Azure AI Making AI real for your business-Applied AI Demo

possible. But it's one thing to talk about these services, it's something else entirely to just sort of see them in action. So I'd like to invite Noelle out on stage to come in and walk through in a little bit more detail exactly how these things work. Noelle?

Hi, thank you very much, Eric. All right, good, so exciting. Hi everybody, thank you so much for being here. I am really excited to share with you some of the things you can do with Cognitive Services and some prebuilt models. Prebuilt machine learning models. The first thing I wanted to draw your attention to, and actually I've been a developer, an enterprise developer for about 20 years and it's just, it's incredible what you could do today that you couldn't do even just a few years ago. So if you take a look at our website, you can actually go out to aka.ms/cognitiveservices and it'll take you right here and you can test some of these prebuilt models right inside the browser. So I just wanted to show it to you. Really quickly, if you take a look, you can just scroll down and see right now it's looking at an image and it provides these tags. This is a web-based service, so you don't have to be a data scientist to get access to this. You just pass it in the URL, you pass it an image, and you can see I just grab another image and it provides a different set of tags and information about that image. Now what happens though when you want to actually do something even more with this information and leverage this model by customizing it? Well, I thought I would do just that. So I started, I actually thought about this project. It's a, you know, we're in Orlando, like Mickey lives here. I thought it'd be cool if I created something that I struggled with, right? How do you tell the difference between Chip and Dale, right? The Chipmunks. How do you even know? Anyone know, by the way? No? All right, good. You'll see. So this, I decided to build an object detection model based on, Yeah, their noses. It turns out Chip has a chocolate chip black nose and Dale has a brown nose. So why not build a model that can detect the two? So I added about 20 images to my model, different tags for each, you know, a tag for Chip and a tag for Dale. I then trained the model and got some performance data on how it works. And then once I got that performance data, right, then I could go and actually just do a quick test. I'm gonna pass in an image and I just wanted to make sure I grabbed an image. I used Bing Search, just grabbed an image that wasn't already in my dataset and you can see, right? It pulls up. Now again, as a, a developer, I'm not a, you know, data scientist in any way. It was really cool that I could do this in a couple of minutes. I literally built this in less than an hour. Then of course I got addicted to it and started messing around with the model, which you will too, I hope. Now, this is incredibly interesting to be able to do as a developer, but there's another use case that every business on the planet needs to take advantage of, and that's the ability to build a conversational bot, and you as a developer can do it in just a couple minutes. So I'm, what I'm gonna do is actually leverage an, an application. You can go here as well. There's a, a link called aka.ms/intelligentkioskdemo, all one word, and you can actually pull up this demo and build this yourself if you want to, but I'm going to leverage a Q&A explorer to build a bot in just a couple minutes. Okay, one minute. So I'm going to type in FAQ. Oh, you know what? I'm actually going to use Ignite. And then it's going to ask me what landing page do you want to demo this on, right? So I'm going to go over and grab the Ignite page and grab the Ignite page and put it in. And then I'm going to pick a knowledge base. Where is the Q&A going to come from? So I'm going to actually pick a URL because it's easy, but I could also choose a TSV file or something else, you know, another Q&A Maker project. So I'm just going to grab this and paste it in and hit create. Now what's happening right now is that it's going through the process of grabbing all of the questions and answers and creating a dialogue for it. That you can then edit. We also have a new utility called Chit Chat that allows you to add a personality to this bot. But most importantly, it's creating a natural language model. So not only do I get to type in, of course, exactly what's in the FAQ, but I can type in variations to what's in the FAQ as well. You can imagine the types of things we can do, right? HR, finance, we can create Q&A bots in minutes for departments that need it the most. So you can see here, right, I've pulled up the site. Here's my little bot right here at the bottom. Just in a couple minutes, I can ask it things like, oh, what do we always ask? Like, what's the agenda? Right? And if I say, what's the agenda, it pops right up. But I could also say other things like, where's my agenda? How about that agenda? Right, so if I just type in, right, where, where's my agenda? That's not specifically in the FAQ on the website, but however I say it, it'll pop up and provide the answer. Yes. Right, so you can imagine, right, with object detection and the ability to check between good produce and bad produce or good silicon and bad silicon, right? Good hardware and bad. Object detection can help you solve this problem as a developer in minutes. And then with Q&A Maker, you can, you can solve for top-of-mind questions for your customers and your employees. So I hope you've seen just in a couple minutes how much fun you can have and that you'll go out and take a look at it soon. Thanks for having me, and I'll give it back to Eric.

Thanks, Noelle Russell Russell. Really great to see