AI innovation is moving exponentially. Having IP cameras is a thing of the past, contemporary businesses are actively moving to advanced AI solutions. Our emerging technology, AI Copilot, is your AI assistant for cameras, working with you to revamp the security, safety, and efficiency of your business. AI Copilot immerses itself in your physical environment, actively indexing it and getting a holistic understanding of your space, regardless of your industry. The more you use and coach AI Copilot, the smarter it gets.
This KB article is the first of three in the AI Copilot series. Click on the following links to explore other articles in the series:
AI Copilot understands entire scenes, enabling new use cases tailored to your business. It will show you everything it understands: a map of all objects in the scene.
How it works
To use AI Copilot for the first time, simply click on the AI Copilot button on Live or Search to see through its eyes, hovering over labels to see recognized objects in space. Then, click on any discovered object to easily gain any intelligent insight.
If you’d like to see what you can accomplish with AI Copilot, go to the Point and Click Workflows KB article.
AI Copilot is poised to address a broad set of environments and use cases. However, in specific instances, AI Copilot may misinterpret parts of the scene. For example, AI Copilot has the most nuanced understanding of common scenes, such as office spaces, parking lots, hallways, and others. However, just as humans enhance their own understanding of new objects and scenarios through coaching, AI Copilot benefits from additional feedback through coaching as well.
In every AI Copilot enabled camera, by design, AI Copilot will invite you to coach it and give it feedback. The more you use AI Copilot, the smarter it gets. This is covered in our KB article, Coach your AI.
Beneath the surface, AI Copilot is grounded in avant-garde foundational vision models, akin to the prowess of ChatGPT.
Here is a comparison between traditional AI and AI Copilot:
Traditional AI is single-faceted and bottom-up, recognizing singular entities like people or vehicles and nothing else. These models are laser focused into ensuring performance on a use case, resulting in an accurate application but one highly limited in scope.
In contrast, Spot AI’s emerging technology within AI Copilot synthesizes and catalogs top-down understanding from images, videos, and textual sources across hundreds of domains. It is multi-faceted, recognizing not only people and vehicles, but a spectrum of arbitrary, granular objects. This conceptual understanding allows AI Copilot to be applied to any new domain, simply through training from everyday human interaction and teaching.
It employs state-of-the-art transformer architectures, benefiting from large-scale multimodal training across a variety of human inputs like text, images, and videos. Through these methods, AI Copilot can intelligently pay closer attention to more relevant parts of images in relation to its internal learning and training.
Integrating Traditional AI and AI Copilot
Your IVRs are intelligently integrating both technologies from Traditional AI and AI Copilot to transform your IP cameras into the smartest AI cameras ever. This integration of the two in AI Copilot facilitates a comprehensive grasp of any scene and any object it has encountered before, allowing it to refine its spatial temporal-reasoning. The result is a platform with profound comprehension and unmatched adaptability tailored to your specific business imperatives.
This feature is available to all users with a Pro license, and will not be available on the Core license package. More information on the licensing packages is available here. While all users will have access to the Base Object classifications, vertical specific classifications require an industry be specified within the organization settings.
Continue understanding AI Copilot by going to the next KB Article: Point and Click Workflows
If you have any questions, please reach out to your Customer Success Manager or email us at email@example.com.
We hope this article was useful to you, please leave us a comment or feedback as it will help us improve our customer support center.