Getting Down to Viz-ness: How Databricks + Tableau Supports Your Organization
Summary
- Platform partnership: Databricks and Tableau together empower users of all kinds to see, understand, and act on lakehouse data securely, governed and performantly on Databricks Data and AI platform.
- Zero-copy integration: The partnership enables direct data connectivity from Databricks to Tableau through DBSQL connectors without data movement, unlocking rich analytics experiences.
- End-to-end intelligence: By combining Databricks governance with Tableau visualization capabilities, organizations deliver business intelligence and AI insights through a unified, secure analytics platform.
Getting Down to Viz-ness: How Databricks + Tableau Supports Your Organization

Talk by Aaron Frein, Senior Product Manager, Tableau (Sponsor Speaker)
The explosion of AI has helped make the enterprise data landscape more important, and complex, than ever before. Join us to learn how Databricks’ and Tableau’s platforms come together to empower users of all kinds to see, understand, and act on your data in a secure, governed, and performant way.
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
FAQs
What is the Databricks and Tableau partnership?
The partnership enables Tableau to connect directly to Databricks Data and AI platform through DBSQL connectors, allowing organizations to analyze lakehouse data with Tableau's visualization capabilities.
What are zero-copy integrations?
Zero-copy integrations allow data to move between Databricks and Salesforce's Data Cloud (and vice versa) without physical data movement, maintaining security and governance on Databricks Data and AI platform.
How can I ensure BI dashboards and AI models use identical data?
This video demonstrates maintaining data consistency between dashboards and models through unified governance on the Databricks Data and AI platform. Using a single source of truth ensures both BI teams and ML engineers work with identical definitions. The platform's lineage tracking verifies data consistency end-to-end.
Should the semantic layer live in the BI tool or platform?
This video explores where semantic layers belong, discussing trade-offs between BI tools and data platforms. The Databricks Data and AI platform enables implementing semantic layers centrally through Unity Catalog and Genie Ontology. Centralization ensures consistency across all tools accessing the data.
How do I ensure agents only see data users are authorized for?
This video covers implementing row-level and column-level security ensuring AI agents respect data access policies on the Databricks Data and AI platform. Unity Catalog enforces permissions consistently regardless of tool. The video emphasizes that security controls apply equally to agents and humans.
Full transcript
All right, everybody. We're going to get going. So, I think headsets are the way to go. Um, it's going to be a little more uh dynamic than I had initially intended just because I don't have the monitor here, so I'll be wandering a little bit just to make sure that I'm hitting on every point in the slides. But first, I just want to say thank you for joining me. I know there's a lot going on today. There's so much to do in the expo hall and I really appreciate you taking the time out to spend 20 minutes with me to get down to business and talk a little bit about how Tableau plus data bricks can really unlock and support your data culture within your organization.
So before I get going. I'm a product manager with Tableau. I've been with Tableau for about seven years and my main focus has been traditionally on end to-end analytics experiences and I'm really excited to start working more with the data bricks team to help bring these two platforms better together. So the first thing that I want to say is just if you've been a Salesforce customer before or been to a Salesforce conference you've seen this slide.
This is our forward-looking statement slide. Most of what I'm talking about today is live in market. So this is a little bit less relevant, but this is just a quick reminder that everything that we share with you, we share in trust and if you should not make any buying decisions based on things that we talk about that are not released yet. So with that, I want to talk just a little bit about the Salesforce Salesforce plus data bricks partner vision which t of which Tableau is a big part, but it's not limited to this.
So the three pillars of this partnership today are our zero copy integrations with data bicks which allows you to bring your data from data bicks into Salesforce's data cloud and vice versa without ever having to move that data. We also have bring your own LLM and agent models or sorry models and agents that allow you to empower the experiences within Salesforce with what you have already invested in in data bricks. But for the most part, what I'm going to be sharing with you is how you can analyze your lakehouse with the power of Tableau.
So at its core, the data bricks partnership with Tableau is founded on the rich connectivity that we have between our two platforms. There are four major features that relate to this connectivity story. To start with, we have our DBSQL connector. This is jointly owned with our partners at data bricks. So we bring this to our customers together to make sure that it is the best possible version of what it could be. On top of that, we have a delta sharing connector available in the Tableau exchange.
This is for those use cases where you want to share data from one Unity catalog to other consumers that might not be part of your governance model internally through the delta sharing protocol. Third, we have a new exploring Tableau integration that we'll be getting into in a little bit. And last, we released in our 25. 1 release the ability to write back to data bricks from Tableau Prep. If you've ever used our low and no code prep engine, this says coming soon right now because it's available in our on-prem desktop version, but we're still working on getting it enabled for Tableau Cloud.
With that, I want to talk a little bit about the Tableau platform and what it is that we at Tableau believe is the future of analytics with our platform. And this is all about how Tableau can continue to be your ultimate ally in accelerating your journey and your organization's journey into using data and AI to make better business decisions and better decisions for your customers and your employees. This can run the gamut with Tableau from the bottom of step in this chain where you have users who are curious about data.
They have an affinity to want to use their data. They might have spreadsheets on their t desktop. They might be sitting on top of some lightweight SQL databases and they want to get some insights out of that. And we're also but we also support you through your journey of data democratization making your organization data driven all the way through to ensuring that data lives in every corner of the organization that you have and every decision that you make.
How we want to do that is by delivering an open and extensible platform that meets you and your users where they are in the center. Tableau lives, but we can connect to data that lives anywhere. Obviously, we're at Data Bricks' conference and so our connectivity with data bricks is really important here, but we can connect to data from any hyperscaler CSVs and flat files all the way up to legacy data warehouses as well as our ongoing commitment to support supporting the commitment or the investments that you and your employees are making in your semantic layers.
So, we have new semantic layer support coming out. We have DBT and the Tableau semantic layer in there. Google looker coming soon. And we also have some interesting stuff in the works with data bricks that we'll be talking more about as time goes on. On top of that, we know that you're investing in your data catalogs and your ELT and ETL tools. We work with all of these partners to make sure that you get the most out of these two things together and can bring those insights into Tableau and also proliferate your Tableau information into those.
You can bring your AI models to extend the value of your insights and make them more predictive and powerful. And finally at the top here you can see that we are fully committed to ensuring that wherever your users are living whether it's in your business applications like Salesforce SAP into your collaboration spaces like Slack and Teams and into your workspaces whether that's the G Suite or Microsoft. We want to make sure that there are easy and seamless ways that you can take the analytics that you've created for your users and distribute them out into the places that they live.
One more view on this is how we're thinking about maintaining the trust and governance model that you're investing in at the layers that they belong. So in your data platform and data modeling layers, we are fully committed to federating and pushing down queries to make sure that the governance that you have instilled into the very framework of your data is respected all the way up into the stack. your semantic information that brings all of that business logic deep into the analytics that you're building and then bubbling that up into the analytics layer.
And we're fully committed to ensuring that your governance story is consistent up and down. So what do we do with all of this information? Why is the Tableau platform so powerful? And the way that I like to think about this is thinking about what is the data culture that I'm trying to build for my organization. And the by data culture, what I mean is how you think about how your organization works with your data. I tend to think about data culture as running on a spectrum from practitioner-led innovation.
These are those cases where many of you probably sit as customers of data bricks. This is when you have the data engineers and data scientists are leading the charge, bringing the data into your data platforms of choice doing the transformations and encoding your business logic there, but also moving quickly and iterating quickly. In these cases, you as your end users typically all have direct access to the underlying data and the models that sit on top of that.
And so enabling you to make the most out of that in your analytics story. Talking quickly makes the mouth dry out. So just thanks for your patience as I took a sip of coffee. On the other end of this spectrum of data culture, we have what I like to think of as operational analytics at enterprise scale. This is the inverse of that practitioner-led innovation. This is where you are distributing analytics to hundreds, thousands, tens of thousands of users who are trying to use that information to do their jobs in better ways.
In this case, it's heavily weighted toward the business users. Those users are not supposed to have direct access to the data that lives in your data platform that's tightly governed and you want to protect those things, not because you don't trust your end users, but because they're not aware, they don't know what they don't know. And so, it's better to keep these things siloed for the right use cases. The pref or the priority in these scenarios is also stability, performance and cost management on top of the governance.
In the middle is where Tableau has existed from our from our start, which is this idea of self-service with strong guard rails. You have the ability to develop data products within your data science and engineering teams, but then distribute those out to a broader set of users who want to build either self-service analytics or distribute analytics on top of those tightly governed models. So, how does Tableau help with that?
First, I just want to reassert this slide really quick. This is one of my favorite slides that I've seen out of our organization in the last couple of years because of how it communicates our commitment to this open and extensible experience for customers. And it's all about enabling you to standardize your governance and your stand and your data in the places where you want to live and allow you to use those within Tableau when the time comes.
So the first of the set of features that I want to talk about just to give you a little bit of an orientation about how to set up your data bricks plus Tableau environment to enable seamless self-service. This is on that first end of the spectrum the practitioner-led innovation. Generally these features are the ones that require your end users to be named users licensed users within data bricks. Data bricks is the sole governance layer of the platform where you are driving your governance.
In those cases, I mentioned earlier we just launched this new integration called exploring Tableau. For anybody who has used Tableau for the entirety of its existence, connecting to data is not complicated, but it's not easy. You have to go into our authoring experiences, whether that's desktop or the web authoring experience, and copy and paste out of data bricks your HTTP path, your instance URL. You also have to manage credentials in a way that requires people potentially moving secrets around in the open, which is not a great story.
What Explore and Tableau allows is this one click. When I am in my Unity catalog, I can launch directly into Tableau without having to do any of that copying and copying and pasting. This makes it very easy for you as a data scientist or a data engineer to pop into our visual analytics tool and start doing some visual analysis right away. It also helps prevent the risks that come with copying and pasting and all of that data transmission that just isn't necessary anymore.
All of this information is well understood on the data brick side and we've made it possible for it to move seamlessly into Tableau's platform. On top of exploring Tableau, we also have our Tableau agent. Tableau agent is the tool that we use or is our chatbot interface directly in authoring that allows you and your users to get going with our product even if you're not really sure how to navigate our user experience.
This is a aentic experience that allows you to build visualizations, calculations, and even modify things that already exist purely in natural language prompts and allows you to have that conversational experience with your data in actual human language as opposed to what we previously had was kind of like a conversation, but it required dragging and dropping pills onto shelves. And lastly, in the self-service analytics, as I mentioned earlier, with Tableau Prep, we now have the ability to write out the results of your transformations directly to your data bricks Unity catalog without any additional steps.
Uh again, this says coming soon because we're working on getting it enabled in cloud, but it's available for our on-prem servers and Tableau Prep Builder desktop applications as of the 25. 1 release earlier this year. For that middle kind of self-service with strong guard rails data culture, we have some really powerful tools around flexible governance. The first of these is our published data sources. The published data source is a curated governed data source that sits on top of your your underlying data store that provides you the ability to have an extra governance layer that doesn't have to be reflected in your Unity catalog.
So you do not need all of your users to be licensed in Unity Catalog to get value out of the data that lives there. With a published data source, you're able to bring that data model into Tableau and provide that access either managed through historically it was the PATS or the personal access tokens. We're moving to the service principle model now that's being released. And so with the publish data source, you can bring in and build a data model that any user could build analytics metrics or additional and or dashboards on top of.
Making published data sources even more powerful is our shared dimensions feature. Shared dimensions is Oh, there we go. I was like that gif was not where I expected it to be. Shared dimensions is our answer to the requirements for multiffact analysis which also enable you to build singular data models that cover the totality of a potential set of data models without having to manage every particular model for a particular use case.
And by that I mean this allows you to map multiple tables that may not have direct relationships together as part of the same data model. Put that into a published data source and every user who queries based on these multiffact published data sources are able to get what they need without you having to manage 5, 10, 15 versions depending on the specific scenario. It really simplifies the governance story and without making any cost on the performance side because these are not physical joins between the objects.
They're logical relationships that Tableau understands how to when a user is actually trying to build a visualization. We only pull data from those tables that are relevant for the specific thing that they're looking for. We don't just query every single thing every single time. Now moving from there to the large that last far right column of data cultures where it's all about operational analytics at enterprise scale.
We have a bunch of tools that you can use to in to accomplish a broader distribution to your end users. The first of these that I want to talk about is Tableau Pulse. If anybody's not familiar with Tableau Pulse, this was released, I believe, beginning of last year and is our metrics focused application that sits alongside all of your dashboards and published data sources that enable you to build KPIs to distribute to your end users.
Each metric can be embedded can either be viewed directly from Tableau or embedded in any of your business applications of choice. And we've added additional discover capabilities that provide you the opportunity to interrogate your metrics further. So if you have a question about what's going on, you can interrogate that in natural language via pulse discover. In addition, we recent we already we have had a mic or a Slack integration for quite a while.
As of last fall, we released a Microsoft Teams app that is live in the market today. This allows you to bring all of your Tableau analytics, whether those dashboards or pulsemetrics or anything else, directly into the place where your users spend the most time, which often is the chat experience that you have. And this is not limited to just viewing a dashboard. This is a living and breathing application that allows you to do those.
If you're looking at a pulse metric, you can interrogate it with those natural language queries directly within Teams. You don't have to pop out into a different application. It's really powerful and we're really excited to see our customers responding to it. Lastly, just a couple of features that I wanted to mention around how you Tableau can help you manage the performance and cost of your analytics as you scale them out.
This goes for all of those use cases, but is particularly relevant for that tightly governed operation or enterprise scale distribution. The first of these is always when architecting your data models, the more that you can push down into data bricks, the more performant it's going to be. When you write calculations in Tableau, that's incredibly powerful. It allows you to get new information going and building them into your visualizations directly in the flow of your analysis.
But that's all stuff that we have to compute in memory. And so once you establish the calculations that really matter to you and your end users, the more you can push those down into your underlying data model and data bricks, the faster that's respon those responses are going to be because they're all managed within data bricks compute environment. Similarly, whenever you can, you still want to even though we have those shared that shared dimensions option, even though we have published data sources, these are governance tools and they might come with a performance tax.
So, when thinking about the performance of your analytics, the more you can materialize those things into data bricks, the faster it's going to be. On the flip side is when you need to control cost and that's where our hyperextract engine is so very powerful. This allows you to This is a a proprietary database technology that is highly optimized for analytics. It is a powerful tool for reducing those high volume queries.
When you have a dashboard that hundreds or thousands of users are loading, you don't necessarily want to send the exact same query hundreds or thousands of times. This is where hyperextracts are a very powerful tool for you to reduce those queries that have to go back to data bricks. They also h we also provide an open and public API that allow you to build and read hyper extracts on the fly. So one of the things that we're working on with and with the data bricks team is exploring the possibility of enabling orchestration within the data bricks platform.
Specifically, as of today, you could as part of one of your workflows when you build a snapshot of new data, build a hyper extract as part of that job and then push it into Tableau instead of doing that and having to manage all of the like background or scheduling that is currently or the legacy model that we have there. So, that is possible with our APIs today and we're starting to explore what productizing those experiences could look like.
The last thing that's worth that I want to highlight here is if you're familiar with Tableau Exchange, we have a concept of what we call dashboard accelerators. And a dashboard accelerator is a pre-baked dashboard and work Tableau workbook that you can deploy into your environment and plug directly into your data so that you don't have to build this yourself. This sits directly on top of your Unity catalog spend tables within data bricks and it allows you to visualize when you're consuming the most like where your spend is going, which warehouses and SKs that's associated with, what time of day, so you can optimize your compute environment to support those needs.
And this is directly available in the exchange today. And that is my time. So thank you so much. Please remember to complete your survey. I'll be here and at the Tableau booth or the Salesforce Tableau booth over in that part of the expo hall if you want to chat more.
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.