Hello. Reinvents. How are you guys doing? Amazing. That's a great answer after lunch too. I uh uh have spoken at a conference before and uh it, it's always amazing to me. Uh how different it is. Uh e every time. Um you know, i, i've been doing this for a long time, different kinds of companies and, uh you know, each has its own flavor and its own flair. Uh but one of the things that i really love about aws is the closeness that we all get to have. And i think we have a lot of very direct interactions and communications. And so one of the reasons that i like coming to re invent, uh it's one of my, one of my favorite weeks is because i get to talk uh to each and every person individually and hear their stories and get, get deep with them. And i, and i wish the same for you. I hope you're having uh a similar kind of a time.

Um uh would you advance our slide one so we can talk about what we're gonna talk about? So, uh today we're gonna talk about generative b i capabilities that are coming to amazon q in quicksight. And um do we have some technical problems over here? Uh yeah. Ok. We're making an adjustment. Uh and my name is zach. Uh i lead product for a i and ml at quicksight and uh i've been building software professionally for uh 24 years. Uh companies like salesforce tableau microsoft start up work. Uh and uh i love this work because i believe that data has this kind of almost magical power to bring us to see things together. And that if we can see the same thing, then we can have a discussion about it, maybe we can make the world just a little bit better place. Uh and i am very excited to be joined on this stage today with a couple of wonderful customers and very data minded people. Uh who, who would you like to say? hello, ed?

Sure. Um hi, everyone. My name is ed. I work at uh godaddy in analytics and i've been around software for similar time period. Um and really have grown to appreciate also the value of, of really how data does bring us together and, and really, um you know, help us have really interesting conversations with it. So um looking forward to having a chat with you today.

Great. And my name is emily revello. I am an analytics manager for huron's care insights team. Um i have a background in math and secondary education um and passionate about helping people solve their problems using data and analytics and quicksight. So, probably gonna teach us some today.

Yes. All right, you too. Uh thanks so much for, for coming out today. We're gonna hear more uh from, from these two in a little bit.

Uh so, so we've got all of this data, right? Um uh who has been gathering data? Yes. Yes. Ok. Some of us have data. Ok? And it's uh potentially very powerful, right? Like we've got all of these ideas about what we can do with the data, but it doesn't always turn out to be as effective as we want it to be. And so many of the customers that i talk to say, you know, it's not getting me the decision input that i wanted. And so some, some question for you, what do you think it is that prevents data driven decision making? Given that we just have these piles of data. Anybody have an idea?

Sorry. Uh socialization. Yeah. People don't know about the data or they don't believe in the data. Yeah. Yes. The lack of good people don't understand it. Uh you know, like the metadata isn't there? Yes. Oh i love that one. Not framing the question in a way that the data can answer it, meaning like they have a business problem that they're trying to solve and then maybe like the data doesn't exist or you know, isn't it yet in the right form? Yes. Was there. Another one. Politics. Yes. Well, there are politics. Yes, one more. Not knowing, actionable next steps. Ok. So i've got all of this data here. You know, there's some finding but you know, what am i going to do with it? That's great. I love that.

Uh so gartner, you know, i think a more general sense says that data skills and staff shortage is the number one impediment for organizations success. And, and they say by the way, uh you know, a lot of the problem is that this time is, is not being used. Well, so 68% of it, you know, is just kind of either undifferentiated work or, or things that analysts are doing. Um and then, you know, for business users, well, 75% of the time they don't have the confidence, they don't understand that they don't believe it's answering their question, they don't know how to use it. So, um so i look at this and i actually see instead a software problem, can the software tools make this easier? And so that's why we created amazon quick site.

Quicksight is amazon's general purpose b i service and uh it works with any kind of data. Uh it includes dashboarding capabilities, reporting all the things that you would want out of a typical b i service. And as you would expect from amazon, it scales to any performance level. So you don't really have to think about or manage performance characteristics, it's augmented and this, of course, is my area, my passion with a i through and through and this i think makes it a little bit different from other products i've worked on at other companies and then also uh it ends up being pretty efficacious. And so i think with a lot of amazon services, you pay for what you use kind of thing. Um you know, there, there's nothing that you're pre buying, you know, licenses to manage or all of that stuff, you just, you know, put users in, you take users out. So uh a lot of people using it today.

Some great customers on this list that i've actually had additional conversations with this week who are floating around, you may be able to talk to directly. Um the dashboarding capabilities are pretty robust and complete uh data visualization and interactivity, all of those sort of things that you would know and you would expect from a dashboarding platform. And then a lot of people, what they're doing is they're actually building applications. And so i was talking to some folks uh just a minute ago, a little bit earlier uh that are creating applications, analytical applications or have data that we want to make available as a service to other people. And that is uh i would say actually a bigger part of what people are doing with quicksight even than just the core analytics uh experiences inside their organizations last year.

At re invent, we announced that we have paginated reports. So these are great for people who have been, you know, for many years working on a, a sort of legacy system on prem, you know, they were looking for a cloud solution. Now, it's an easy way, put all your b i one service in order to sort of be able to reuse all the data and um uh sort of business logic that you've encoded.

Uh so my area in a i uh one of our key capabilities is what we've been uh calling quicksight, q and quicksight q was launched originally uh back in 2020 so quicksight q enables business users to ask and answer questions of data. And last year at re invent, we announced that we were bringing new natural language capabilities for business users who are not technical to be able to do some more advanced things. So they can do things like get a forecast. Ask a why question, why were sales down in february of 2023 and automatically get an analytical response.

So cues in the product, all of this uh sort of a i revolution is happening around us. I mean, what a year, how many of you have active a i initiatives in one form or another? Yes. Ok. That's like 75% of the room. That's, that's amazing. And um you know, when i took this job, i was thinking, all right, are we you know, this natural language thing is it's been really hard and you know, are we really, are we gonna make the, the run at it? And so i had this conversation with the folks and i said, you know, can, what makes us believe? Like, can we get there in the end? What brought me was this belief based on our approach, the background, the things that we had that we could bring to bear that we could actually build something that would solve this natural language problem, making it easy for people, which has, has been super challenging. And along comes these sort of generative large language foundational models and they bring a whole new tint to it.

So now we've got our inbuilt models in quicksight that know how to do language understanding and they know how to map your intense onto your data. We've got forecasts and outlier detection and other analytical capabilities in built a quick site that we can leverage and we've got these new large language models. So the question is, what can we do with all of this stuff? These are a is, don't they look cute and they look helpful. The question is though, what can they really do?

So it turns out after examining these things, i came to the conclusion that uh these are really best for assisting our work at this time. So these large language models are uh effectively under the covers predictors. Um you know, they they forward predict what text is going to be. And uh we, we know that they are trained on variety of like very broad uh world information. So they have a lot of conceptual ideas that they can get into and they have a lot of power that they can bring to bear in terms of showing us new things, new ideas that we haven't thought about.

So they can su information for us, they can do that kind of work. They can sort of teach us specific skills in specific areas where we're not an expert. And then they can do this almost magical sort of set of generalization. So they can say, oh, you know, i think this may be going on based on what's happening. And uh and that, that's actually fascinating because it can bring us some new ideas.

However, they are not the perfect solution for every problem. So we do not want to use them to replace human judgment. Human judgment is still the primary key for making decisions. We do not really want them doing math at this point, that's an evolving space. We know there's a lot of good sort of technology that's coming in in this area. But at this point, there are better ways to do math.

There are systems like Quicksight that already produce statistics and, and sort of BI analytics reliably uh that for example, are um things that you can sort of depend on. There's also uh this sort of well-defined task. So if you have a machine learned model that does forecast, that you've trained, that is probably gonna be better at forecasting than a large language model. You can ask them to forecast, but they aren't trained as specifically and just like a person who's gone to school for a specific kind of degree program, they're going to be better at that task.

Um and then finally, there is a step of verification. So before we send things out into the world, we want to be sure that they're true or at least to our human judgment. So we've announced Amazon Q with generative BI capabilities. So Amazon Q is the umbrella assistant offered by AWS that can help you to complete work in all different parts of your AWS experience. And in Quicksight specifically, it provides and AI accelerated dashboard authoring experience that helps business analysts build dashboards faster.

It provides AI answers to questions of data on demands for business users to be able to get the insights, they need to drive a decision and then it has the ability to do. And I think this is really interesting and I hope you do too when you see it some storytelling. So the the AI can actually assist business users in examining their data and producing stories that are helpful than to explain to other people what might be going on in order to drive a decision.

So the other experience has three parts. You can build visuals, you can build calculations and then you can refine the visuals in order to achieve a precise outcome. The AI answers to questions of data on demand comes in two parts of the experience. So first you can ask Q to summarize your dashboards and it will automatically extract key insights from the dashboards and make them easy to see.

Second, there is an enhanced Q and A experience. We've rethought how Q and A can be and we've created something that solves common business user problems by suggesting questions and providing a what's new experience that lets them know kind of what data they can ask about to make that obvious a multi visual answer which breaks down all of their data to give context and then explains with narratives what's going on in the data so that they can confidently understand the data.

And then finally, we support vague questions and these are the types of questions that business users frequently ask and you know how, how many or or how much or why kinds of questions where many of the details are missing that would be required in order to formulate something, you know, as complicated as a SQL query, which many of us can write, but many of them don't know about. And just in case we interpreted it wrong, we also offer alternatives.

So we tell them, hey, there could be multiple parts of your data that, that match to what you're asking for, for stories. There first is the story capability. So story is a brand new kind of thing. We're bringing a QuickSite, it is not a dashboard, it is not a report, it's more like a document or slides. And so you can build these stories using the AI can introspect on your real actual data, it can build narratives, it can explain what's in the data and it can draw conclusions and make suggestions.

Then you can use the AI to refine the content that you've created. And then because these stories live within your BI service, they participate in your existing governance model.

Uh but I'll, I'll give a quick uh intro to myself. My name is Ed. I lead some um data analytics teams at, at GoDaddy. My uh uh colleague Mayer helped me put together a story today and I wanted to have a little bit of background and context um for uh quick side uh GoDaddy.

Um so Richard Hand, I guess real quick if you guys uh have heard of GoDaddy and, and know that we're not only uh domains website company, oh, a little bit fewer than I thought. Ok, great. We'll learn something together.

So, what I'm showing here is, is a little bit of a kind of a snapshot of some numbers of the business. Uh and while everyone recognizes us as a player in domain names, um the impact that we have is, is far beyond that. So we have, you know, over 20 million what we call everyday entrepreneurs. Um these are your local florists, um your pickleball paddle manufacturers, um your favorites, craft vendors. Uh and we support, of course, the, you know, the 84 plus million domains that they, that they register.

Uh but beyond that, there's also this commerce business, right? So this thing called gross merchandise value of about 33 billion uh which is the value of goods and services transacted, right? Um by our customers on our platforms. So our commitment goes just beyond, you know, beyond um domains, but also websites and commerce. We even have a uh payments business with branded hardware, right? So really our strategy go at it is kind of simple and it's really underpinned by data. And that's sort of what is I think interesting for us here.

Um and so we aim to uh you know, champion every single entrepreneurs, every single uh sort of micro segment and offer them the guidance they need and provide data uh to help our customers succeed. Um and we're not just providing those as sort of tools but really creating kind of an ecosystem uh where they can thrive.

And so with that, our vision is really to see a world where the entrepreneur really is at the center um of the global economy, right? And so we, we think of the entrepreneur and the data surrounding their journey as sort of a a wheel. Uh and if you can think about, you know, kind of all the complex data, uh if you interview how small businesses and what that looks like, you can kind of maybe um relate to that.

So from establishing strong digital identity uh with essentials like domains and logos to uh online ubiquitous presence um with the hosting websites, um you might, you know, know about those things, but what we might not know is there's also an area that we're diving deep into, which is called connected commerce and it's tapping into the bustling world marketplaces.

Um whether it's uh bleeding into the digital and the physical uh with online offline payments, um and, and just streamlining payments in general. And so we're really trying to be in every stride of the entrepreneurial journey. And so you can imagine if you're running a business like that, all of the complex permutations of the data being produced in each step of the way. And that's really the true immense problem that we are trying to tackle.

So data is great, but it's really insights, right? Uh that supports our customers success, right? And so we do that through many things, including our, our care guides. Um and so we took a look and we asked our, our analysts community uh about the pain points and how do you reduce the time to insights, let's say, delivered through care or any of our other uh analytics teams.

And so the survey came out with a few answers around, you know, various um levels of uh you know, kind of time that it would take long poles if you will uh of bringing time uh the time insights down, right? And so you can see that exploring the data might be a day or so, preparing the data might actually take a week or so, um visualizing it and preparing a interesting story which we'll talk about might take about a week, um maybe validating this a little bit shorter and then deploying might take another week, right?

So you can see sort of that's a lot of time being uh being spent in trying to bring an insight right to our business. Um meanwhile, uh as you guys probably can relate to, there's a bunch of changes, people are slicing the data slightly differently. One bu is saying one business unit is saying, hey, can you do this to it? And meanwhile, there's uh you know, adding fields and these kind of things, multiple round trips, right? You probably have experienced this in your analytics is multiple round trips of of changing the the the slice or the view of the data.

And so with that in mind, what has been sort of our our data journey or how is it, how we evolved the platform to sort of try to address that? So starting in kind of the early two thousands, you know, we like many of you, we started with relational databases uh built largely on the microsoft stack. And then moving into 2011, we embraced the massively parallel processing uh and then also incorporated terra data of course.

And then actually ala elation uh data governance quite early and then tableau for di visualization. And then the period for about 2013 to 2022 about the past uh 10 years, uh we embraced, of course, big data and things like pig hive, even red shift hadoop and spark um you know, were there, right?

And then we just last year moved uh really to the modernized aws cloud. As many of you are on and really introducing things like data, a quick uh quick side and athena and those things. And so now what our challenge is, is really to say, how do we again reduce that time to insight um with this new, these new capabilities while not sacrificing things like governance and the and the things that we like to see in our, in the correctness of data and and the usage of it.

So going back to those, you know, things we found out about time to insights, what we're hoping to see what we're, you know, beginning to see with aws stack is how we're revolutionizing that, that story for us um starting with exploring the data um using things like data brew and preparing the data with blue studio and data brew, providing sort of clean and organized data sets for folks to use and really what we'll talk about here in a second is the building our visualizations and really the insight stories, right? that, that i, you know, for, for me, it's been more fun and more more engaging with people um to use quick side and q combined with what you'll see in, in generative b i.

Uh and lastly, of course, deploying the production, we hope it's gonna be more seamless and with some process automation reducing some of those deployment times. So in essence, we're hoping that the a w stack sort of end to end sort of helps us uh not only accelerate insights but keeps good things like governance in place.

So before I dive in, let me just kind of mention to you what we think are the profound impacts that we'll hope to see um from enhancing our um visualization strategy uh with uh with quicksight. So, you know, firstly, we're hoping that we can um make a really game changing reduction in the number of ad hoc reports. Remember all those changes i was mentioning, uh we're hoping to eliminate some of those with some of the features, you know, zach showed.

Uh we see a lot of this happening i in our organization, you may have seen the same, but the other is um you know, really close to heart is we do a lot of i'll say manual uh anomaly detection of our business performance data, right? Every, every single stakeholder, all the way up to the CEO literally on a daily basis wants to see, you know, what is behind the anomalies in our performance in the business, right?

And there's some really neat insights uh capabilities that the tool has to be able to, to be able to help us there. I just had someone on my team yesterday call me and tell me about when can we do this. So I'm gonna try to walk you through uh uh some, some screenshots not quite as exciting as uh as zach's video. Uh I had to do some neat things with the data to be able to show this to you.

But I wanted to show you a little bit of sort of three personas that we see in our organization um using some of these features. So first in the analysis view, uh we have a care analyst who is preparing a dashboard, uh which breaks down what we call interaction. And these are these moments when a care guide is communicating with a customer through an interaction, what we call platforms such as voice chat or messaging.

And one of the uh you know, typical questions is right, which of these platforms is performing the best, right? From a revenue perspective. Um and this is usually when a customer is um you know, buying a new product or renewing uh a subscription or product when they're on the phone or, or interacting somehow with one of our um care guides.

And of course, you won't see this in, in a video, but you'll see this kind of on the side there. How easily it is to add elements to this simply by using natural language and then to sort of add that. So we've done things like use calculated fields and some of the capabilities that just showed to add, you know, new and recurring revenue um slicing by country, some trends and a couple of sanki diagrams down below.

Um and really adding the visuals, you know, is really as simple as, as we've just seen is as simple as just saying something like, you know, count of interaction id by platform per month and then really taking then a look at how um quicksight interpreted that uh and then reviewing that and it's essentially making sure that that was what you wanted in your analysis, right?

So in some ways, it's kind of another version of gen a i sort of prompt engineering if you will from a visualization side and then doing all the tuning you might need to do with uh anything else in the visualization that's appropriate.

So second, um you know, in the same view of the analysis, you have sort of the um the reader view or the the business stakeholder, um this person is consuming what was just done. Um but it was interesting here is, you know, this uh executive summary feature was enabled by the, by the analysts in the previous view.

And so we're sort of able to get this view of course, which i'm blurring out a bunch of, you know, key insights on, but essentially you're able to sort of see quickly, where should i dig in? Right. How many times have you looked at a dashboard and say like, what am i looking at? What should i do about this? What's next? Right.

And so there are some ways to sort of get to that more quickly that we see and by doing this, this this person can go in and pin this copy it and kind of share it out. And typically for us that might be a product manager or some other leader in the company who has a little more context around the data in terms of what investments we might make.

You know, the data might show that one platform is performing better than the other in revenue or something else. But there's something else underneath it, maybe it's not shown in the data that that person has to sort of uh give some insight into.

So finally, you know, this person looks at it and says, well, the one unan unanswered question that they thought was, well, this is great from a revenue perspective, but from the customer's outcome, uh i'm really not seeing everything, i'm not seeing the whole picture.

And so this person says, well, let me go out and add something else in this case, maybe net promoter score and without having to do another round trip, you know, with the analyst team, they can go ahead and, you know, ask a natural language question or put in a natural l language phrase like average nps score by platform without unattributed orders.

So even with a neat filter in there, uh and then get, you know something out which is helpful for her further analysis.

So finally, the third and final persona is someone that is preparing all of this for, let's say a business review, right? And so what we were seeing is in the, in the demo, of course, is someone that's interested in taking the insights from our team and putting it all together into a story.

And then really, as i was saying before, sort of having a conversation with everyone interacting with this both from a qualitative quantitative, you know, sort of ana uh analysis kind of perspective, but also with sort of the generated kind of creativity from uh from ll ms, right?

So you sort of see one from a strategy perspective. It's not always just about the numbers is always about, you know, sort of creative ideas and those things and you can sort of have that sort of assistance, you know, in the storytelling feature. And we're really, really anxious to see how that, how that pans out for us.

So you can see here, there's a, there's an insight that it might pop out that we might not have thought about in terms of a certain channel that might be, you know, driving better spending. And also there might be um other suggestions on what to do that might be again, more creative in terms of experimentation or things that we might be doing.

But there might be other ideas that, that um they're coming out out of the, out of this kind of storytelling um uh feature. So, oops, there we go. Anyway, i uh that's all i have.

Um you know, i'm really anxious to bring up uh emily up here and let you guys know that, you know, we've just started the journey uh from someone like emily who's been on the journey with quick, a little bit longer. We're looking forward to learning from more from her as well as any of you that have been on this journey with quick site and uh i'll hand it off to emily. Awesome.

Thanks, Ed and Zach. Um again, my name is Emily Arlo and I'm with Huron. Uh for those of you who don't know who Huron is. We are a consulting group out of Chicago, Illinois. Um but with lots of other offices throughout the US and internationally a little bit as well.

Um we have almost 65, 5000 employees and serve over 2000 different organizations every year. Um those span lots of different industries. Uh but today we're gonna focus on our health care industry, which is where me and my team sits.

Um so specifically I work on the Huron Intelligence team that has one of the, one of our data sets is a commercial and Medicare claims data set. Um it has deidentified patient information so I can't tell it's me Emily in the data set, but I can follow the patient's lifespan throughout the course of the data.

So I can see what I did back in 2018 all the way to yesterday when I saw my doctor. Um it has over 290 million patients in it for over 6.2 million doctors, physician assistants, nurse practitioners, any type of provider you would see for a healthcare need.

It spans the care continuum, uh which is important when we'll get into the dental. I have some live data we can share. Uh but spanning the care continuum, we can see ems and urgent care visits to an inpatient hospital, stay for a knee joint all the way to rehabilitation or skilled nurse, nursing facilities or hospice.

So, um you're probably thinking, what are you doing with this? What what types of questions can you answer? And the one we're gonna focus on today is the out migration patterns of patients.

So little personal story when I moved last year with my toddler, I called up my uh local provider and said, hey, I'd like to, you know, get him in for his pediatrician visit. He was nine months old. So if any of you have kids, you're getting 78 visits in that first year.

They told me great, we'll see you in 11 months and I said he needs three visits in that time. So, um the ty those are the types of questions we're trying to solve and answer and look at. Where are those access issues occurring? Is it happening for pediatricians? Is it happening for a specialty cardio visit?

That patients are driving three hours for trying to narrow down because that answer is different in every single market and for every single hospital. So that's one way we're using the data and I'll show a sample of that in a bit.

But where are we with our QuickSight adoption? So, like Ed said, um we have been using QuickSight for a little over four years, a little under four years now. Um initially, it started as some very ad hoc reports um just throwing something in a visual.

So my team wanting to have something in a dashboard instead of having several different pivot tables in an Excel file. Um it quickly expanded to res subscription reports when we a client who really liked what we were doing and wanted it on a regular basis, which then expanded into embedded where we're focusing today um and taking some of those very standardized visualizations that we've done over and over again and putting them inside of a software product.

So you can imagine that we have lots of different users who have lots of different needs and use cases from authors to readers to executive business users. So QuickSight has been helpful for that.

How it has helped our team uh specifically was that taking data from super super custom of everything is different and locked in a spreadsheet. And there's multiple different sources of truth and transforming it into integrated data experience.

So that being those embedded embedded reports or also just those subscription reports or um sample visuals that look about the same for every project. And we can tell our internal consumers kind of what they can expect when they buy data from our team.

Um, very, very, very simple data pipeline for what this looks like for us is on the bottom there. But, uh, we'll query data from multiple sources including Redshift, um, and store it in an S3 bucket. Use a Glue crawler, put it in Athena, put it into QuickSight.

You'll notice though on the far left or far, right. Um, that the insights are still getting kind of stuck in PowerPoint.

Uh we get to that point where we have a dashboard, a reader. Um, and a business user can look at it and understand it. But, uh, at the end of the day, that chief strategy officer of the hospital just wants one slide that tells them exactly what they're looking at.

Um, and until stories which we're about to talk about, um, it was all stuck in powerpoint. So how is it helping our customers, um, both internally and externally?

Um, the story i wanna leave you with today, if you remember anything, is that the barrier to entry for every different persona type is decreasing with these generative b i capabilities.

So, first for my team, specifically, i author dashboards every single day, um, i teach my analysts how to author dashboards every single day. Um, and curating the content has gotten faster since this generative b i went into, uh, preview earlier this year.

Um, actually a couple of weeks ago, i took it on a test run and built a dashboard. You know, i know how to use quicksight very well. But i use the generative b i authoring experience and it decreased the amount of time it, i needed to make the dashboard by an entire day.

Normally, it would take me three to build it, look it over, understand everything that's there. And this time it took me two. So even saving one day can be powerful for the story experience and you'll see my pipeline shifted a little bit on the right there of keeping insights inside of quick sites.

Um we're empowering our subject matter experts to be able to build and share compelling insights inside of quicksight. So it's taking those dashboards that have 20 or so visuals, multiple tabs, lots of different information and filters and boil it down to four or five different text blocks that show the exact um insights you see there as a human, as well as some next steps. And so what's uh that we can take action on?

So with that, i'm gonna jump over to the demo. Um the computer locked, sorry. Perfect. Uh so like ed, i did record a video here, but i wanna bring us back to what this data is.

So uh first caveat, this was real data that i took and put in the chicago market. So we are looking at actual claims that did occur for one of my clients. Um but we're looking at fake hospital names and fake cities.

So if you scroll really hard there, um you'll find like the shed aquarium is one of the lat lungs, obviously not a hospital. But um the actual insights and concepts that we're seeing here are real data.

Um this is one of those out migration data sets like i was talking about before. Um and it takes all patients who live in now, chicago and finds out where they got care within a 3.5 year time frame. How much of it stayed in the market? How much of it left the market? What types of things left all answers we're gonna solve for.

So i did build the full dashboard using jared b i. Um but we'll walk through a couple of different visuals together.

So the first one, the first question i might ask is is out migration going up or down. So trend of encounters in and out of market pretty simple question.

Um it gives me a table back. I kinda want that to be a line graph and we can see right off the bat um that the trend of line graph shows out migration is increasing. It's that dark blue line there.

So we wanna kinda understand what, what is leading to that next question i might ask is encounters by latitude and longitude. Where are these patients going? What hospital systems uh what cities outside of chicago?

I get very pointed here in tq i want it as a points on a map graph and add that to my visual as well. If i change my mind after i've put the graph on the analysis, i can actually change the color here like zach did in his demo and change it just as so the in and out of market.

Um so those dark blues are still the out of market and we can see if you're familiar with chicago patients are going to lots of different suburbs as well as up to wisconsin and, and over to other cities in chica uh in illinois.

Finally, my last question is, well, what are they leaving for? So i can ask q encounters by service line colored by out migration as a stacked bar chart looking to see that ratio of how much is staying in versus out of market and add that to my analysis.

So pretend we built 20 or so visuals here together. Um but we'll pop over to the reader view of this and we've got a full dashboard now.

So i'm gonna shift personas a little bit to one of my coworkers. Hannah. Hannah is an analyst on my team. She works with this data every single day and talks to hospitals about the insights here.

So hannah would log into this dashboard and see right away as a subject matter expert that cardiology is in both of the top left boxes. It's a top service line, it's a top service line leaving the market. But what hannah doesn't see is what types of cardiology.

There could be a lot of different things leaving, are they leaving for a heart transplant? That's very specific? Are they leaving for hypertension? That's something primary care wise.

So hannah can ask you what are the top cardiology subs service lines going out of the market again, pretty broad, wants to understand what the top things are leaving the market.

Q populates for her both a list of all those subs service lines, a visual of all of them as well. And we can see that it was hypertension, which is a primary care concern and it gives her some of those. Did you mean questions? And here's what we pulled out as well.

So hannah feels pretty confident that she understands there's a data story here to tell her hospital client.

So hannah's gonna build a data story about all the things going out of her market. We're gonna click the scroll page here and again, give it a pretty general and generic prompt.

Build me a story about why patients are going out of market for care. Pick relevant data showing what is driving their behavior and explain where they are going and what types of services they might be seeking. Again, very, very general. We wanna understand what the a i is showing us.

I'm gonna pause here for a second and point something out that i found really cool when i was doing the demo of this, but we're picking a bunch of visuals out of this exact dashboard that i am here. But you'll see and it's blurred because it has all my client list in it. You'll see the list on the left there has all the other dash boards that i as the reader have access to.

So maybe i have access to three or four different dashboards. Maybe one of them is about cardiology itself. I can pull in visuals of relevant data from other areas of my quick site and pull it in here as well.

But back to my re invent demo. So hannah picks in a bunch of these visuals and it populates a story as we saw in zack's demo, you can change the style of it there. I i like this one. It matches our, our themes in huron.

I can add my emily arvio name there and i can customize it to me. It gives you an introduction, scrolls through several of the different graphs that i've chose and then it shows you those top cities where patients are going out of market.

And again, i want to pause here and highlight something. Another thing that i found really cool and exciting, but this visual you'll see is a list of chic like suburbs near chicago, right?

Um what you don't see is chicago. But when we saw our mat visual, there was a giant bubble for chicago. So this visual is already prefiltered in the dashboard. I filtered it when i was the author and it keeps and sticks that filtering for the business user who is building the visual as well.

And then finally, it gives you some insights. And the one i highlighted there is these cities like they attract patients, do their proximity, patients are staying near the market. The biggest thing we saw in that uh reader experience was that they were leaving for hypertension, the primary care service and these patients are kind of staying in market but having to drive to suburbs for care.

So with that, i'll bring back up ed and zach. Thank you. Nice job. Great talk. Thank you. Well done, well done eds.

All right. Um so uh we're just gonna finish it up here. So uh so we looked today at these three experiences that come with amazon q in quicksight, this a i powered dashboard authoring experience to enable business analysts to build dashboards faster.

The a i answers to questions of data on demand for business users to get the insights they need in order to inform a decision and the a i assisted data storytelling which enabled business users to then share their findings with others in order to bring people on the same page in order to move a decision forward.

Uh we can share more. Thank you, share more information with you on generative b i.com. Uh if you'd like to know how to get all of these things. They uh are currently a public preview which good news means that they are free for quicks site q subscribers.

And if you don't already have quicks site q, you can subscribe to quicks site q as an add on to quicks site for a 30 day free trial. So you can try all of this stuff in quicks site for uh free pretty soon. Uh right after the conference if you choose to do so, uh and there are a number of other quick site sessions.

Uh i'd like to point your attention to just in case. Uh you have not already completely filled your schedule. And if you wanna hear me, give this talk again in a little while, uh but without my wonderful core, uh you can, you can do so over in uh the venetian.

Uh we have a community. Quicksight is uh a great community of people. It's more than aaa service, it's more than a product. It's the experience of working together to solve these data problems. And you can join, all you have to do is uh go up to the community. Very friendly, very helpful, can answer questions, get you rolling.

Um and you can connect to the three of us, please. Uh this is just our linkedin. So feel free to take a quick snap if you want.

Uh and then on this next slide, uh of course, everybody, please fill in your surveys. Uh we we do, uh, listen and, and look at all of your comments and, uh, they, they matter to us.

Uh thank you for taking the time today. I know that you had many different options during this time period and i'm flattered that you chose to spend your time with the three of us. We're gonna stick around here and take questions for about 10 minutes.