A simple ai agent so you can learn…it’s not as hard as some make it out to be and doesn’t have to be intimidating either #aiagent

Let's build an AI agent together. A lot of people think it's really complicated, but it's actually not, so let's get into it. The first thing I'm going to do is create a project within Azure AI Foundry. You're going to see it's really easy to do, and I'm going to name it LinkedIn Generator because I actually want an agent to create content on LinkedIn for me, that way I don't have to do it. Resource here is going to take a little bit of time to load, so just be patient, and then now we can jump into the agent playground. Before we can actually build an agent, though, we have to deploy a model. So I'm going to use GPT for Omini because I know it's fast and it won't cost too much money either, which is obviously something important. At least to me, I want it to be cost-effective. We can just deploy the model. This part is pretty easy, and once the model is deployed, we can actually ask the out-of-the-box agent what it actually does, and this is going to give us a pretty standard example, like you asking ChatGPT, obviously, like, what do you do? But this agent isn't supposed to be your standard assistant. I want it to actually do an action for me and be a LinkedIn agent, so what we're going to do is actually give it those instructions. The cool thing here is that you can see, like, now the agent understands what its task actually is. You can actually see the thread itself, how many tokens, and how fast it actually took to respond to my query. The really cool thing here, though, is that we're going to be able to add knowledge, like a knowledge base of my previous LinkedIn post so that the agent actually understands how I like to write, what my writing style is, and add an action so that it can actually send me the email without me having to prompt it. It'll just automatically send me an email in the morning, so follow along for part two to build out this agent even further.