From the course: Agentic AI: A Framework for Planning and Execution

How agents are used in industry

From the course: Agentic AI: A Framework for Planning and Execution

Start my 1-month free trial

How agents are used in industry

“

- We've been exploring agents over the last few videos, and I hope you've been able to pick up some of the new concepts that make them exciting and powerful. In this video, I'm going to switch gears a little to explore a few real world use cases for agentic AI. I'll start with Covariant. They created a model called RFM-1. It's a robotics foundation model that provides agentic functionality giving robots the ability to reason much like humans do. They respond to their environment containing details, like text, video, measurements, or preset robot actions, and they can determine the correct action to do while also executing on it. So a human could interact with a robot by chatting with it with abstract queries like, "Are there any fruits in the bin?" or "Pick up all the red apples." And beyond chatting, agentic workflows are then possible with this robot model. Another example comes from Palo Alto Networks, who created an agentic AI, nicknamed Sheldon, to provide personalized employee support. So for example, here's a support request for IT, and I've anonymized it for privacy, where the person mentioned that they never received a keyboard. An agent understood the sentiment in the text and DM'd the person. This triggered a support ticket that got them a new keyboard. Another example is Akira AI, who use agents in a multi-agent payroll system that uses a human in the loop for supervision. But ultimately, it includes unique specialized agents that have specialized roles all overseen by an orchestrator or supervisor agent, which is controlled by a human. It's a young and growing ecosystem with more agents coming online all the time. But agents aren't for everyone, and in the next video, we'll go through a decision framework on when to use and when not to use agents.

Contents