AWS Mid-Year Leadership Summit 2025 | Jose Kunnackal, AWS

Clips
News
More from AWS Mid-Year Leadership Summit 2025

Jose Kunnackal

Director, Amazon QuickSight

AWS

From dashboards to decisions: How AWS is transforming data access at scale

Interactive data insights are rapidly redefining how businesses navigate decisions, turning static dashboards into dynamic tools that inform action in real time.As enterprise demands outpace traditional business intelligence capabilities, platforms such as Amazon QuickSight are stepping in with a more intuitive, accessible approach. QuickSight blends ease of use with deep analytical power — making it possible for users at any skill level to dive into data, explore trends and ask complex questions through natural language

play_circle_outline AWS Halftime Report discusses updates in 2025 at Seattle headquarters.

play_circle_outline Democratizing Data: Jose Kunnackal Explores Amazon QuickSight's Generative BI Features for Broader Business Insights and Access

play_circle_outline Launch of Scenarios enhances user experience with Agentic AI technology in QuickSight.

play_circle_outline Enhancing User Experience: Simplifying QuickSight for Non-Technical Users and Exploring Its Comprehensive Offerings for Developers and Businesses

play_circle_outline Enhancing Business Intelligence: QuickSight's Impact on Performance, User Management, and Data Accessibility in Real-World Success Stories

play_circle_outline Future innovation aims to combine structured and unstructured data for richer insights.

play_circle_outline Enhancing Analytics Workflows: The Critical Role of Data Cleaning, Management, and Metadata in Effective Data Utilization

Info
Transcript

Jose Kunnackal, AWS

Jose Kunnackal

Director, Amazon QuickSight AWS

In this in-depth interview from the AWS Mid-Year Leadership Summit, Jose Kunnackal, director of Amazon QuickSight, joins theCUBE’s John Furrier to explore how AWS is redefining business intelligence through generative AI and agent-based analytics. Kunnackal breaks down the expanding QuickSight portfolio, including Q Business, Q Developer and Scenarios, and how these offerings are enabling customers to derive faster insights from both structured and unstructured data.

The discussion traces QuickSight’s evolution from a cloud-native BI tool into a gene... Read more

explore Keep Exploring

What updates and announcements were made by AWS in 2025? add

What are the key usability features and customer demands that led to the development of QuickSight? add

What challenges arise when scaling a large number of personnel in organizations like Amazon, and how does QuickSight's Q and Scenarios with Agentic AI aim to address user inquiries? add

What are the features and components of QuickSight and its related offerings? add

What are the features and capabilities of QuickSight that support user access and data management for business reporting? add

What are the current capabilities and features available in QuickSight and Q Business for integrating and utilizing data within a business? add

What is the importance of data quality and metadata in AI systems, particularly in the context of AWS services like QuickSight? add

bolt Powered by CUBE AI

Jose Kunnackal, AWS

search

>> Hello, I'm John Furrier with theCUBE. We are here at Seattle's headquarters for AWS for a Halftime Report on all the updates in 2025. It feels like the full year with the tons of announcements. You can almost run re:Invent Now, but it's still six months away, so it's halfway midpoint throughout the year. And we got a great guest, Jose Kunnackal is the director of Amazon QuickSight, a fast-growing product, something I can't wait to get my hands on. Jose, it's great to see you. Thanks for coming on Halftime Report. There's so much going on packed in one year. I can't even keep track of all the announcements.

Jose Kunnackal

>> Oh yes, it's been great. I think we've seen a lot of momentum and the year's been exciting to say the least. We launched something called Scenarios, so one of our advances in Generative BI, so that's been very interesting to a lot of customers and we've also had a lot more going on.

>> Yeah, I talked to Matt Garman and I was telling him on his LinkedIn, another region gets launched, more data center CapEx investment for them to fuel the demand for more large scale systems, more updates on the models. Swami's posting about all the greatness on Q. QuickSight seems to fit that narrative where you're starting to see, I won't say democratization, that's been used, but you're seeing non DevOps people come into the cloud to move fast with their data. That's the theme. Unlocking the value of data and workflows and they want to get it fast. But there's two markets. There's the DevOps side, provisioning infrastructure. Now, you have a whole nother usability feature.

Jose Kunnackal

>> 100%. I think a lot of the emphasis has been on how do you get users to really find value from their data? And the DevOps side of it, making that as simple as possible definitely is something we want to do. But at the same time, with a product like QuickSight, what we've seen is over the years, there's been a lot of customer demand for scaling insights to everyone. So one of the reasons why we started out with QuickSight to start a long time back was this question from customers of how they scale insights to every user. So data and insights are not just for the CEO or the managers, it's for every user because every user needs to know, hey, what's happening? How many pieces of inventory do I have left? What do I need to do? And that became difficult with traditional products, especially in the BI space because these were all built on some kind of server. And when you have 20,000, 50,000, Amazon's got hundreds of thousands of people who are working on things, it becomes very challenging for you to scale that. So, QuickSight and we brought in Q. And so the whole notion of Q was how do we make it simple for users to ask that question, get an answer? And then I mentioned Scenarios, where with Scenarios, we are bringing in Agentic AI in order for users to ask that question and then continue that. So let Q and let the system think about the problem, come back with a plan and then work through it.

>> Lay out the structure of the portfolio because sometimes, it can be confusing. You got Q, you get Q for developer, Q for business, you got QuickSight. How do I understand the map? How should people think about QuickSight? Is it all of that? Is QuickSight everything or is it Q? Explain the hierarchy of the offering, the portfolio.

Jose Kunnackal

>> Yes, that's a great question. So we've had QuickSight for a number of years, which is our cloud-native BI product. It offers everything from dashboards to Pixel perfect reports to embedded analytics that you can add into your own products. In QuickSight, we launched Q back in 2021, which was a natural language interface. So that would allow you to ask questions of your data, so your database tables and tabular data. And over the last few years, we introduced a larger portfolio of Q services, which is Q Developer, which is helping developers code faster and better, especially with generative AI coming in. And then you've got Q Business which allows you to go against your unstructured data. So think your data in files, in SharePoint. So there's a lot of enterprise knowledge that's not in databases, not in data warehouses or data lakes. And these are just files and pieces of information that exist across the board. So whether it's in a third-party system like SharePoint or just files sitting on your desktops. And it became very important and it became possible for Generative AI to unlock this value. And so that's where we saw the potential of Q Business. So Q Business connects to a number of sources and it supports the permissioning of data by user so that you see the data that you're entitled to, which is very important because opening up all of the data to all of the people is simply confusing but also massive governance nightmare.

>> One of the things, just to take a sidetrack on this is that go back 10 years, okay? 10 years ago all the rage was you coming out of high school was, I want to get a CS degree because people saw that having a computer science degree or coding was a superpower. Okay, great. And that's true. Now, you're starting to see people saying, I don't really need to get a computer science degree, but I want to have a coder. How do I find a CTO or a coder? Now you have the era where it's available to code. So if you know the business logic, you can actually have a coding assistant. So I think we're going to hit another wave of innovation where people who understand the business piece will have coding capabilities. That's why I love Q Developer and that's the hottest area in Gen AI right now. So that's a phenomenon. Okay, so put that aside for a minute. We'll come back to that. The other thing is that the BI world of analytics, it's been around for a while, as you mentioned, that's colliding with generative AI and there's an intersection of business analytics and business intelligence with generative AI. So how is gen AI impacting analytics? That's the data. Data feeds the AI, AI's changing the face of business, everyone's implementing it and the people who now can code don't need to hire a coder or can learn a little bit and code more. So do more with more. Talk about that dynamics. I think this is the key point in the industry where the enablement is going to explode and create a huge aperture of value creators.

Jose Kunnackal

>> That's definitely a factor that's impacting the world of BI as well because you will find that traditionally, data was in a data warehouse. There is a BI person, a SQL person who's writing some queries and then translating that into something that you can see. Now, there is a difference between a generative AI or an LLM writing SQL against some data, to finding a business answer to a question. So what were my sales in North America? When you ask that question it actually needs to go through a number of business rules before it gets you the right answer, otherwise, it's going to get you an answer because the LLM can go against that. So, what's happening in the world of BI is that agents and LLMs are actually helping make that process much faster. So with Q and QuickSight as an example, what we find is that customers or business users are able to ask questions of data, whether it's in a dashboard or what we call a topic, which is essentially a link to a table. You're able to ask that question and get an answer instantly. And that was a process that would've taken maybe weeks because every BI team involved in this world is overloaded with questions. And traditionally, what that would've also led to is building yet another dashboard. And what you don't want is that sprawl of dashboards, but you do want answers. So when we built Q in QuickSight, we did focus a lot on making sure that when you ask that question, we give you... Traditionally, you get a single visual. So you ask me, "What are sales?" It says, "Hey, sales are $5 million." Now what you want is, okay, what were the sales? What is the trend? All of that. And you'll find that with Q and QuickSight, we've been able to put in a unique multi-visual response that provides that. But you ask about that part that you mentioned about anybody can be a coder. So that's where our recent launch of Scenarios really comes in where with Scenarios, you have a data analyst by your side. And the reason I say that is you can ask a question like what if we were to increase a free trial by two months? As a product person, let's say you had a product and you had a free trial. Now with Scenarios, you'll find that it is able to go and build a plan around it. So it finds data that's in the system for you and says, "Hey, these are the pieces of information that might be relevant. This is what I'm planning to do." So it shows you a plan and as a business person, you know the plan, but often you don't know how to run, let's say a regression or do some forecasting. You're able to do that using Scenarios. So Scenarios will do the work for you. So in many ways, it is accelerating the work that you would've done, and at the same time, it is working in tandem with the business user and the business teams that are coding some of the logic.

>> On the business logic, I was having a conversation with a bunch of thinkers on a retreat this past weekend, young tech entrepreneurs and leaders as well as old, old experienced like us. The debate was, of course, I love the AI infrastructure, more GPUs, more XPUs, better models. But it came back down to the app layers where the business logic is, and that's where the gold is. Agentic is going to enable that. On our last interviews, we talked a lot about agentic. How is agentic feeding into it? Because you've got Scenarios, the business uses the target. How does a company leverage that? Because if I'm in a company, I might not give the access to the cloud to the business user. I'll give them to the DevOps team for sure, they're provisioning the resource on AWS. How do you connect that agent enablement into the business logic or the business side on QuickSight? Because I see QuickSight as an interface to the cloud. So take us through that. I'm a customer. I got the console. Not has root access, but you don't want to give out those credentials. But how do I enable QuickSight for my team? What's the best practice? Because this seems to be the better way than just sending a request into the DevOps team and saying, "Give me a report."

Jose Kunnackal

>> Definitely. So with QuickSight, you do have an interface that's completely off. You mentioned the AWS console. So it is off the console. So it's like a regular SaaS application that you access. So you can get in there, you can ask the question... I'm sorry, you can get in there and you start from the top. So you can get in there and provision access, and you do have the ability to connect your enterprise identity system. And then you can grant different levels of access to different users. So like you said, there are going to be a set of people who need to manage the data or set up the business rules, and then there are going to be a set of people who can ask questions. So as an example, if you look at Whole Foods, Whole Foods recently switched to QuickSight and the thing is, they do have store managers, like over 25,000 users across their stores. These are supervisors in their stores who have access to QuickSight. So these folks need access to dashboards, they need to be able to quickly look at the data and they get access at a certain level, so we call that a reader. And then you also have people who are building the dashboards and all of that. They get author-level access. But the idea is that you have this easy-to-use interface that's off-console and you can go and ask those questions. Now, the moment you ask another question, which is how do you make this useful from a business standpoint? So that's where we see another service. We talked about Q Business earlier, which connects to your enterprise data. So at re:Invent last year, we did bring in the notion of connecting Q Business and QuickSight. So now, think of what can happen if you're able to ask a question of your data, but it takes not only the numbers into consideration, but also your notes from, let's say your notes that are in S3 or SharePoint or in Confluence, pick your product. So when you connect that data and then you ask a question like let's pick the same question, what was sales in North America? And then you get a response that contains both the visuals but also, context saying that you know what? Sales and let's pick a state, California are going up because of a certain demand. And that's in your notes. And now, that bubbles up into the question. So really, the point is how do you let a business user ask that question and read from both the unstructured data or the qualitative data and then the quantitative or the structured data altogether? And by getting users into the QuickSight interface, you do that. And we've been building QuickSight for quite some time. So one of the core things has been building IT cloud native. So again, going back to the scale point, every user needs to have the performance that makes them productive. And if you look at how I mentioned Whole Foods, the previous tool that they used took like, I don't know, about a minute simply because it was set up at a different time. Now, dashboards take about four seconds. So store manager, cutting down time by 60 seconds to four seconds is a massive difference and you're saving a whole bunch of time that they can-

>> There's a huge productivity gains. Clearly AI's checked the box on that. I think there's no debate in the industry. Certainly AI's got the productivity. I want to ask you about how... People ask me, how do I get there? And you have two groups, the data team and the business users, the business teams and whatever. The question comes up is, I get AI, but I know I've got to do the workflows, I got to get the data. How do I get the data in? And a lot of people are preparing things on spreadsheets knowing it will be automated. So is there a readiness best practice for companies? Let's just say pick a company out there that says, "Hey, you know what? My folks are using this cloud-based spreadsheet." Or it could be anything, but we have all the data, it's manual, but it's good manual work, it's inbound logistics. We'd love this process, it's well documented. We're happy to fill out these fields once, data entry. Once it's in there, you now have part of the data. That's the trend we're seeing. How is that translating with QuickSight and how do you see agents coming and saying, "Okay, we'll take it from here." That seems to be what people want. They go, "I'll do the work, now what?"

Jose Kunnackal

>> That's a great question. And AWS has a very large set of analytics services that provide you all the way from the catalog to data processing to data transformation to QuickSight. And what we see is a trend where exactly what you said, there are companies where people are doing this in Excel files. So, let's pick a media and entertainment company, very traditional. They've used spreadsheets because that's the way they work. But to get to that spreadsheet from their artist and data actually takes about a month. Now, enter a product like QuickSight. What we've seen is that by transforming that spreadsheet into something that can be input into QuickSight, because QuickSight connects to S3 and by using some of the underlying services in AWS like Glue and others, you're creating a structured transformation from a spreadsheet-based system. The end user can still use a spreadsheet, but you're transforming that spreadsheet into a pipeline that now goes all the way into QuickSight. And in the case of that customer that I was just talking about, we went down from a month to less than a week because you still need checks that they have manual processes that need to be done, but that kind of transformation helps. And then you have the part where you can also on AWS think about how the other infrastructure pieces fit in order to get you to that agentic future. Because QuickSight's AI capabilities are also built on Bedrock. So by using that common foundation, you are able to build other services that are bespoke to your needs if needed, or use something that's in QuickSight. So you could just ask questions of Scenarios or Q&A in QuickSight-

>> So the benefits on top of that data. Just back up real quick. Scope the timeframe you mentioned the time it takes, is that setting up the pipeline? Was it a manual process now it's automated or not?

Jose Kunnackal

>> In that case it was automated. Yes. It is automated now. So previously what happens in many of these organizations, and we work with many of them because business teams typically have a mantra that if something's working, they don't want to quickly go and change it because it helps continuity. So what we've been seeing is with QuickSight, we are seeing many instances where they were traditionally on spreadsheets, but it's quickly moving over. At the same time, we also have very large enterprises. So pick Capital One, JP Morgan spoke with us at re:Invent last year. We do have a number of different companies that are working with us in order to transform their enterprise architecture. So rolling this out at scale, but at the same time you also get adoption from business teams that are just finding this tool to that question earlier. You don't need to go to the console. It's very easy to get in. So they're just going at it from a business productivity standpoint.

>> Talk about QuickSight from its origination. Obviously, a lot's changed. How would you describe QuickSight today for someone who says, really, I want to lean into this, I'm sold on the benefits, I'm willing to do the work. What is QuickSight today and where has it come from?

Jose Kunnackal

>> That's a great question. Early on when we started with QuickSight, QuickSight was probably the straggler among cloud-native BI services. Sorry, that's the wrong word. QuickSight is probably one of very few cloud-native BI services. What you see now is that a lot of the other services are trying to approach a cloud-native interface. So QuickSight, we've been building this foundation for about 10 years now. And on top of that, we have brought in dashboards, pixel-perfect reports, embedded analytics, and all of the generative BI capabilities. So if you go into QuickSight back in 2023, we introduced the notion of using Q in order to build your dashboards, in order to build calculations, in order to ask your questions as a business user and also create stories. So a number of LLM powered capabilities that came in via Q, which really benefit the business user. And what that translates to is the fact that this is a unique product that provides you all of these generative capabilities, but also it's very approachable, so you don't need to be an expert to start using QuickSight.

>> The developers always fascinates me. What are some of the trends you're seeing with customers? What's some of the adoption use case? Can you share a story of someone getting in? What was their experience? What was their approach and how does that relate to folks who want to use it now?

Jose Kunnackal

>> Yeah, yeah, definitely. So I think one of the stories that I can share is ETL Bank in Brazil, which rolled onto QuickSight I'd say roughly 24 months back. And they shared their story last re:Invent, and they went from zero to about 30,000 users in nine months. And typically, something like that takes years. I think in the past, they've ramped up to that number in I think, seven years or so. So the fact is that you're able to ramp up into that product at an enterprise scale with everybody starting to consume the product very, very quickly. And that requires a number of different dimensions. Everything from the foundation of security and governance, making sure that when you're rolling something out in an enterprise, everybody needs to have access, but they also need to have secure access. And at the same time, you also need to have visualization capabilities. So at we announced the integration with things like high charts, which make sure that you have a whole variety of visuals to go from. You have the Q&A capabilities and pieces that really make it approachable to business teams to get on board. And then you also have the features that you probably will not find in other VI products. So if you look at the integration with Q Business, the ability to go against your structured and unstructured data, that really does not exist. If you look at the Scenarios capability, the Agentic AI where you ask, it's not just a question and answer, but you say what you want in terms of an outcome and you have a data analyst that is really Q, working with you and walking you through what needs to be done, doing the regression, doing the forecasting. So I'd say that ramp up is fast and we have a number of enterprises that we are working with in order to move that really, really-

>> What are some of the cool things you're working on now in terms of what's next with QuickSight? I can imagine building an app on top of it. Back in the old web days, you used to call them shortcuts, a link to a page. But you got dashboards, okay, check, you got that. You got the Q for business, you got the developer side of it. Is there a future where I can build apps on QuickSight today and what's that look like in the future?

Jose Kunnackal

>> That's a great question because I think where you're getting to is you spoke about the developer part of things where things are moving really fast and as much as you have a CS degree, you can build even without a CS degree because you just need to know what you want to build. So one of the things that we are working harder and harder on is how do you get the unstructured data and structured data together, because that's really the foundation, the knowledge of your business. So there's a gold mine in every business team in terms of the knowledge that people have and knowledge that teams have. And then there's the question of what capabilities do you have? So today in QuickSight, you've got the dashboard building capabilities, you've got the Q&A capabilities, and in Q Business you do have a number of capabilities around asking questions of search and summarization. At re:Invent, we added over 50 actions. So these are things you can do in third party systems. So whether it is sending an email, sending a Slack message, connecting to Salesforce or Zendesk to update a ticket, there's a whole bunch of things there. And then we also worked on, we have Q Business also had some automation capabilities that we've been adding. So there's something called Q Apps, so a variety of capabilities in there. So we are trying to see how do we best bring them together in a format that's more and more consumable for the business user.

>> It's almost funny. AI is so magical because when people see how fast it can pipeline, all these hard stuff that can go back a decade, they'll be like, okay, I got to do a schema on this. I got to do all this work. People almost have the philosophy now, and this is why I think data lakes are hot, and I think it's going to change because when they realize what that data is used for, but the mindset now is, I just dump all my data into the cloud. I mean that in a positive way. Push it to the cloud and magic happens. Okay. So okay, it's not that simple, but take me through that. But the mindset of actually releasing the data, it could be log files from a website, traffic to database information, SQL files. So just, if I just pump everything into the cloud, that's the first step. That's where people's heads are at. I'm oversimplifying it, but take me through the next step. Take me through the progression because, okay, step one, get the data in the cloud, unstructured and structured. What happens next? What's the next progression for the customer? It all depends on what their use case is, but what's the general, high-level sequence of events.

Jose Kunnackal

>> There's a general saying that I think you'll find on LinkedIn and other places, which is that your AI is only as good as your data because data provides a lot of context, and if you provide too much unrelated context, you're just trying to muddle too many things together. So one of the fundamental things is when you have data in that form, how do you make it clean or how do you get it together in a format that makes sense for you as a user and also for any sort of agentic system? So I mentioned earlier how there are other AWS services. So within QuickSight, you do have options to go and essentially tag the data, provide metadata. So in QuickSight, in Q, we provide the notion of something called Topics. So with Topics, you can add custom metadata. So if you are a mortgage broker as a simple example, you might have terms that you use in the mortgage industry that people don't use outside. Being able to add that additional metadata is very important. So in Q Business you have the same thing where, as you bring your data-

>> Like jargon.

Jose Kunnackal

>> You can call it jargon, but people use that on a day to day-

>> Words that mean something in context.

Jose Kunnackal

>> In context, yes. So because when a business user asks a question, they are going to have that jargon or the specific words that they use, that's not going to change. So you can't ask them to say, use pure words in order to get responses.

>> Yeah. The joke in cybersecurity was no terrorist writes bomb on the manifest. But you have to infer that. But there's a lot of context and inference going on. So the cleaning is a big problem, opportunity to fix.

Jose Kunnackal

>> It's an opportunity. And we do have services in AWS, so there's the SageMaker Unified Catalog, there's a catalog, there's other services like Glue. So those allow a business team to go and specify the rules and clean that. And then in QuickSight as an example, you do have the option to go and put in that metadata in there, join the data as needed. So that's all the data team in some ways. And many of these teams also have the notion of a producer and a consumer. So there are central data teams that do this and business teams that consume it. But once you have it in the system, then you're able to start asking questions of it.

>> Jose, I want to get this, I've come full circle because I said earlier, the business intelligence, the analytics market, which has been around for a while, there's best practices', there's use cases, cultural job descriptions, and people do things. The generative AI is where the breakthroughs are coming in. The intersection of both those. You got breakthroughs happening, which is step function change, needle moving moments. What's the innovation equation that you're seeing right now where you can have the best of, I understand why I have analytics, but now, I have new disruptive enablement for breakthroughs for better things?

Jose Kunnackal

>> Yes.

>> What's going on in that intersection of those two markets?

Jose Kunnackal

>> I'd say that you'll see a mix of sentiment there in that market because you do have places where generative AI can directly play a part. So I'll give you a simple example. We talked about Scenarios. One of the teams inside Amazon use Scenarios against their ticketing data. And ticketing data can be messy because they have a number of tickets. Actually, to be clear, this was, I think feedbacks. So they have a lot of feedback and they're looking at the date, a bunch of things, but they also have raw text in there. So traditionally, a database system would find it difficult, but something like Scenarios, which is agentic now is able to parse that. It is able to correlate tickets or feedback suggestions with what components did that come from? What devices did it come from? But it's also able to mine the unstructured feedback. So the text that you have in there, it's able to mine that in order to make sure that you are getting the output of both the structured and unstructured data using an agentic system. So you're seeing differences there where these systems can get to it, but at the same time, not all data can be accessible that way. So some of the work will come in just making sure you have all the data piping together in order to form that unified lake as you were talking about.

>> I think if you can build a feature called the data cleaner, just press a button, just clean the data. That seems to be the heavy lifting. I got to ask you, from a top-selling or hottest feature, what's the best adopted or use case feature product component of QuickSight that's got the most traction right now for folks who are leaning into it, while preserving the future innovation headroom that's out there? What would you say that the main thing is?

Jose Kunnackal

>> I'd say that in the BI world today, and with QuickSight specifically, you are seeing a lot of interest in what we call the generative BI capabilities. And that includes Q and Scenarios. And the reason for that is that in a industry where you've traditionally had to go and click, click, click and build a dashboard, or that's for the analyst or for the business user, have to either ask a question, send an email to get a response. On both sides, AI is making a difference. For the analysts, it is helping them build those dashboards faster. Calculations as an example, were a huge thing. Even in Excel, if you try to write a calculation, it takes you a while to get it right. So having AI help you with it, having Q help you with it, that's a big deal. And for the business user, having the ability on the Q side, so I look at it as a collection of features because it's not that you use the Q&A, but not the dashboard building, but those capabilities are generally super exciting.

>> Okay. I'll rephrase. Outcome, what's the number one outcome they want? Just ease of use?

Jose Kunnackal

>> Faster insights today, right? That that's the number one. Because traditionally, many of these things have taken very long, so nobody likes to wait two weeks for a response.

>> And to your other point, no one wants 10 more dashboards.

Jose Kunnackal

>> Nobody wants 10 more dashboards.

>> So you have dashboards of dashboards. So I think gen AI could abstract away that.

Jose Kunnackal

>> Yeah, definitely.

>> One hot trend I want to just bounce off you and get your thoughts, is that we're seeing the words preemptive, predictive. You're starting to see even in security, preemptive minus zero days they're calling it, where they can see attacks while they're forming the pattern that's going to be a ransomware attack, be prepared. Alerting the security teams like it's coming. In business, I could see preemptive predictive data, generative helping, hey, you might have a down quarter. That's where the dream of agents comes in, where the data is actually getting smarter. It's almost reasoning moves to intelligence. What's your thoughts on that?

Jose Kunnackal

>> I'd say that that is definitely an area which has a lot of potential, and traditionally, we've used different names to talk about some of these things. So even within this product in QuickSight, you do have the notion of what we call anomaly detection. So it can look against, for example, your sales data, your customer data, and you connect it and it can keep track of, are there any interesting patterns? So exactly the same thing that you talked about in terms of security. And on the Q business side of the house, we have seen that there are things like actions, where you can take an action. So as we are bringing these different capabilities together, there definitely is the question of, can I see some of these things and can I take decisions? So in the BI market, they call it decision intelligence. So can I make decisions based on the intelligence? And then in some ways, you do want the human to actually make that decision because you don't want the system to automatically keep chopping around. So that's where, again, connecting it back to Scenarios in Q, what we've seen is that there is a lot of potential for agents to play that, but there's also the question of you don't want these systems to just operate. So Scenarios is unique in that it allows you to look at, inspect the plan and then work with the plan. So connecting all of that, lots of potential in that area.

>> I just wrote a blog post leading up to the two events that happened this month, Snowflake and Databricks, the big data conversations going on there related to this. I introduced the idea of the agent systems intelligence layer, what called the green layer, the green field, it's a layer of innovation. But this move to intelligence is going to enable the user to take more authority or have agency. You're starting to see the empowerment side of it. Any thoughts on or reaction to or commentary on the impact to the user? Because remember, whoever can see the value disrupting point to innovate, if they don't have agency or authority, they can't implement it. They know about it. So you're starting to see on the business side, huge power shift to those people who have the domain expertise, who know how to manage the data, leverage the tools and the tooling.

Jose Kunnackal

>> I think in that space, bringing together... Again, we are on this journey to bring together structured and unstructured data. And as we are bringing that together, we definitely see that more context is helping the user take better decisions. So I'd say that yes, all of that context will help, but it is something that we'll continue to think about and invest in.

>> All right. Well, I really appreciate you coming on and I have to ask, okay, we're at the midpoint. Look back between re:Invent and now, what happens? What happened? And then what's the focus going into re:Invent for you?

Jose Kunnackal

>> Well, I think re:Invent and now, we've been innovating quite a bit on the Q Business front, on the QuickSight front, on the agentic front. Heading up to re:Invent, like you said, it's a fast-moving world where I think everything's changing, the LLMs are getting more powerful. Developer tools are getting more powerful. So my hope is that we'll have a lot more exciting things for you in the next six months.

>> I'm looking forward to playing with QuickSight and getting those dashboards out to all of our people. Appreciate it, Jose. Thanks for coming on.

Jose Kunnackal

>> Of course. Thank you.

>> Okay. We are here at AWS' headquarters for our Halftime Report, two days of coverage, filming and broadcasting. I'm John Furrier, your host at theCUBE. Thanks for watching.