Iveda Solutions, Inc. (IVDA Quick QuoteIVDA - Free Report) has announced a major enhancement to its IvedaAI platform with the introduction of real-time zero-shot AI detection. The new capability allows users to instantly create and deploy custom AI detection models through simple natural language prompts.

The technology enables retailers and security operators to activate sophisticated video analytics simply by typing a word or phrase such as "shoplifting," "suspicious behavior," or "graffiti." Once a prompt is submitted, the platform immediately generates a tailored AI model and begins analyzing live or recorded video feeds.

Eliminating Traditional AI Training Requirements

The launch marks a significant shift in how AI-powered video surveillance systems are developed and deployed. Traditionally, organizations seeking to create detection models needed to gather large datasets, label thousands of images, train algorithms over extended periods, and invest substantial technical resources before deployment.

Iveda's new approach compresses that entire process into seconds. Instead of relying on pre-built training datasets and prolonged development cycles, the platform converts natural language instructions into active AI detection models that can be applied instantly to existing camera networks.

Per David Ly, founder and CEO of Iveda, the advancement represents a major step forward in real-time AI video analytics. He noted that users can now deploy specialized detection capabilities immediately, without waiting for data collection or model training, enabling organizations to respond more quickly to emerging security challenges.

Retail Loss Prevention Emerges as Key Use Case

One of the primary applications of the technology is retail loss prevention, particularly the detection of shoplifting. The company noted that shoplifting has historically been difficult for traditional AI systems to identify reliably because it can take many forms across different store environments and involve a wide range of concealment methods.

Conventional AI models often struggle to account for the countless ways theft-related behavior can occur. Building comprehensive training datasets capable of capturing every possible scenario has proven difficult, limiting the effectiveness of many automated surveillance solutions.

IvedaAI seeks to overcome this challenge through its zero-shot detection framework. By combining advanced Vision Language Models with the company's existing portfolio of pre-trained object detection technologies, the platform is designed to understand context and intent rather than relying solely on object recognition. This enables the system to identify behavioral patterns and activities that may indicate theft or other security concerns.

The company disclosed that the technology is already being evaluated by one of the world's largest fast-fashion retailers, which operates thousands of stores across more than 90 countries, as part of efforts to strengthen loss prevention initiatives on a global scale.

Early Testing Highlights Behavioral Analysis Capabilities

Initial testing has demonstrated the platform's ability to identify activities that may signal security incidents before they occur. During evaluations, prompts related to suspicious behavior successfully highlighted individuals looking through windows and repeatedly examining staff-only entry points.

According to the company, these results illustrate the platform's ability to recognize intent-driven behavior and pre-incident activity that might otherwise go unnoticed by traditional analytics systems. Rather than focusing exclusively on static objects within a video frame, the technology evaluates actions and contextual cues to generate more meaningful insights.

The system can also be deployed to identify a range of security-related events and behaviors, including shoplifting, loitering in sensitive locations, vandalism-related activity, graffiti attempts, smoking in restricted areas, and aggressive conduct. All detections become available immediately after a prompt is entered into the existing IvedaAI dashboard, without requiring additional infrastructure or software environments.

Flexible Cloud & On-Premise Deployment Options

To accommodate different operational and security requirements, Iveda is offering the technology through two deployment pathways.

Organizations seeking rapid implementation can utilize a cloud-connected configuration that leverages large language model processing for live frame analysis without requiring substantial hardware investments. This option is intended for customers looking to activate the technology quickly and efficiently.

For enterprises operating within highly secure environments, the company also offers a local deployment option powered by its proprietary Cosmos-Reason engine. This on-premise configuration supports fully local processing within closed-network environments, eliminating the reliance on cloud connectivity while maintaining a secure architecture with no open ports.

Iveda also announced plans to begin shipping its next-generation Cosmos-Reason-2 engine during the coming quarter. The upgraded platform is expected to provide enhanced inference performance and improved detection accuracy.

Built on More Than a Decade of AI Development

The newly introduced capability builds upon more than 10 years of AI research and product development by the company. During that period, Iveda developed and refined a variety of object detection technologies, including weapons recognition, facial recognition and license plate identification.

The addition of Vision Language Model reasoning extends these capabilities by enabling more sophisticated interpretation of complex environments and behaviors. According to the company, the technology is designed to maintain performance in challenging real-world conditions, including low-light environments, crowded locations and dynamic retail settings.

Iveda has stated that it continues to integrate emerging AI frameworks, including technologies from NVIDIA, as part of its broader strategy to keep the platform aligned with ongoing advancements in artificial intelligence and machine learning.

Expanding the Future of Video Analytics

The introduction of real-time zero-shot AI detection represents a new approach to deploying video analytics, allowing organizations to create custom surveillance capabilities using simple natural language instructions. By removing many of the technical barriers associated with traditional AI model development, the technology aims to make advanced video intelligence more accessible and adaptable for retailers and security operators.

As organizations increasingly seek flexible and scalable AI solutions, Iveda's latest enhancement positions the company to address evolving surveillance and operational requirements through real-time, context-aware video analysis that can be activated in seconds.