 Good morning fellow nerds and welcome back to AWS re-invent. We are live from the show floor here in Las Vegas, Nevada. My name is Savannah Peterson. Joined with my fabulous co-host, John Furrier. Day two, keynotes are rolling. What do you think? This is the day where everything comes to the cork, it's popped off the bottle. All the announcements are flowing out. Tomorrow you're going to hear machine learning from Swamy, a lot more in-depth around AI probably, and then developers with Werner Vogels, the CTO, who wrote the seminal paper in early 2000s around web services that became AWS. So again, just another great year of next level cloud, big discussion of data in the keynote, a bulk of the time was talking about data and business intelligence, business transformation. Easier is what people want. They want the easy button, and we're going to talk a lot about that in this segment. I'm really looking forward to this interview. How about the easy button? We all want the easy button. We want the easy button. I love that you brought up Champaign. It really feels like a Champaign moment for the AWS community as a whole being here on the floor, feels a bit like the before times. I don't want to jinx it. Our next guest, Scott Castle, from SciSense. Thank you so much for joining us. How are you feeling? How's the show for you going so far? Oh, this is exciting. It's really great to see the changes that are coming in AWS. It's great to see the excitement, the activity around how we can do so much more with data, with compute, with visualization, with reporting. It's fun. It is very fun. I just got a note. I think you have the coolest last name of anyone we've had on the show so far, Castle. Oh, thank you. I'm here for it. I'm sure no one's ever said that before. So just in case our audience isn't familiar, tell us about SciSense. So SciSense is an embedded analytics platform. So we're used to take the queries and the analysis that you can power off of Aurora and Redshift and everything else and bring it to the end user in the applications they already know how to use. So it's all about embedding insights into tools. Embedded has been a real theme. Nobody wants to, I keep using the analogy of multiple tabs. Nobody wants to have to leave where they are. They want it all to come in there. Now, this space is older than I think everyone at this table. BI's been around since 1958. How do you see SciSense playing a role in the evolution thereof? We're in a different generation of analytics. Yeah. I mean, BI started, as you said, 58 with Peter Lund's paper that he wrote for IBM kind of became popular in the late 80s and early 90s. And that was Gen 1 BI. That was Cognos and Business Objects and Lotus 1, 2, 3. I think like green and black screen days. And the way things worked back then is if you ran a business and you wanted to get insights about that business, you went to IT with a big check in your hand and said, hey, can I have a report? And they'd come back and here's a report. It wasn't quite right. You go back and cycle, cycle, cycle, and eventually you'd get something. And it wasn't great. It wasn't all that accurate, but it's what we had. And then that whole thing changed in about 2004 when self-service BI became a thing. And the whole idea was, instead of going to IT with a big check in your hand, how about you make your own charts? And that was totally transformative. Everybody started doing this and it was great. It was all built on semantic modeling and having very fast databases and data warehouses. Here's the problem. The tools to get to those insights needed to serve both business users like you and me and also power users who could do a lot more complex analysis and transformation. And as the tools got more complicated, the barrier to entry for everyday users got higher and higher and higher to the point where now we look at Gartner and Forrester and IDC. This year they're all reporting the same statistic between 10 and 20% of knowledge workers have learned business intelligence. Everybody else is just waiting in line for a data analyst or a BI analyst to get a report for them. And that's why the focus on embedded is suddenly showing up so strong because little startups have been putting analytics into their products. People are seeing, oh, this doesn't have to be hard. It can be easy. It can be intuitive. It can be native. Well, why don't I have that for my whole business? So suddenly there's a lot of focus on how do we embed analytics seamlessly? How do we embed the investments people make in machine learning, in data science? How do we bring those back to the users who can actually operationalize that? And that's what Sison does. It's interesting, Savannah. Data processing used to be what the IT department used to be called back in the day. Data processing. Now data processing is what everyone wants to do. There's a ton of data. We saw the keynote this morning at Adam Sileski. There was almost a standing ovation. Big applause for his announcement about ML-powered forecasting with QuickSight Cube. My point is people want automation. They want to have this embedded semantic layer in where they are. Not having all the process of ETL or all the muck that goes on with aligning the data. All this, look at a lot of stuff that goes on. How do you make it easier? Well, to be honest, I would argue that they don't want that. I think they think they want that because that feels easier. But what users actually want is they want the insight right when they are about to make a decision. If you have an ML-powered forecast and Sison's has had that built in for years now, you have an ML-powered forecast. You don't need it two weeks before or a week after in a report somewhere. You need it when you're about to decide, do I hire more salespeople or do I put 100 grand into a marketing program? It's putting that insight at the point of decision that's important. And you don't want to be waiting to dig through a lot of infrastructure to find it. You just want it when you need it. What's the alternative on a time standpoint? So real time insight, which is what you're saying. What's the alternative if they don't have that? The alternative is what we are currently seeing in the market. You hire a bunch of BI analysts and data analysts to do the work for you. You hire enough that your business users can ask questions to get answers in a timely fashion. And by the way, if you're paying attention, there's not enough data analysts in the whole world to do that. Good luck hiring them. Time to get it. I really empathize that when I used to work for a 3D printing startup and I have just, I mean I would call it PTSD flashbacks of standing behind our BI guy with my list of queries and things that I wanted to learn more about our e-commerce platform and our marketplace and community. And it would take weeks. And I mean this was only in 2012. We're not talking 1958 here. We're talking a decade in startup years is 100 years in the rest of the world life. But I think it's really interesting. So talk to us a little bit about infused and composable analytics. And how does this relate to embedded, yeah. So embedded analytics for a long time was I want to take a dashboard I built in a BI environment. I want to lift it and shift it into some other application so it's closer to the user. And that is the right direction to go. But going back to that statistic about how, hey, 10 to 20% of users know how to do something with that dashboard. Well how do you reach the rest of users? Yeah, when you think about breaking that up and making it more personalized, so that instead of getting a dashboard embedded in a tool, you get individual insights. You get data visualizations, you get controls. Maybe it's not even actually a visualization at all. Maybe it's just a query result that influences the ordering of a list. So like if you're a CSM, you have a list of accounts in your book of business. You want to rank those by who's the most likely to churn. You get that, how do you get that most likely to churn? You get it from your BI system. So, but then the question is, how do I insert that back into the application the CSM is using? So that's what we talk about when we talk about infusion. And Sysons started the infusion term about two years ago and now it's being used everywhere. We see it in marketing from Click and Tableau and from Looker just recently did a whole launch on infusion. The idea is you break this up into very small, digestible pieces. You put those pieces into user experiences where they're relevant and when you need them. And to do that, you need a set of APIs and SDKs to program it, but you also need a lot of very solid building blocks so that you're not building this from scratch. You're assembling it from big pieces. And so what we do at Sysons is we've got machine learning built in. We have an LQ built in. We have a whole bunch of AI powered features including a knowledge graph that helps users find what else they need to know. And we provide those to our customers as building blocks so that they can put those into their own products, make them look and feel native and get that experience. In fact, one of the things that was most interesting in the last couple of quarters is that we built a technology demo. We integrated Sysons with Office 365, with Google Apps for Business, with Slack and MS Teams. We literally just threw an NLQ box into Excel and now users can go in and say, hey, which of my salespeople in the Northwest region are on track to meet their quota and they just get the table back in Excel. They can build charts of it in PowerPoint and then when they go to do their QBR next week or the week after that, they just hit refresh to get live data. It makes it so much more digestible and that's the whole point of infusion. It's bigger than just the iframe-based embedding or the JavaScript embedding we used to talk about four or five years ago. APIs are very key. You brought that up. That's going to be more of the integration piece. How does embeddable and composable work as more people start getting on board? It's kind of like a flywheel. How do you guys see that progression? Because everyone's copying you, we see that. But this means it's standard. People want this. What's next? What's that next flywheel benefit that you guys are coming out with? Composability, fundamentally. If you read the Gartner analysis, when they talk about composable, they're talking about building pre-built analytics pieces in different business units for different purposes and being able to plug those together. Think of containers and services that can talk to each other. You have a composition platform that can pull it into a presentation layer. Well, the presentation layer is where I focus. For us, composable means I'm going to have formulas and queries and widgets and charts and everything else that my end users are going to want to say almost minority report style if I'm not dating myself with that. I can put this chart here, I can put that chart here, I can set these filters here and I get my own personalized view, but based on all the investments my organization's made in data and governance and quality so that all that infrastructure is supporting me without me worrying much about it. Well, that's productivity on the user side. Talk about the software angle, development. Is there low code, no code? Is there coding involved? APIs are certainly the connective tissue. What's the impact? Yeah, how hard is it? Oh, so if you're working on a traditional legacy BI platform it's virtually impossible because this is an architectural thing that you have to be able to do. Every single tool that can make a chart has an API to embed that chart somewhere but that's not the point. You need the lifecycle automation to create models, to modify models, to create new dashboards and charts and queries on the fly and be able to manage the whole lifecycle of that so that in your composable application when you say, well, I want a chart and I want it to go here and I want it to do this and I want it to be filtered this way, you can interact with the underlying platform. And most importantly, when you want to use big pieces, like, hey, I want to forecast revenue for the next six months. You don't want to be popping down into Python and writing that yourself. You want to be able to say, okay, here's my forecasting algorithm, here are the inputs, here's the dimensions and then go and just put it somewhere for me. And so that's what you get with SciSense and there aren't any other analytics platforms that were built to do that. We were built that way because of our architecture. We're an API first product, but more importantly, most of the legacy BI tools are legacy. They're coming from that desktop, single user, self-service BI environment and it's a small use case for them to go embedding. And so composable is kind of out of reach without a complete rebuild. But with SciSense, because our bread and butter has always been embedding, it's all architected to be API first. It's integrated for software developers with Git but it also has all those low code and no code capabilities for business users to do the minority report style thing and assemble analytics components into a workable digital workspace application. Talk about the strategy with AWS. You're here at the ecosystem, you're in the ecosystem, you're leading product and they have a strategy, we know their strategy, they have some stuff but then the ecosystem goes faster and ends up making a better product in most of the cases if you compare. I know they'll take me to school on that but that's pretty much what we report on. The Mongols do a great job. They have databases. So you kind of see this balance. How are you guys playing in the ecosystem? What's the feedback? What's it like? What's going on? AWS is actually really our best partner and the reason why is because AWS has been clear for many, many years. They build componentry. They build services, they build infrastructure, they build Redshift, they build all these different things but they need vendors to pull it all together into something usable. And fundamentally that's what SciSense does. We didn't invent SQL, we didn't invent Jackal or Dackle. These are underlying analytics technologies but we're taking the bricks out of the briefcase. We're assembling it into something that users can actually deploy for their use cases. And so for us, AWS is perfect because they focus on the hard bits, the underlying technologies. We assemble those, make them usable for customers and we get the distribution and of course AWS loves that because it drives more compute and it drives more consumption. How much did they pay you to say that? That was a wonderful pitch. We always say, hey, they got a lot of great goodness in the cloud but they're not always the best at the solutions and that they're trying to bring out and you guys are making these solutions for customers that resonates with what they got with Amazon. For example, last year we did a technology demo with Comprehend where we put Comprehend inside of a semantic model and we will compile it and then send it back to Redshift. And it takes Comprehend which is a very cool service but you kind of got to be a coder to use it. I've been hearing a lot of hype about the semantic layer. What is going on with that? The semantic layer is what connects the actual data, the tables in your database with how they're connected and what they mean so that a user like you or me who's saying I want a bar chart with revenue over time can just work with revenue in time and the semantic layer translates between what we did and what the database knows about. So it speaks English and then they converts it to data language. That's exactly right. Yeah, it's facilitating the exchange of information and I love this. So I like that you actually talked about it in the beginning, the knowledge map and helping people figure out what they might not know. I am not a BI analyst by trade and I don't always know what's possible to know and I think it's really great that you're doing that education piece. I'm sure especially working with AWS companies depending on their scale that's got to be a big part of it. How much does the community play a role in your product development? It's huge because I'll tell you one of the challenges in embedding is someone who sees an amazing experience in outreach or in seismic. I think it's good to say I want that and I want it to be exactly the way my product is built but I don't want to learn a lot. And so what you want to do is you want to have a community of people who have already built things who can help lead the way and our community, we launched a new version of the Sysons community in early 2022 and we've seen a 450% growth in that community and we've gone from an average of one response per post. Oh yeah, 450%, I just want to put a little exclamation point on that. Yeah, that's awesome, wow. We've tripled our organic activity so now if you post this Sysons community it used to be you'd get one response maybe from us, maybe from a customer. Now it's up to three and it's continuing to trend up. So we're seeing that. It's amazing how much people are willing to help each other if you just give them a platform to do it. Oh, it's great. I mean, business is so competitive. I think it's time for the Instagram challenge, the reels. Lay it on him, John. So we have a new thing we're going to run by you. We used to call it the bumper sticker for re-invent. Instead of calling it the Instagram reels, if we're going to do an Instagram reel for 30 seconds, what would be your take on what's going on this year at re-invent, what you guys are doing, what's the most important story that you would share with folks on Instagram? I think it's really, what's been interesting to me is the story with Redshift Composable. Redshift Serverless, one of the things I've been seeing. We know you're thinking about Composable a lot. It's in there, it's in your mouth. So the fact that Redshift Serverless is now going to be coming the de facto standard, it changes something for my customers. Because one of the challenges with Redshift that I've seen in production is if, as people use it more, you've got to get more boxes. You have to manage that. The fact that Serverless is now available, it's the default, means that now people are just seeing Redshift as a very fast, very responsive repository. And that plays right into the story I'm telling, because I'm telling them it's not that hard to put some analysis on top of things. For me, maybe it's a narrow Instagram reel, but it's an important one. And that makes it better for you because you get to embed that, and you get access to better data, faster data, higher quality, relevant, updated. Yep. And as it goes into that 80% of knowledge workers, they have a consumer great expectation of experience. They're expecting that 0.5 MS response time. They're not waiting two, three, four, five, 10 seconds. They're not trained under the OLAP expectations. And so it matters a lot. Final question for you. Five years out from now, if things progress the way they're going with more innovation around data, this front end being very usable, semantic layer kicks in, you got the Lambda, and you got Serverless kind of coming in and helping out along the way, what's the experience going to look like for a user? What's it, in your mind's eye, what's that user look like? What's their experience? I think it shifts almost every role in a business towards being a quantitative one, talking about, hey, this is what I saw, this is my hypothesis, and this is what came out of it, so here's what we should do next. I'm really excited to see that sort of scientific method move into more functions in the business, because for decades it's been the domain of a few people like me doing strategy, but now I'm seeing it in CSMs, in support people, in sales engineers, in line engineers, that's going to be a big shift. Awesome, thank you. Scott, thank you so much, this has been a fantastic session, we wish you the best at SciSense. John, always a pleasure to share the stage with you. Thank you to everybody who's attuning in, tell us your thoughts, we're always eager to hear what features have got you most excited, and as you know, we will be live here from Las Vegas at Reinvent, from the show floor, 10 to six all week, except for Friday, we'll give you Friday off, with John Furrier, my name's Savannah Peterson, we're theCUBE, the, the, the leader in high tech coverage.