that reasons probabilistically.

AI guardrails, gateways, and API protections close the gap. Guardrails apply explicit verification to model behavior, inspecting prompts, tool invocations, and outputs against agency policy. Gateways act as policy enforcement points for model traffic, brokering access and containing incidents before they reach downstream systems. Often overlooked, but critically important in the age of AI, API protections carry least-privilege authorization into the machine-to-machine calls that agentic workflows generate.

• Presented by: [Presenter TBD]
• Location: [Room TBD]

Qualys Session:

Governing Cyber Risk in the Frontier AI Era

EO 14409 and CISA BOD 26-04 mark a structural shift in federal cyber risk governance. Remediation timelines for the highest-risk vulnerabilities now compress to as little as three days, and CVSS-only prioritization has been formally retired in favor of a risk-based model weighing asset exposure, KEV status, exploit automation, and technical impact. Against a 55-day average remediation rate across federal agencies, this represents a material gap between policy and operational reality — one with direct implications for compliance posture and leadership accountability.

This briefing examines what BOD 26-04 and EO 14409 require of FCEB agencies, the risk this shift creates for agency leadership, and the operating model needed to close the gap: hyper-prioritization aligned to the directive's four risk variables, zero-day remediation capable of meeting compressed deadlines, and continuous AI-speed detection that keeps prioritization current as threats evolve. Attendees will leave with a clear view of what changes now and what it takes to stay ahead of both adversaries and auditors.

• Presented by: [Presenter TBD]
• Location: [Room TBD]

2:25 PM – 3:05 PM | Breakout Sessions – Round 3

Wiz Session:

Leveraging Continuous Risk Management with Code-to-Cloud Defense to Support cATO

Traditional Authorization to Operate (ATO) processes rely on static, document-centric reviews that stall software delivery and mask emerging operational risks. Continuous Authorization to Operate (cATO) shifts risk management from point-in-time compliance bottlenecks to continuous, real-time security posture validation.

This session presents a practical approach for achieving scalable cATO by integrating automated risk assessment across three core categories:

• Secure Software Supply Chain (SSSC): Shifting security left into developer workflows (IDE, PRs, and CI/CD pipelines) to align with Secure by Design best practices to remediate vulnerabilities, misconfigurations, and secrets before deployment.
• Continuous Monitoring (CONMON): Delivering unified, multi-cloud visibility and graph-based context to aggregate indicators of risk into "toxic combinations" to enable real-time risk evaluations and remediation prioritization based on threat likelihood and potential mission impact.
• Active Cyber Defense (ACD): Connecting runtime threat signals with code-to-cloud context, accelerating blast radius and root-cause analysis to rapidly detect, contain, and mitigate true threats within operational environments.

Attendees will gain actionable strategies to automate NIST SP 800-53r5 control validation, streamline evidence collection, and accelerate secure software delivery to enhance true mission readiness and cyber resilience.

• Presented by: [Presenter TBD]
• Location: [Room TBD]

Zscaler Session:

Accelerate AI, Not Risk: Zscaler's End‑to‑End AI Security Platform

As public sector and highly regulated organizations rush to harness the revolutionary power of Generative AI, the mandate is clear: we must accelerate innovation, but never at the expense of security. In this session, discover how the Zscaler platform serves as the ultimate enabler for responsible AI adoption—providing the strict guardrails, data loss prevention (DLP), and governance required to democratize AI securely.

With deep expertise in guiding C-suite leaders and delivering on federal mission objectives, Chad Tetreault connects public service insights to the practical deployment of advanced security architectures. This session focuses on strategically aligning AI initiatives with core mission objectives, preventing data leakage to public LLMs, and securely unlocking the full potential of AI-driven transformation while mitigating risk.

• Presented by: [Presenter TBD]
• Location: [Room TBD]

3:10 PM – 3:50 PM | Breakout Sessions – Round 4

GuidePoint Security Session:

[Session Title TBD]

[Session description TBD]

• Presented by: [Presenter TBD]
• Location: [Room TBD]