Sentry AI is moving its physical security platform beyond AI-powered video monitoring with v11, a release that combines autonomous AI agents with human guards. The update introduces Agentic Patrol, sensor and access-control integrations, and Guard Dispatch, creating a coordinated security workflow designed to monitor properties continuously, investigate incidents and route physical responses when human intervention is required.
Sentry AI is positioning its latest software release as a shift from AI-assisted surveillance toward an agentic AI security workforce, combining autonomous monitoring with human guards rather than attempting to replace physical security personnel outright.
The company’s Sentry AI v11, announced September 15, introduces Agentic Patrol, broader integrations with access-control and sensor systems, and Guard Dispatch for coordinating human responses to AI-detected incidents. The platform is available now to new and existing customers, according to Sentry AI.
The distinction is important because conventional video analytics typically operate as detection systems: cameras observe activity, software identifies a potentially relevant event and an operator investigates the resulting alert.
Sentry AI wants its agents to take a more active role.
With Agentic Patrol, AI agents can proactively patrol properties, analyze activity and investigate potential threats instead of waiting for an alarm to trigger a response. The company’s broader platform combines video intelligence with access control, sensors and human guard operations.
That creates a model closer to an AI-enabled security operations center than a standalone computer-vision application.
From cameras to a connected security environment
The new release expands the data available to Sentry AI’s agents.
Rather than interpreting camera feeds in isolation, v11 can connect information from cameras, doors, gates and other sensors. The goal is to correlate multiple signals to establish a more complete picture of activity on a property.
For example, suspicious movement detected by a camera could potentially be evaluated alongside access-control events or activity at a gate. In an agentic architecture, that information can become part of an investigation rather than simply generating another alert for an operator to review.
The approach reflects a broader enterprise AI transition toward agents that can gather information, reason over multiple inputs and initiate predefined actions.
McKinsey’s 2026 research on agentic AI found that organizations expect adoption of AI agents to increase substantially over the next several years, while also identifying governance, identity and runtime controls as increasingly important as autonomous systems gain access to business environments.
Physical security presents a particularly consequential version of that problem because an AI system is not simply producing a digital answer. Its output can influence whether someone is dispatched to a physical location or whether an incident is escalated to authorities.
Guard Dispatch connects AI with people
Sentry AI’s answer is a human-in-the-loop operating model.
Guard Dispatch connects AI-detected incidents with human security personnel. New tracking and mapping functions provide operators with visibility into incident locations and guard positions, allowing them to identify and coordinate available personnel.
That effectively links three stages that are often separated across security technology stacks: detection, investigation and physical response.
Sentry AI’s own platform materials describe a similar workflow in which AI alerts are reviewed by remote security personnel before incidents are escalated to roaming guards or other responders.
The model also addresses a practical limitation of autonomous security systems. AI can monitor large numbers of cameras continuously, but physical intervention still requires people when an incident involves confrontation, public interaction, emergency services or other circumstances requiring judgment and presence.
Sentry AI CEO Uday Kiran Chaka describes the division of labor as continuous AI awareness and response coordination combined with human judgment and physical presence.
The economics of an AI-human security workforce
Sentry AI is also making an aggressive economic claim around the model.
The company says the combined AI and human approach can deliver security at roughly one-fifth the cost of on-site guards, while avoiding replacement of existing hardware and software infrastructure. It also claims that its agentic AI handles approximately 90% of the work of an on-site guard.
Those figures are Sentry AI’s claims and should not be interpreted as independently established industry benchmarks. Actual economics will depend on property type, camera coverage, guard staffing, response requirements, integration costs and local labor markets.
Still, the economics illustrate why physical security is becoming another potential application for agentic AI. Continuous monitoring is labor-intensive, while the number of cameras and sensors organizations can deploy continues to expand. An AI layer that can triage routine events could allow human personnel to concentrate on incidents requiring physical or contextual judgment.
The same architecture also introduces new requirements around reliability and governance.
McKinsey’s 2026 AI Trust Maturity research found that nearly two-thirds of organizations surveyed identified security and risk concerns as the leading barrier to fully scaling agentic AI. The study also found that 74% considered inaccuracy a highly relevant AI risk and 72% cited cybersecurity.
For physical security, those concerns translate directly into operational questions: What events can an AI agent investigate autonomously? What confidence threshold should trigger escalation? Which actions require human approval? How are decisions recorded? And what happens when camera, sensor or access-control data conflicts?
Sentry AI’s architecture leaves humans responsible for the physical response, but the increasing autonomy of the software makes those controls increasingly important.
A broader agentic AI category emerges
Sentry AI’s v11 also illustrates how agentic AI is spreading beyond conventional enterprise software.
In cybersecurity, IT operations and business applications, AI agents increasingly act on data and software systems. Physical security adds the real world as another operating environment, with agents interpreting sensor data and coordinating people who can intervene physically.
That makes integration a central part of the platform strategy.
Rather than asking organizations to replace cameras, access-control systems or other infrastructure, Sentry AI says v11 works with existing security environments. This can lower the barrier to deployment because customers can add an intelligence layer without rebuilding their physical security stack.
The competitive landscape includes AI video analytics providers, security operations platforms and autonomous security companies. Vendors such as Knightscope are also combining AI-driven monitoring with human security personnel and field operations, demonstrating that the market is moving toward hybrid security models rather than a simple AI-versus-human choice.
For Sentry AI, the next stage is therefore not simply better object detection or more accurate cameras. It is creating an operational layer that can continuously observe, investigate, coordinate and escalate.
That changes the definition of an AI security system from software that watches cameras to software that participates in the security operation.
The key test will be whether those agents can maintain accuracy, explainability and reliable escalation as they operate across increasingly complex environments. In physical security, where an incorrect decision can have immediate real-world consequences, the value of agentic automation will ultimately depend as much on controlled human intervention as on autonomous detection.
Market Landscape
Physical security is increasingly converging with AI agents, computer vision, IoT sensors and remote security operations. Sentry AI’s approach sits between traditional guarding and fully autonomous security: AI handles continuous monitoring and investigation while human personnel provide judgment and physical intervention.
The broader agentic AI market is also moving toward governed autonomy. McKinsey projects that agentic AI could account for about 15% of enterprise cybersecurity budgets within three years, up from roughly 4%, as organizations invest in identity, governance and controls for autonomous systems.
For physical security providers, this creates an opportunity to consolidate cameras, access control, sensors, remote monitoring and guard dispatch into a single AI-coordinated workflow.
Top Insights
- Sentry AI v11 adds Agentic Patrol, enabling AI agents to proactively monitor properties and investigate potential security threats.
- Guard Dispatch connects AI-detected incidents with human guards, combining autonomous monitoring with physical response capabilities.
- New integrations correlate camera, access-control and sensor data to provide agents with broader real-time situational context.
- Sentry AI claims its AI handles about 90% of an on-site guard’s work, but that figure remains a company-reported metric.
- Agentic security introduces new requirements for human oversight, runtime controls, reliability and auditable response workflows.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
