 Hey, hey folks, can you hear me? Yeah? We're excited to be here today As she said we are an AI company and there's never been a better time to talk about AI than today Maybe just to get started. How many guys have tried Chagipiti in the past year? Okay, it's quite a big number and it's just about to get bigger We are about to see like billions of people transitioning away from Traditional pointing click interface to a natural language interface That means you type in something you hit enter you wait for a response Repeat a cycle until you get what you need Think about billions of users using conversational AI every single day That will make for trillions of interactions every single year And so you're gonna end up with a lot of unstructured data a lot of text and that might feel very messy very overwhelming But in reality, this is a gold mine of information if you just know how to look at it So the question is how do we find out what the heck is going on in these conversations? to start with What are we users doing? Are they trying to generate a LinkedIn post? Are they trying to summarize a document or perhaps just testing out some code? We don't know Also, what are our users most common questions like help me with my own one next one on one with my boss? Or perhaps You know draft an email for a new client. I get no idea Finally and most importantly are our users satisfied with the alarm experience Are they having a pleasant experience or is it a waste of time? And if it's a waste of time, why is the user experience so bad? Answering these questions at scale is not easy There were a few ways to get started pretty good ones But none of this is allowing you to really understand what's happening in the conversations. Let me tell you why First thing you might want to manually analyze every single user conversation one by one And that's obviously like a great starting point when you're at the start You've got a few beta users. You're gonna get a lot of insights, but it's impossible to scale It's impossible to scale because it would be too time-consuming once you have thousands of users Next you might be thinking what about standard product analytics, you know checking often people again How long does they look them for where they click and so on and so forth? Again, that's a good starting point, but it's not good enough because it's on a scratch on the surface You get a sense of what the user is Doing, but you don't know what the topic and you don't know what the intent Finally, you might be tempted into relying on explicit feedback from the users We have all heard about thumbs up thumbs down and PS scores one to five one to ten whatever The reality is very few of you have ever taken the time to provide feedback Certainly not me and I would bet most of you don't do it So this is you know in theory a great way to get started, but in practice you don't get enough Today I am very happy to introduce you to Nebulae which is the ultimate platform to automatically understand what your users are doing Let me tell you how we operate in three steps First thing first segmentation every time you pick a conversation. It's made up of multiple segments That's where the user intent changes. So the first thing we do is putting it up This is an example. This is just one user. It's one conversation, but the user wants to do three different things So these are three different segments and so as a result of that we are taking the conversation as input and splitting it up in three different pieces Once the segmentation process is over, it's time for analysis That means drilling down on each user segment to find out what's going on And whether it's a one-it-wonder or an extended exchange. We want to automatically Understand as much as possible what the user is trying to do That means automatically understanding what's the intent of the user, what's the topic of the conversation and whether the user satisfied or not So let's look at an example together. We've got the sorry just going back Our first example here is a person asking for some information And it's a great experience because she gets exactly what she needs right away one prompt one good answer Second example is a little bit different She's looking to generate a LinkedIn post and this time around as you can see there is a lot of back and forth The user is trying to get where she wants, but it takes some effort. So that's an okay conversation It's not smooth. So it's not great, but it gets to the point The final example the user is looking for information again But this time around she just can't get there. She tries multiple times, but she never gets a good answer So that makes for a bad user experience Once we have kind of identified what's the profile for each individual user It's time for the third and final step in the process, which is synthesis This time around the idea is aggregating insights across thousands of different of users So to be able to answer the questions that we were Asking at the beginning of the conversation So for instance now we know how many people are satisfied how many people are dissatisfied with the experience and why that's the case We do know what are the most common queers from the users. We do know how to make data land better and much more In other words, thanks to Nebulae What used to be like a messy unstructured amount of data now makes sense and You basically know what the hack is going on We help you Improving your alarm product kind of making it better and making it more user-friendly and making it such that you can improve your user experience so That's the product. It's live So I'd be very happy for any of you guys to come and visit at the booth a seven just around the corner Product is live very happy for you to give it a spin Especially if you do use it alums or conversational AI or charge a pity we just over there and I'll see you later