AI Video Security in 2026: What’s Real, What’s Next, and What’s Hype?

Join Scylla AI for an expert-led webinar on October 29, 2026 at 12PM EDT exploring the technologies shaping physical security in 2026 and where they can deliver practical value.

Reliable detection remains the foundation of AI video security. Vision-language models (VLMs) can add context to detected events, support video search, and assist with alarm triage. Their value lies in complementing established computer vision pipelines, while keeping security teams in control of response decisions. As “agentic AI” claims gain attention, how can security professionals distinguish useful capabilities from overstated promises?

Join Albert Stepanyan, Co-founder and CEO of Scylla AI, and Armen Ghambaryan, Lead Deep Learning Engineer at Scylla AI, for a discussion of detection, contextual intelligence, and responsible automation, supported by practical technology demonstrations.

During the session, we will discuss in detail:

How video analytics is evolving from predefined object and event detection toward systems capable of interpreting richer visual context and what that means for security teams.

Where vision-language models can help interpret detected events, answer natural-language questions about visible conditions, and give operators additional information to assess an alarm.

Video Search and Alarm Triage: How contextual analysis and natural-language search can help teams find relevant footage and review alarms more efficiently and where verification and human judgment remain essential.

Agentic AI: Practical Automation or AI-Washing? What makes a system genuinely agentic, how it differs from detection, contextual analysis, and predefined workflows, and why autonomy in security raises questions about reliability, accountability, and control.

Deployment That Fits Your Security Operations. What to consider when integrating detection and VLM capabilities with existing cameras, VMS, and workflows, including deployment flexibility, data sovereignty, and human oversight.

Live Technology Demonstrations. See examples of Scylla’s detection capabilities and VLM-assisted contextual analysis, illustrating how these technologies can work together to support incident verification and operator decisions.

Speakers

Albert Stepanyan

Co-Founder and CEO at Scylla

Armen Ghambaryan, Ph.D.

Lead Machine Learning Engineer, Scylla AI