Skip to main content

What's New and What's Next: Building Impactful AI/BI Dashboards

Summary

  • Dashboard capabilities: Latest Databricks AI/BI Dashboards features include key drivers analysis, cross-filtering, and seamless Genie integration for interactive exploration.
  • Performance optimization: Data authors learn techniques to tailor visualizations for different audiences while optimizing cost and performance at scale.
  • Unified analytics: AI/BI Dashboards and Genie bring together data exploration and self-service analytics on the Databricks Data and AI platform.

What's New and What's Next: Building Impactful AI/BI Dashboards

Watch: What's New and What's Next: Building Impactful AI/BI Dashboards
Talk by Eason Gao, Product Manager, Databricks ; Rory Jacobs, Sr. Manager, Engineering, Databricks
Ready to take your AI/BI dashboards to the next level? this video dives into the latest capabilities in Databricks AI/BI Dashboards and how to maximize impact across your organization. Learn how data authors can tailor visualizations for different audiences, optimize performance and seamlessly integrate with Genie for a unified analytics experience. We’ll also share practical tips on how business users and data teams can better collaborate — ensuring insights are accessible, actionable and aligned to business goals.
Intelligent analytics for everyone: https://www.databricks.com/product/business-intelligence
BI meets AI: https://www.databricks.com/resources/ebook/business-intelligence-meets-ai
Explore the only unified platform for agent systems: https://www.databricks.com/product/artificial-intelligence
See all the product announcements from Data + AI Summit: https://www.databricks.com/events/dataaisummit-2025-announcements

Chapters

FAQs

What are key drivers and how do I use them in dashboards?

Key drivers automatically identify the factors most influencing a metric, enabling users to understand why business outcomes changed. This reduces the need for manual analysis and speeds up decision-making.

How do I optimize dashboard performance at scale?

Techniques include smart data aggregation, caching visualizations, and designing for different audience segments. The Databricks Data and AI platform enables these optimizations while keeping costs controlled.

Can I integrate Genie into my dashboards?

Yes. Genie natural language search can be embedded directly into dashboards. Users can ask questions and get answers without navigating multiple pages or creating custom reports.

How do I ensure my dashboards are accessible to different users?

Data authors can tailor visualizations for different audiences using role-based design patterns. The Databricks Data and AI platform provides governance tools to ensure users see only appropriate data.

Full transcript

0:00 Introduction to AI/BI

Right. Uh, hey everybody. Uh, thank you for bearing with us. I know it's end of this conference Thursday afternoon, but super excited for you all to be here. Uh. I'm the product manager at data bricks for AIBI dashboards. I'm joined by our engineering manager for the team. Um so between the two of us you know all your dashboard feature requests because somehow got funneled you know through our backlogs. So really excited for you all to be here today.
Um the plan is to share with you guys you know a little bit more detail on what you saw on the main stage. So things you saw like key drivers cross filtering right that you saw in the keynote this morning. We'll get into a little more detail. Um we're going to talk about performance. Um and then we're definitely going to do a demo and answer some of your questions. Um so legal had us put up a forward-looking statement.
So I'm going to pause there for two seconds. Um, and please know that there's going to be a survey throughout this entire conference. You've gone to other sessions. I think they've probably shown you this already. Uh, but please do that. Um, but yeah, let's get going. Um, point today is to talk about what's new and what's next in the IBI dashboards. Um, like I mentioned, the plan is a quick intro on dashboards in general, a look at what's new and what we've been doing in the past year.
Um, talking about optimizing dashboard performance because, you know, as we get to a more mature position in dashboards, you're going to want to be sharing these to more and more users. So making sure cost doesn't scale like linearly with usage by actually finding more you know optimizations is going to be really important. I'm going to demo everything look at the road map and then of course take your questions. Okay so brief refresher on AIBI.
Um I think I'll choose how much time I spend on this depending on kind of how we're feeling in the room. So um can I get a show of hands if you use AIBI yourself? Oh so a good number of us. Okay. Um, and are you guys in so how many of you guys are like prototyping and kind of considering AI at this stage? Okay. Okay. I see a good amount. Okay. So, um, I won't spend too much time on this then, but like you saw this morning, AIBI is a product line of two products, Dashboards and Genie.

1:33 Refresher on AIBI value

We're bringing them closer together. And you know the reason why we made this investment in the first place and I think alluded to this in the keynote is you know obviously we have great BI partners in data bricks like your Tableau your lookers you know and I'm sure a lot of your organizations use those tools too. Um but we heard demand for a new native BI solution because you know we heard from customers that they felt we were in a position to do three things really well.
The first one was uh people wanted to leverage you know the governance you have in UC you know the fact that you can do all of your compute management and cost management in one platform and that really reduces the complexity of your tech stack. So people said hey like my data is already in data bricks can I just really easily visualize it and build a dashboard share it out without having to go to another tool. Um and that's the first motivation for how why we built the IBI.
The second part was scaling data is just getting harder and harder. So you know data volume keeps going up but if you to get that to another tool you have to choose like am I going to direct query am I going to do an extract model but in either of those cases you're limited by some kind of performance or scale or cost right but with native data bricks you don't have any of those limitations you know you know our compute model there is no additional view and licensing fees right and you can just query as much data as you want as you would in a notebook and in a SQL editor and finally you know this wave of AI I think we're all seeing even of our own experience using chatbots.
You know, AI gets better with the more context it has. And a lot of that context lives in dashboards. It lives in notebooks. It lives how you do data engineering and it lives in kind of how even how you do rag models for example, right? So, um, a lot of those other proprietary BI tools only have context on the data that you give it, but it doesn't have your entire data estate. So, AIBI and Genie specifically is in a better position to leverage all that data.
So, that's why we invested in AIBI as well. Okay. So, you know, when you put all those pieces together, one success story we saw last year was that wine shipping is one of the many organizations that have fully adopted AIBI as their primary BI tool. So, 75% of their entire staff now logs into data bricks, you know, on a regular basis and make datadriven decisions. Um, and we're really excited about this trend. We know not all customers are kind of at this stage yet.
You know, a lot of you guys are probably in a transition state where you're going from, you know, like a PowerBI or a Tableau and you're like, maybe I'll use some data bricks, some PowerBI, maybe I'll use all data bricks in the future. What does that migration look like? Um, but you know, when you get to the end game and you have all of your, you know, insights in AIBI, you really do see a lot of benefits. So, with GA last year, um, and we've heard since GA that you guys asked for three things and a lot of this was motivated by, hey, like I want to use you guys more.

4:29 Roadmap and goals

Like, I'm bought in on the cost thing. I'm bought in on, you know, the unified governance, but you don't have through or you don't have multi- page or I can't send subscriptions in this way that I need. So, it's a non-starter. So, more coverage of conventional BI features. The first ask that we heard. Now, the other one is people asked for a more efficient author and consumption experience. You know, we have an opportunity here to build BI from the ground up.
We don't have to do it the way it was built 20 years ago. And that meant you know you know using more efficient authorship patterns like AIdriven SQL writing AIdriven like viz viz building right oneclick forecasting like you saw on the main stage today to really make your authors a lot more efficient uh but also a lot more kind of skilled because they don't need to go and understand all the details of data science for example to do forecasting.
Um, and it's the same for the consumption experience, right? So, traditionally, you built a dashboard, somebody had a follow-up question, they went back to the analyst, they add a new panel, but Genie allows you to have this like kind of well-rounded closing the loop within the data product itself. Um, and finally, you know, just more options to share of your organization. And that was the motivation for data bricks form, right?
If you have everything you've done in data bricks, but you know your business users aren't comfortable coming into the workspace or you as an admin aren't comfortable bringing your business users into the workspace because they're going to write crazy queries. Um, people ask for a better way to manage that. Um, so today we're going to talk about how the work we're doing is going to address all three of these. But ultimately the vision of AIBI is we want to shift you guys from a model where anytime somebody has a question you built a new dashboard to now actually curating dashboards and genies by topic or by team and then this becomes like a long living asset where you can iterate on you can improve and you don't have this like data sprawl in your organizations anymore where you have like tons of deprecated dashboards and everyone's always building new stuff.
Okay, so what's new? Um let's talk about that plethora of data cap dashboard capabilities first. Um the first one and I'm actually curious at this point because um the way we get these feature requests is that a lot of our customers talk to our account teams and the account teams literally slack and I and go hey like do you guys have calculated measures? No, when's that coming? And then we start keeping a log of what are the most common asks?

6:27 New dashboard features

And that informs our road map. So um did any of you guys ask for calculated measures and dimensions? Okay. Awesome. Okay. I see hands. Okay. Yeah. So we built it. Um calculated dimensions are now available generally available in dashboards. Um so you know real real quick is you know within the data tab now and you'll see this in the demo. You can create custom definitions for just aggregated metrics or entire new fields.
Um so you can do this after you do the querying. And what this really enables is one it's just a lot more flexible in terms of how you express data but there's a huge performance benefit as well because every time you have a dashboard data set that is something that we have to execute when you load that dashboard. So before if you wanted an average and then like an aggregation but you also wanted to filter by individual rows those are separate data sets you have to do because once you aggregated you lost all the detail.
Um so with calculated meions dimensions uh you can do that post-processed so that you get a lot fewer data sets and that also has cross filtering capabilities as well. So you know really simple example is if you want to do profit margin you no longer have to create a new data set and write custom SQL just to calculate that. You can just have an orders table and you can write the formula for profit margin and then that is aggregated calculated for you at the end of a data set.
Um, we also did multi-page layout with filtering that applies across all of them. So, so did anybody ask for multi-page or or wanted multi-page since G? Sweet. Okay, so multi-page is also available. Uh, the key thing to note here is that right now we have a 10page limit on multi-page. Um, and then a 100 visualizations within each of those pages. Um the other thing to note is that there is a performance benefit here in multi-page as well because you know we saw an opportunity to b basically help you only conditionally render what's needed uh for the publish view.
Um so you know a common example is you have basic queries to do you know general high level KPIs but then you have an expensive query that digs into details right so you can put that expensive query in a details tab have everybody land on the summary tab first so until somebody goes into the details tab you you'll never have to execute that query so we really recommend actually breaking up your dashboards into multiple pages um because then you can conditionally control kind of how much you how much compute you use.
Um we also did new visualizations. Uh so you know a couple of key ones like maps were not available last year this time but now you've got point you've got coral pleth we've introduced box plot to show you better distributions. We've introduced seni to show you kind of how data flows between stages. We did forecasting um and right now the team is working on funnel. So there's definitely still a long list as we're being worked on.
Um so know that new visualizations are always coming as well. Um that's just kind of what we had time to cover in detail when kind of one per slide, but there there's so much more coming, right? Like just to name some of the things on this list. You know, tagging and certification are on the way. You know, version controlling with git folders for you guys who want to version control and kind of manage dashboards via CI/CD is available in public preview now.
You know, you can now restore from the last published version. Um and you know, one thing we're always surprised by is nobody really asks for undo redo. Um, I don't know if that's like something people actually just don't need as much anymore. Um, but we personally really want an undo redo. So, undo redo is in development now and that should be out in the next month or so. So, you can finally like command Z undo some changes on the canvas.
But today already you can undo a dashboard to the last published version. So, if you don't like the local changes you're making, but you want to just go back to what your viewers see, there's already an easy button for you to do that. Um, actually one one more note on this is that we really love for you guys to keep working with us and telling us what you still want to see from dashboards, right? Like if you're looking at, you know, incorporating dashboards into your workflows, but you're currently doing another tool like you want to do it in the IBI but you can't because you needed maps or you need a multi-page, talk to your account teams, get connected with us, um, get that feedback in and it really does inform our road map every single quarter.
Okay. Um, so that's the kind of the plethora of BI capabilities. But the other thing we wanted to talk about is integrating AI. Um so the first thing is last year we said you've got dashboards, you've got genie separate surfaces but we want to bring them together so you can use them in one place. We did do that this year. So in this gif here you can see genie is now natively embedded into dashboards and there's two ways to do this.

10:38 Integrating AI assistance

Uh one is when you publish a dashboard now you can just say hey spin up a genie based on the context of this dashboard and we automatically do that for you and that's what you see here. Now obviously if you have existing genie spaces and you've done a ton of work you know going through making sure the instructions are right you're happy with the performance of that space you can also connect an existing genie space and that will be what's used to populate this window so you know this is the starting version of this feature and there's going to be more integrations coming you can imagine you know I'm going to ask a query I'm going to generate a new chart and put it onto my canvas those flows are all things that we're investigating as well but today already when you look at a dashboard you can invoke a genie ask follow-up questions and I'll be able to look at things like oh like these These are your visualizations.
These are the aggregations you have on those visualizations. So these must be what you're referring to when you ask a question in this way. Um that's with a viewer. But we also wanted to make your authors just a lot more efficient. Um so there's two things that we put together to give you key drivers and forecasting. So the first thing is you know similar with notebooks and SQL editor. Um it's been a couple of years now where you've had the ability to use the assistant to help you write queries.
You can do that in dashboards as well, right? So you can expect those benefits across the platform. Uh but also on the canvas side, we've given you natural language and the ability to just shortcut clicking through like this field on X, this field on Y, aggregate by week over here, color by this field. Um that it can all be done in English. So um you can now do both of those pieces entirely with natural language to help speed up and kind of reduce the learning curve of the tool.
Uh, but the really cool thing about these two is that if you have the ability to spin up new SQL and you also have the ability to generate new visualizations based off of AI, uh, you can do one-click visualization generation now. So, if you know you want to forecast something, we will both give you the new SQL and also spin up the new line chart at the same time. So, you have one-click generation of that insight by leveraging both of these mechanisms under the hood.

12:53 Sharing and embedding

Um, and finally, we've unveiled new paths for sharing. Um so there's two things that we heard that make sharing difficult in AIBI last year. Uh the first one was my business users work in things like Google Drive today where my entire organization exists in the search dialogue right so like I can find my boss I can find my skip I can find my team all my mailing lists are here so if I write a doc I can share it with anybody but they come into data bricks and they're like I can't find anybody here you know I feel like I'm the only user on an island in data bricks because my admin hasn't provisioned the rest of the organization so I feel stuck.
Um so what we did there is that there's a new there's a new ended uh mechanism called automatic identity uh that is now available starting with Azure where you basically connect data bricks to your Azure ADP um and then all of your users who exist in Azure and all those groups are now available to be shared to from within a dashboard share dialogue. So now you can finally find everybody in your organization and once you share with uh that dashboard with them what they get is an email and an invite to view that dashboard and if they're not already provisioned into data bricks they'll be just in time provision when they come in and there'll be a provision at the account level and then the admin can further put them into groups right you know and do all that management but we've kind of eliminated that barrier for for sharing.
Um the other part is like hey like if I can't bring my end users into data bricks because the workspace is scary like they're going to run up cost because they don't know what these buttons do. it doesn't feel like a BI tool like there have viewer licenses in Tableau they have this great explorer experience right um so what we did is that there's a new consumer entitlement now in the workspace and how it works is you provision a user into database and when you add them to a workspace instead of saying hey now you can use SQL you can say hey you are just a consumer in this workspace and what that practically looks like is the workspace is going to be super clean so you get only dashboards and genie on the left you get your list pages for those two assets and the only things available in that list are the things that are shared with Um, so it's a much cleaner experience and it's a first step towards data bricks one which is on the main stage which is not available yet but this consumer entitlement already is and one will directly kind of piggyback off of the mechanism of this consumer role.
So consumer users will get the experience of one when that is released. Um, so you can already start playing with this consumer role that's generally available. Um, but this these are kind of hairy topics like identity is never like quite as easy as it sounds like on stage. So there's a dedicated session this afternoon to talk about that. So I left um you know a screenshot there for you guys to know. Um and yeah like I said you know that mechanism consumer piggybacks with data bricks one once available.
Um but today you can already something you can try and you know we obviously love feedback on that mechanism as well. Um the last part of sharing is embedding. Um, so embedding was actually our single most hotly requested feature when we G8. Um, and embedding at first what we heard from you guys is I want to wrap this in an iframe with some sort of other component and put it into a custom application. And we said, okay, here's an iframe, right?
Super easy. You from the share dog, you grab that iframe, you put it into a confluence, you put it into a wiki, um, and it works. And that worked for some users. But then as soon as we released that, we heard, wait, wait, hold on. What I meant was I want to put dashboards in front of my vendors on my own customers because I have this really interesting data that I really want to share and be kind of like show off like how more modern that my tool is against my competitors by putting dashboards and analytics like natively into my vendor portals.
Um but the key thing there was those users are not in your IDP and you don't want to provision all of those users into data bricks and suddenly give them access to say all of your tables, right? Um, so what we did there is that now there's a private preview for what we're calling embedding with app delegated authentication and real simple it just means you can provision a service principle. You authenticate your application against data bricks using that SP and at that point um all of your kind of end customers can impersonate that SP to load that dashboard.
So they no longer need data access and they no longer need um to have data bricks accounts. Um but you know you're probably thinking but I don't want vendor A to see data from vendor B, right? So that's obviously something that's really important. So as part of this mechanism, we also allowed you to pass key value pairs. Essentially, you can parameterize the request within your applications based on the viewer looking at that dashboard.
Um, and then you get the ability to filter the data depending on who's viewing. Now in both of these cases, the same caching mechanisms that we have performance uh is applied and also you know it's just super easy to kind of look at, you know, all the lineage auditing right is all included as well. So that's a really quick look at kind of the latest things we've been doing. Uh but we know performance is also top of mind for a lot of you.
So for that I'm going to hand it. Yeah. All right. So uh performance best practices. I had 10 tips to start with. I had to boil it down to four to fit. Um so if you have more questions after this, send them to EN. No, just kidding. Come ask me afterwards or ask your account rep. There are four basic things I want to go over. Um, we want to leverage client side compute um to avoid round trips to the server to speed things up.

17:21 Performance best practices

Uh, we prefer dashboard filters to parameters. So, don't use parameters for filters. And I have slides on each of these in a second. Um, you want to use custom calculations to prevent data set sprawl. And you want to trim columns in your data set to reduce network load. And let's go through each of those really quick. All right. So, client side compute. You want to keep your data sets at the moment under 100,000 rows and 100 megabytes in size.
If you filter on result sets that small, everything occurs already in the browser. So of course you have to run them the first time to load them into the browser. But now when users change their filter dropdowns, zero queries are executed. There's zero load on the warehouse. There's zero cost. If they're bigger than 100k rows or 100 megabytes, you get multisecond interactive latency, right? it has to go run the par run parameter changes the code for example or it's too big it must run the query on the server side so of course that's going to be slow um if you keep it under of course you're going to get that sub-second latency and kind of as a follow-up point to that this is why this slide is here prefer filters to parameters filtering happens client side the data set is small the parameters always run server side you can see the difference in the example here on the left hand side maybe it's might be your right I don't know selectab table where parameter that hits the server every time because it changes the SQL itself so of course we have to run it on the server if you just add a filter control to the dashboard it happens client side if it's small yeah I just I just want to say one thing to emphasize what what you just said there because you know on stage when you look at AIBI demos there's always this like magical like pick a couple drop downs and the dashboard is instantly loading Right?
And I've been on calls with some customers saying like, "That doesn't happen for me." Like, "Well, why isn't it happening?" Well, it is those two things. It is either your data is too big, so it's over 100,000 rows or 100 megabytes, or you're using parameters. And if you can get out of those scenarios, you get the same magical kind of like interactive feedback that you get that you see in all the demos that we do. And I've definitely gotten tickets that I've routed to the team where they're like, "It was fast and now it's slow."
And invariably, they tripped over one of these limits. It's not very hard to go from inside the limit to outside the limit. Yeah, super common. Um, we are working, you didn't list it on the road map, I think, to increase that 100k limit rather dramatically. Probably like 10x or remove it alto together. We shall see. Um, the data limit will probably remain for now. All right. The second thing I've heard a lot of complaints about is data set sprawl.
So, let's suppose you wanted to create a profit margin. That's really easy. It's, you know, revenue minus cost divided by revenue. And then you say, I need to segment it by region or state or industry. And the first thing people do is they start cloning their data set and making one for each. Um, that's lots of data sets to build and maintain. But instead, you can create a custom calculation. And now you can segment by those things.
So one data set can now power many widgets, which also works well with cross filtering. I'll show that in the demo in a few seconds. And of course, fewer data sets, fewer queries, faster download, faster dashboards. And the last one, I've seen this one a fair bit. You want to trim your columns down. Fewer bytes over the wire means faster loads. Um, we had a ticket the other day, someone did select star, like a table with like 5,000 columns.
And of course, it was super slow. And they're like, why is my dashboard unbelievably slow? It's like they're downloading an incredible amount of data. you really want to do select ABC or whatever you need from that table and just pull that. Otherwise, you're gonna have performance problems. It's sad but true, right? Everything is going over the internet. All right, so this gets the fun part where we're going to go to the demo and hopefully I can drive successfully from my laptop here and hopefully it's still connect network.

21:47 Product demonstration

It is. All right, so I've got a multi-page dashboard. So, one of the things he talked about, one of the things that we launched this year was multi-page. So, we have a summary and then detail pages. So, sales overview, we've got a region breakdown, got account details. One of the things we just launched, if I go back to the summary tab here, is we now have global filters. So, you can click the little filter icon on the left there, it opens up this panel.
I can do nice things like select the last 90 days, and I can pick the industry. I'll pick tech because most of us here are in tech. All right. And so now you've got this dashboard and you can see if we like jump to say the region breakdown those carried over. Why did that? That's odd. Maybe there's no data for those. Yep. Yeah, there was no data for that filter. But you can see there was a map there. If we clear out the tech, you can see there's a lot more data.
Right. The other thing I want to call your attention to while I'm on this slide is we've built a Sanki diagram. So that's nice. You can see I've got a filter set to the Midwest. I should be able to clear that somewhere. I wonder where that's coming from. Right. And so now in your Sanki diagram, Midwest Tech, all of these things are over here. I'm in edit mode, so uh that may be a little bit of our pain here. And you can even like narrow things down further with segmentation which is kind of nice.
There was a cross filter set. All right. So you can see all of these things here. So now you've got seni diagrams these box plots showing min and max for each region cortiles all of those good things. All right jumping back here a little bit. The second thing I think you just talked about forecasting which is quite nice. We have s sales revenue over here. You're looking at that trying to wonder what the future is. And are we just struggling here?
All right. So, we've got this nice thing. Should be able to hit this forecast button. I'm doing this live, so we'll see what happens. That went well. Oh, it's thinking. So, that should be running. And now you're getting this nice forecast. The dotted line is the forecast region. The shaded bars are your error bars. I'm going to publish this dashboard here which is if you've used it there's a kind of a draft and publish mode.
So authors are seeing the draft version. Publish is taking a snapshot ready for the viewers and you end up with the same dashboard over here for viewers. Right? And you can see all of these things, right? And so we'll jump in here. And one of the new features that we also have, we talked about using AI a second ago. And uh, of course, I jumped off to the wrong thing. Sorry, mouse errors. You can now hit explain this increase.
And it should run. Why did February have a slight bump? And it will ask Genie why it had a bump. And so it's running live. Let's see what it says. This is running over my cell phone, so we'll see how that goes. All right, it actually came back timely manner. Too bad. All right, so you can see it says, hey, look, the time period uh it went up 600K due to enterprise purchases, but that was slightly reduced by uh a decrease in performance and standard.
And you can see it gives you a nice little bar chart here showing what went up. Enterprises at the top, standards at the bottom, and it gives you more charts to kind of help you out. All right, more fun. If you haven't done this, we have a schedule. So, let me add a little filter here. Let me just add this one. Filter. All right. You can now hit schedule and we can create new schedules. Schedule. And new feature here got use default values, use filter selections.
So this is nice because we've gotten a lot of this request. Who here has had this request where they're like I want to send this dashboard to three different people and they need three different regions or three different somethings. Is that common? Maybe not. All right. Nobody. Uh so now I can create that schedule. And if I hover over the U filter icon, you can see it says changing dates there at the bottom. 90 days.
So everybody, you can now subscribe different people to different variants of this without having to duplicate the dashboard and set different default values. All right, the last thing I wanted to show here was embedding. So if you hit the share button, this is our most popular request and I'm going to show how easy it is. Like less than a minute, you can embed this dashboard in Confluence. And I'm going to do this live, too.
Um, so where' my mouse go? Bottom left corner. Yes, I see it on the bigger screen. I see. All right. So, if I hit embed dashboard, you see JavaScript. We're not doing code. We're going to do iframe. Let's just take all of this. It'll be easier. All right. That we can trim it in a second. Screw it. Okay. So, there we go. And let's go over to our Confluence page. And I have a Confluence page prepared here. Lovely. I think I got disconnected from the network.
So, I'm going to reload it at me. You can search for an iframe. And we can just simply paste the URL in here. Add some garbage at the end. I overdid that. doesn't want the full iframe. They just want the URL. Yeah, I copied a little too much. I included all the parameters. So, let's do that. All right. And have still a little too. Hopefully, that got everything. And we'll set the width to 100%. And the height to a,000 pixels.
And hopefully you should see live. It's populating and loading. Your dashboard is now in this confluence page. Update. You can see we really do have a live confluence page with multi-page with global filters. All of your data is right here. So if you really wanted to give this to your users in Confluence, that took like a minute. You're on stage with me fumbling about. So that is the last bit of things that I wanted to demo what you wanted to talk about next.
I was going to do road map. So um Okay. Y I will scroll you back over. Sounds good. Somewhere. Great. I think right here. Thank you. All right. So you saw the demo. Um and I just want to recap that you know we are seeing a lot of traction in the IBI especially from last year to this year. You know Rivian is one of the customers that you saw you know at the end of keynote there you know where they're really able to now enable you know technical and business users at the same time to get insights from their data.

29:17 Future outlook and Q&A

Um so we'd love for everybody to get there you know with Rivian with wine shipping these customers um and leverage AIBI you know in organizations. So we really believe in the mission that we do um and we want to work with all of you to understand you know what else are you trying to see from the product what would it take for you guys to get there in your organizations and realize all these unification benefits. So what's what we're currently thinking what we heard from you guys and I see is like phones already up ready to take a picture of the road map is is uh a couple of things right so first one is we heard asks around kind of a native TV monitor display mode but so today you know there isn't really an easy way to just refresh a dashboard on its own um so it's not really good for kind of a a personal display trading desk you know room uh setting so we want to give you that um drilling down and drilling through so like that will only keep getting better and you will get more capabilities, level of detail, hierarchies, you know, kind of next level of interactivity and expressibility will come.
Um, what you just saw there is sort of on an initial version of report bursting that showed where, you know, you can manually go through, pick up some values and then send that to an audience. Uh, high volume means you'll be able to upload a table where every single row is going to be a different recipient and then you can say like, hey, team A needs values A, B, and C to be set for these fields. Team B needs this. And you can have 30, 40 teams, right?
And then so you'll be able to just iterate through that list and burst dashboards to much much larger audiences. And when you up when you update those tables, we will you will also be able to update the schedule. So there's just so much less maintenance to do there. Um funnel visualization, you can already see a mockup of what that will look like. Um you know that's in the final stages of development. Um we're looking at alerts integration.
So a common thing we hear is hey like if my line goes above the certain threshold, I want to send this copy to someone who should know about it, right? um you can already do that today with how we've integrated with workflows. So today you can already you know as part of this this kind of new release um is the ability to send a dashboard burst as part of a workflow task and workflows also have kind of if conditions. So you can theoretically already say like run a query if the numbers above this versus this dashboard as a follow-up task.
It kind of gives you the full power of workflows if you want to do that. But we want to give you something native as well because I know not every user is kind of comfortable configuring workflows. they might feel more comfortable working in dashboards. Um, on the performance side, we're looking at, you know, optimized query as a native button, a performance profiler that you might already be familiar with, a SQL editor will come to dashboards.
Um, and some native performance improvements, right? So, we know 100K works, you know, most use cases, but there are genuinely cases where you do have more than 100,000 rows, but you still want the past performance. Uh so we're looking at ways you know like under the hood to implement how we're doing our caching to give you a higher limit you know without like crashing everybody's web browser. So a lot more coming on that front.
I would love feedback on these if this feels like the right direction we're going in. If there are things you guys want to see that isn't on this list. You know obviously like undo redo small things like you know usability improvements are kind of too small to be listed on slides like these but just know that our entire front end team is kind of working on that as well. Um so thank you so much. Um, thank you for all the partnership, all the feedback, and we'd love to take questions.

Learn more about the Databricks Data and AI platform.

The information provided herein is for general informational purposes only and may not reflect the most current product capabilities or configurations.