In partnership with

Richard Stiennon Agentic AI Security Panel Series

Richard Stiennon leads each session as a discussion among technology vendor founders, all of whom have deep expertise in the topic. There are zero product pitches or any marketing messaging: the aim is to reduce the knowledge gap that is rife among CISOs and other practitioners.

Panel Topics

14 initial sessions on various Agentic AI Security topics. Each panel has its own page with the session details and open discussion.

  1. Measuring Access Control for Enterprise AI Agents How do you prevent autonomous enterprise AI agents from illegally accessing sensitive corporate repositories or dangerously escalating their privileges? This high-impact session explores robust technical frameworks for quantifying data restrictions and auditing permissions without bottlenecking your daily workflow velocity. Join top cybersecurity leaders to master the delicate balance between necessary agent autonomy and strict corporate data boundaries. View panel
  2. Measuring Identity & Authentication for AI Agents When autonomous AI agents act independently on behalf of human users, traditional enterprise identity protocols instantly crumble. This expert panel dives deep into modern machine-to-machine authentication frameworks designed to safely verify agent identity and manage cryptographic keys. Discover the cutting-edge standards required to secure automated enterprise workflows and trace digital lineage across complex multi-cloud environments. View panel
  3. Measuring Agent Resilience to Malicious Inputs & Prompt Attacks Prompt injections, data poisoning, and jailbreaks are actively disrupting autonomous enterprise systems and severely compromising corporate safety. This urgent discussion breaks down the technical methodologies used to stress-test and accurately measure agent resilience against sophisticated adversarial inputs. Learn how to effectively benchmark your defensive guardrails and harden core reasoning pathways against these emerging exploit techniques. View panel
  4. Testing Agent Kill Chain Containment Scenarios What specific fallback mechanisms activate when an autonomous enterprise AI agent goes completely rogue or gets malicious code compromised? This critical session simulates worst-case security failures to test automated kill-switches and network isolation protocols. Join leading security engineers to learn how to design unbreakable enterprise containment strategies that instantly sever malicious access lines during an incident. View panel
  5. Testing Agent Observability & Forensics Because autonomous AI systems reason dynamically, traditional static logs utterly fail to capture critical post-incident forensic data. This deeply technical panel explores advanced operational frameworks required to trace complex agent memory states and decision-making timelines. Discover actionable strategies to bring full operational transparency to black-box systems and confidently audit multi-step agent actions in real time. View panel
  6. Comparing Human in the Loop & Review Frameworks Does injecting continuous human oversight into your architecture inevitably destroy the operational velocity of autonomous AI systems? This panel evaluates the absolute efficiency, security overhead, and enterprise scalability of modern Human-in-the-Loop and review frameworks. Learn exactly where to place approval gates within critical decision pathways to maximize system safety without sacrificing any necessary organizational throughput. View panel
  7. Assessing Agentic Supply Chain Security Your deployed agent is only as secure as its third-party plugins, foundational open-source models, and connected enterprise APIs. This eye-opening session dissects hidden vulnerabilities within the broader agentic ecosystem, focusing heavily on dependency risks and unauthorized data-leakage vectors. Learn how to comprehensively map, audit, and secure your automated software supply chain against sophisticated upstream exploitation. View panel
  8. Benchmarking Agent Sandboxing & Containment Techniques Isolating highly active autonomous agents requires technical network security measures far beyond basic, standard containerization practices. This deeply technical discussion benchmarks the latest virtualization tools, secure runtimes, and restricted micro-virtual machines. Discover which specific sandboxing environments provide the absolute strongest defense against host compromise, lateral network movement, and severe privilege escalation attacks within your infrastructure. View panel
  9. Reviewing Governance Frameworks for AI Agents With regulatory pressures rapidly mounting across the globe, enterprise technology leaders must urgently establish clear compliance standards for autonomous AI. This forward-looking panel reviews emerging global governance frameworks, corporate accountability models, and risk-management strategies. Discover how to align agent capabilities with strict legal mandates while successfully fostering a proactive culture of rapid, competitive organizational innovation. View panel
  10. Evaluating Agent Swarm Detection & Mitigation Coordinated networks of autonomous agents introduce unprecedented threat vectors, ranging from hyper-automated fraud to distributed denial-of-service attacks. This cutting-edge panel explores real-time behavioral monitoring and anomaly detection strategies built explicitly to neutralize malicious agent swarms. Learn how to identify synchronized automated behaviors instantly before they completely overwhelm your enterprise's digital infrastructure or compromise critical business operations. View panel
  11. Profiling Runtime Behavioral Assurance for AI Agents Static pre-deployment code analysis cannot predict how an autonomous agent will dynamically behave when facing real-world enterprise variables. This session introduces cutting-edge runtime behavioral profiling frameworks designed to catch anomalous agent actions instantly. Learn how to continuously monitor underlying machine intent, detect unexpected behavioral drift, and enforce strict, real-time policy boundaries during live, mission-critical operations. View panel
  12. Profiling Agent Performance vs Safety Tradeoffs Can an autonomous agent remain completely secure without becoming painfully sluggish or heavily restricted in its operational capabilities? This strategic session tackles the inherent friction between raw operational performance and rigorous security guardrail enforcement. Explore data-driven case studies balancing automated throughput with risk mitigation to maximize both your competitive market edge and your defensive posture. View panel
  13. Reviewing the Agent Build vs. Buy Dilemma Should your organization build custom agent architectures completely from scratch or rely heavily on commercial, off-the-shelf platforms? This highly practical session evaluates the unique security implications, proprietary IP ownership rules, and long-term maintenance costs of both approaches. Gain clear, actionable insights into customization limitations and vulnerability patching overhead to guide your technology and safety roadmap. View panel
  14. Assessing Agent Deployment Scenarios & Time to Value Deploying production enterprise-grade autonomous agents requires successfully navigating a highly complex landscape of cloud, on-premises, and hybrid container architectures. This panel analyzes real-world deployment patterns, directly mapping critical security friction points against actual time-to-value metrics. Discover how leading organizations bypass integration bottlenecks to safely accelerate their journey from initial proof-of-concept to secure, measurable enterprise ROI. View panel

Current Panelists

The founder roster for this series is being confirmed and will be announced here.

About the Host

Richard Stiennon

Chief Research Analyst, IT-Harvest

Richard is the author of Guardians of the Machine Age and Chief Research Analyst at IT-Harvest, where he tracks every vendor in the cybersecurity industry. A former Gartner research VP, he has covered the industry for more than two decades.