Welcome to part two of our series on Unlocking the Power of Predictive AI, where we explore AgilePoint’s AI Control Tower. In part one, we introduced the AI Control Tower and discussed its Predictive AI Agents for binary, multi-class, and numeric predictive models.
In this post, we’ll dive deeper into Predictive AI Agents for Image Classification within the AI Control Tower. Like predictive models, image classification agents follow a similar framework, with four key agent types:
- Runtime Flow Adaptation: Modifies workflows based on predictions.
- Sub-process Initiation: Automatically triggers sub-processes.
- Notification: Alerts users when specific conditions are met.
- Basic Image Classification: Simply performs classification without taking further actions.
Key Differences for Image Classification Agents
Although these agents are conceptually similar to those for predictive AI models, there are three main differences that set them apart:
- Image Input Configuration: Unlike predictive models, image classification agents require an image as input. This means you’ll need to configure the image input source during setup.
- Single Image Processing: While predictive AI agents can handle multiple configurations, image classification agents are limited to processing one configuration and one image at a time. However, stay tuned—bulk image classification is coming soon!
- Varied Responses: The responses from image classification agents can vary significantly and hence a lot more complex and the configuration is more technical in nature. Unlike the structured and predictable responses from predictive AI models, image classification responses depend on factors like how the model was trained, the development process, and the AI framework used.
Learn More in the Video
To get a detailed walkthrough of these differences, watch the following video, which covers these three points in depth. If you haven’t seen the first video on Predictive AI Agents for Predictive Models, I recommend watching that one first.
Please note that these videos are meant for AgilePoint app designers who wish to go deep into the concept of AI Control Tower and their configuration. If you are interested in getting just a high level overview of Image Classification AI Control Tower, please click here.
Design Time – Configure Image Classification AI Control Tower
Run Time – Image Classification AI Control Tower in action