Skip to main content

AI-Powered Marketing Data Management: Solving the Dirty Data Problem with Databricks

Summary

  • AI-driven data quality: Acxiom's platform uses anomaly detection and ML-driven transformations to clean dirty data (incomplete, inconsistent, inaccurate), improving campaign effectiveness and AI agent accuracy.
  • Privacy-first governance: Built on Delta Lake and Unity Catalog, the platform ensures GDPR, CCPA, and HIPAA compliance through role-based access, encryption key management, and fine-grained controls.
  • Performance at scale: The multi-tenant architecture on Databricks delivers 75% faster data processing while maintaining secure governance and enabling marketing teams to trust data for decisioning.

AI-Powered Marketing Data Management: Solving the Dirty Data Problem with Databricks

Watch: AI-Powered Marketing Data Management: Solving the Dirty Data Problem with Databricks
Talk by Ankur Jain, Chief Cloud & Data Modernization Officer, Acxiom ; Steven Kostrzewski, VP, Cloud Product Management & UX, Acxiom
Marketing teams struggle with ‘dirty data’ — incomplete, inconsistent, and inaccurate information that limits campaign effectiveness and reduces the accuracy of AI agents. Our AI-powered marketing data management platform, built on Databricks, solves this with anomaly detection, ML-driven transformations and the built-in Acxiom Referential Real ID Graph with Data Hygiene. We’ll showcase how Delta Lake, Unity Catalog and DLT power our multi-tenant architecture, enabling secure governance and 75% faster data processing. Our privacy-first design ensures compliance with GDPR, CCPA and HIPAA through role-based access, encryption key management and fine-grained data controls. Join us for a live demo and Q&A, where we’ll share real-world results and lessons learned in building a scalable, AI-driven marketing data solution with Databricks.
Here’s more to explore:
Unified and open governance for data and AI: https://www.databricks.com/product/unity-catalog
See all the product announcements from Data + AI Summit: https://www.databricks.com/events/dataaisummit-2025-announcements

Chapters

FAQs

What is the dirty data problem in marketing?

Data is often incomplete, inconsistent, or inaccurate across sources, limiting campaign effectiveness and reducing the accuracy of AI agents and personalization engines.

How does Acxiom's platform improve data quality?

Anomaly detection identifies outliers and errors, ML-driven transformations normalize and standardize data, and the Acxiom Real ID Graph ensures consistent customer identity across sources.

How does Acxiom ensure compliance at scale?

The platform implements role-based access controls, encryption key management, fine-grained data controls through Unity Catalog, and data lineage tracking to meet GDPR, CCPA, and HIPAA requirements.

What performance gains does Acxiom achieve on Databricks?

The multi-tenant architecture on Delta Lake and DLT delivers 75% faster data processing while maintaining secure governance, enabling marketing teams to work with clean, trusted data.

Full transcript

0:00 Introduction and overview

Do you want to kick it off? Sure. So, I guess we have some obligatory slides we need to show at every tech conference. Um, do you want to read through these or do you want us to just flip through these? Where's our data bre? I think she's in the front row. Okay. So, forward-looking statement. Uh, you're going to see some things that are generally available. You're going to see some things that are in private preview and public preview and so on.
I think the takeaway here is make all your purchasing decisions on things that are real and generally available so that we don't get in trouble and neither do you. Um, make sure that you complete your survey. If you haven't already downloaded your mobile app, do that after the presentation so that you're watching what we're saying and not fiddling with your mobile app. But make sure that you're filling out a survey afterwards.
Super helpful to get feedback on how we're doing what we presented and whether it was valuable to all of you. So, now that all of that stuff is out of the way, um. There won't be a spelling test, I promise. Um, I'm VP of product management at Axiom and this is my counterpart. I am chief cloud and data modernization officer at Axiom. I lead the engineering for both products as well as client solutions at Axiom. So, what we're going to do today is we're going to talk to you all a little bit about how um what Axiom is, what we do, what our industry is so that we can kind of ground everybody on our business.
I'm going to take you through that journey really quick in the beginning. I'm going to turn it over to my friend who is going to go through data bricks use inside of our architecture across our products and solutions and then we'll wrap it up with a live demo at the end um if the demo gods are in our favor. Um if not we'll do a little recording and show you exactly how things work inside of the data bricks console. Um if you have any questions fold them till the end.
We'll save time for those. Okay. So who is Axiom? Um, Axiom is part of a larger advertising holding company named Interpublic Group and our whole business is on data. We've been in the data industry for the last 60 plus years. Uh, we have our own proprietary identity graph data products and those are used across a number of clients and industries in order to help businesses find new customers. And so the core of all of the things that we do with technology really relates to getting identity resolution correct.

1:47 Acxiom company profile

When we talk about identity resolution, there are a couple of things. There's identity resolution that you might be doing with string matching within your own first-party data in your own CRM. For example, our identity resolution is a much more robust than that. And this is really around how we stitch together privacy compliant data to come with an identity spine that is used across all of our client solutions. So there's really a lot of business benefits here uh around identity resolution.
There'll be copies of these slides afterwards, but really the key is around how do you connect the data that you have about your customer to data that uh can help you reach additional customers. So you might have a whole bunch of information about people that buy and serve and uh you market to. This is really around how do you extend the reach of all the data that you have and connect to other prospects across your markets.
And that is what our real ID and our data technology does. And data bricks is really at the core of how all of this works. And so on the left you might have your information about your particular customer and you might have a whole bunch of information that customer has provided to you or that you've inferred. And on the right is all of the extended set of information that is available about that customer and prospect.
And it's super important that this data stays organized, governed and protected. And so that's what we're going to get to in the next section. Now when we talk about identity and data management and data governance, we have integrations and partners across the whole advertising and marketing ecosystem. You can see a whole bunch of source providers and data providers on the left as well as all the paid media and advertising.
So again, just to ground you in what we do for a business um and how we connect value to our clients and it's not just about all the first-party data, but it's really about all the enrichment data, lot of data. So that's why governance and that's why the use of data bricks is so critical to our foundation because it helps us to stay compliant. It helps our clients to stay compliant and also to be able to reach out to people in a privacy compliant way across the world.
So little bit of stats, everybody's got to have a vanity slide. And so we have a huge addressable market and a lot of additional attributes that allow our clients to reach their prospects at scale. And where this comes in is into the concept of the solutions that Axum delivers for our clients. So we do a range of things around modernizing your marketing, unlocking your data's potential and integrating with prospect data, uh helping you acquire new customers, maximizing your investments, and moving beyond cookies and other identifiers to help you reach your prospects and your customers in the digital world.
Now technology-wise, all of this is connected and powered by our own internal platform that we call Interact. Interact is a purpose-built advertising platform that allows our agencies and our clients to manage their advertising campaigns from planning all the way through to measurement and execution in between. This is allowing uh all of our clients to manage their data within our agencies in a collaborative fashion.
And so today, this has been something where we've been going through and iterating and innovating. Uh, and this has been available to all of our employees around the world. And we're pretty excited that coming soon, safe harbor, uh, we'll be able to announce a little bit more about how this is going to be accessible to all of our clients and potentially to all of you later on this year. So, uh, before I turn it over to, just really important, all of the things that we design and all the things that we build and the solutions that we provide adhere to a whole range of different compliance standards.
Uh, a few of them that are up here on the screen, obviously GDPR, CCPA, and a whole bunch of other way means of protecting data are super important as well. Um, so with that, a little bit of Harbor Cruise and Axiom, and I'll turn it over to go in more depth. Thank you. Before I go, um, are there any marketers in the room? All right, we we have quite a few because you know there is definitely some marketing jargon that we may be using.

6:03 Modern data stack architecture

So just apologies in advance and if anything doesn't make sense, if any terminology doesn't come across, you know, please raise your hand, feel free to ask questions, you know, we want it, we want to make it very interactive. Um, so let enough of Axiom, let's talk about data, you know, that's why we are all here for this summit. Uh so I'm going to dive a little bit deeper to kind of give you a view of how Axiom you know all these things that was talking about resolving the identity appending third party data assets to enrich your customer profile um how do we kind of move data along how do we solve messy data problems so I'm going to give you a taste and this will go a little bit technical so um hopefully everybody kind of um you know is is is with me in these u diagrams um So I joined Axiom about a year ago and started meeting with a lot of clients you know just to kind of understand what kind of challenges when we talk about data when we talk about processing mass processing massive volumes of data sets these are some key themes you know which I kind of captured and I'm sharing this with you uh number one is modernization you know how can we build modern solutions which are cloudnative um which are serverless which kind of give you the platform from scalability and elastic elasticity standpoint.
Um, a lot of times we meet clients who are on various different platforms. Some are on Amazon, some are on Snowflake, some are on data bricks. How do you kind of stitch this whole complex ecosystem together which brings data from one platform to another in a seamless fashion, right? And there are many more. I won't drain the slide, but just wanted to give you the the taste of the challenges. I'm sure this all resonates with you as well within your organizations as as you work with your clients or as you work with your own internal customers or within with different departments within your your own company.
I'm sure a lot of these will resonate. Um now Axiom as a services company a lot of times we have to deploy these solutions in clients own environments as well right? So aligning our tech stack with their tech stack that's another uh big challenge as well. Right? So these are some of the challenges which I wanted to kind of show you the these are like basically the backdrop of how we think in terms of solving big massive data problems and this is the solution.
This is the framework that we have kind of created. uh the name is very generic modern data stack but essentially it's it's basically a combination of best of breed software tools which are available like data bricks in the market and then bringing in our products real ID infobase interact the slides that you saw um in sections how do we kind of bring it all together to create a modern marketing data stack for our customers on which we can process govern collect store you know all these different type of operations ations that we do in building a complex um you know data ecosystem and it's all powered by real ID.
Real ID think of it as how do I how do I identify my customer across various different channels. So for example I walk into a store buy a product then I go to 1800 you know this number and make a call to say okay this product is not working as you advertised and then the problem doesn't get resolved so I go to social media. So how do you know it's the same customer who is interacting with you across various different channels and that's the power we bring in this modern data stack to not only identify your customer but also add additional attributes so that you can market to them with much more intelligence.
Right? So that's basically the modern data data stack. Various different platforms are supported. And when we think about this modern data stack, it's not just about marketing because marketing is one of the function which an organization which a business kind of thinks about. If you're building a customer 360, your sales should be able to use it. Your finance should be able to use it. Your customer service should be able to use it.
It's all about monetizing these data assets across the organization. Right? So it gives you a solid foundation. you you must be hearing a lot about genai I'm sure in different various different conferences within your own organization as well and over next three days you'll hear a lot about genai as well so if you have the data right all these things will start to fall fall into place right analytics a IML all these will start to fall in place and so this becomes the foundation for all your a IML initiatives this becomes the foundation for all your data governance initiatives this becomes the foundation for all your secure data distribution across your various different departments within the organization as well.
Right? So these are basically some of the u key themes but here is the architecture diagram. I know it's a very very busy slide but this is how we think when we solve complex data problems. U towards the left you have all these different sources. talked about first party, second party, third party data set. So these are your different data assets you know which you may be acquiring different data sources which may be coming in as a structured data semistructured data unstructured data sets different frequencies in which you have to acquire them you know some could be batch some could be API based meaning you're connecting directly to those applications and pulling the data and then some could be streaming where you have let's say mobile applications which are streaming data on a real-time basis so how do you acquire all those different data sets and then in the in the middle is the store and process which is essentially where the spine of this modern data stack kind of comes in.

11:00 Data ingestion and platform integration

Giving you the ability to spin this up on any platform and then essentially giving you the capability to store that data, process the data and things like those. Then comes the consumption. Once you have processed the data, once you have harmonized your data, how do you consume it? And consumptions are various different patterns. You may be, you know, building reports, you may be serving this data for downstream applications.
You may be um you know doing data science on top of it to build some genai models and then finally you activate this data right once you have made sense of this data sets then how do you activate this data across various different channels this could be media this could be email this could be direct mail so various different types you know for those marketers who are in the room they may resonate marketing activation data activation these are basically the common themes so high level if you take away anything this is basically The complex platform the modern data platform that we have built which is all powered by data bricks can be deployed on any cloud platform whether it's Google Amazon Azure we support all these major platforms plus governance is baked in plus you know security is baked in plus privacy is baked in so let me walk you through some key themes here u what this architecture actually enables and maybe couple of examples of how we have brought this to life for a couple of our clients as well Right.
So um modern ways of ingesting data um and I'll give you an example lot of times still today lot of clients send us data through SFTP you know so they send us the files we process those files we take those files and then run them through our identity and data asset thing then we push the result back again through the SFTP channel right and this is the number one problem clients are like how can I be uh how can we process this in a much more real-time fashion.
Right? So, modern ways of ingesting data is different type of frequencies but then giving you the capability to natively share. You may have heard about delta share that data bricks talks about, right? Which is zero copy data sharing mechanism. You don't have to physically you know push these files basically through how in an automated fashion you have this stream of data coming in as a trusted party for you to consume.
Right? So these are the type of mechanisms that we deploy in in modern ways of ingesting our data. Um I talked about various different platforms right Google, Azure, um GCP, AWS all these basically these are basically um the platforms that we support and data brick support and how we can bring up you know this platform across various different um for various different clients. Um the other theme that we are hearing from our clients is that clients are not comfortable sending their first party data.
First party data is the data that they capture about their own customers right because there is sensitivity around it. It's PII and especially after cookie deprecation or thinks about cookie deprecation clients are also thinking their own first party data is the gold mine. Why should they send it to somebody else? That data should reside within their own four walls. Right? So how do we basically build this solution? How do we take this framework and deploy in our own clients environments?
I'm sure you must be thinking about the same thing in your environments as well that how can I deploy this solution across platform. So this this framework gives us the capability to integrate this solution to build this framework across various different platforms. The third thing I'm sure you may have heard about composible architectures a lot. You know this is a terminology lot of different platforms are kind of using um and there are a couple of topics you know which um comes to mind when we talk about composibility.
One is like that zero copy data sharing right how do we minimize the making redundant copies of the data in many different places for example CDP platforms um leading platforms like Adobe Salesforce action IQ um you may have a customer data warehouse and then you deploy a CDP platform traditional ways you copy the data from your warehouse into the CDP right so many different copies of customer data that you are creating composible architecture is store it once use it many times right so how these CDP platforms can come and read that data from platforms like data bricks which may be running on AWS so that you have your customer warehouse which gives you the insights about who these customers are and then how you build those campaigns and activate those campaigns through a CDP platform so that's the composability aspect right so this enables this platform this framework enables that um the other thing I want to bring up here is clean room.
How many of you have heard about this concept called clean room? I'm sure marketers have but those of you who haven't. Clean room is essentially a way of sharing data in a privacy safe manner. Meaning if I'm sharing data about myself, I don't want don't want a third party or a partner to know about the PII about the customer, right? So how do you anonymize that customer and then how do you share it in a compliant way?
So that's basically what this clean room is. So you will be hearing a lot about um clean rooms from data bricks in this summit. Uh but this is essentially how brands you know launch co-arketing efforts um you know with their partners using clean room and sharing data in a privacy safe manner right so this is another concept um that we have kept in mind when we talk about processing data. So let me give you an example of um a big large global retailer you know where we deployed this whole modern data stack um leveraging data bricks as the underlying uh platform and um there were two goals you know from you know from this use case.

17:25 Client case studies

Number one was how do I acquire new customers, right? Um acquiring new customers meaning how do I market to new customers so that I can attract them to buy products, you know, from our brand, right? So this customer acquisition is a very marketing ccentric term and just wanted to kind of give you an idea of when we say acquire new customers meaning they are trying to acquire new customers who they can market you know uh for their products and then boosting retention meaning how do I retain my existing customers?
How do I deliver them personalized experiences or how do I become relevant to them when they are thinking about certain products, certain services? So customer retention that's another big topic when CMOs talk about so customer acquisition and customer retention very classic problems. So this is where you know Axiom kind of deployed this modern data stack powered by data bricks and I'll show you the architecture. This is a very high level architecture.
Um essentially uh client was like we don't want to minimize we don't want to send duplicate copies we don't want to be in the business of shipping data through SFTP so what we enabled here is delta share client is shipping us the data client is sending us the data through delta share which then comes to our platform which is running on AWS client is on Azure so you see this cross crossplatform data sharing which is happening and then once the data comes to Axiom we perform all our hygiene, all our identity resolution services, all our data enrichment services which I showed you in the previous architecture which also kind of talked to you all that happening in a clean room type of a fashion with governance bolted on top of it through Unity catalog and then um you saw one slide about interact.
Think of interact is where marketers actually come and start designing their audiences so that they can actually build campaigns to reach out to them. Right? So that's really where the customer acquisition and then customer retention happens because that's where you design those experiences. That's where you do that lookalike modeling from a data science perspective to see okay if these are my five customers the kind of behavior that they show how do I find 50 others right so that is called lookalike modeling and that all happens through audience console within the interact platform once you have done that then you then you can activate those campaigns across various different channels so this is an example where we deployed this for a global large retail at a global scale to bring the power of data bricks to help solve those uh business problems.
The other example I want to show you think about the same thing but in a regulated industry you know especially like healthcare there you cannot expose any PII um this all has to be governed in a certain way because the laws the compliance laws are very very strict on what you can what you cannot do. So, Axiom has a data assets you know which is called infobase which is mainly for um unregulated industries where you can use this but we have special data sets which are available which which we call as Axium health which is mainly targeted toward health healthcare companies and this is an example where we kind of brought IPG health and Axiom health together to kind of provide uh similar kind of use cases in a regulated industry which is actually all HIPPA compliant right so here we're not talking about an individual here we're talking about personas because everything is deidentified everything is anonymized so how do you kind of do the similar kind of exercise in a in a HIPPA compliant fashion so this is an example where IPG health you know is all powered by data bricks it's running on Azure platform it's helping pharmaceutical companies it's helping health care providers it's helping insurance companies to design their products and services and market to those personas you know in a very safe and secure compliant manner right so this is another case study of how we deployed this modern data stack in in various different platforms so hopefully this resonates with you that how Axm is helping our clients in solving complex data problems messy data problems crossplatform you know leveraging data bricks so with that I'll hand it over to over to you sir Great.

21:17 Governance and security

Thanks. So I think even before surveys have been submitted, the crowd favorite is any sort of architecture or architecture slide. Any diagram and the phones come out and the cameras start going. Um I'm guilty of doing the same thing. So here's one more for you as well. So when we talk about um how we manage classification, how we manage governance. Now, Unity catalog has amazing capabilities and even, you know, leading edge AI detection around automatic classification.
Well, that's great if you're just working within Unity catalog or you're just managing your data within the data bricks interface. When we talk about building products and building solutions, we have a lot of other pieces of architecture underneath the hood. And so what's really important for us is that we have a central way to define the metadata that our applications can reference in order to enforce our set of custom rules and standards across any of the underlying tech stack.
And so when our engineers came up with their solution and their architectural approach here, we wanted to leverage the capabilities inside of Unity catalog. And so what we did uh which I'll show you in a minute here through our UI and data bricks and in our application is that we came up with a set of policies and a policy contains a set of classes and roles. Roles are assigned to users and we're allowed to we're enabling the definition around which users can access different tagged types of data or different classified types of data within data bricks.
And so if I'm a user and I'm an analyst, but I'm not allowed to see any PII, let's say I'm working with credit bureau data, and that data is protected until a certain point in your activation use case, we're able to tag that data in a certain way so that no users can actually see the values of the data while they're running models and building audiences with the underlying data until they get to the point in their workflow where that data needs to be revealed to a publishing partner or a print shop.
So there are different business cases that we had to build into the classifications of data, the types of data that we wanted to be able to govern and enforce and all that gets stored within our own metadata repository that then is fed into Unity catalog via SQL so that Unity catalog can reference the definitions in that context. We also have other pieces of architecture that reference this as well. So through the concept of iceberg sharing and allowing other platforms to reference this set of metadata, we're able to define this once and have this enforced throughout data bricks or through any other client application that needs to connect to it as well.
So classification governance settings inside of Uni Catalog. I'm going to take a risk here. We're going to try and do a live demo. So let's see if this works. Okay. So, what I'm going to show you all is a prototype of something that we built. And this prototype is going to show up. There we go. Um, over on this screen. And let me make this larger. I know we're coming up on 10 minutes. If I have time to use my agent inside of our application, I might do that.

24:54 Live demo and conclusion

Um, okay. So, I'm inside of Axiom Cloud here. All of this is powered underneath the covers through data bricks. And so I have my workspaces, I have my cataloges, I have all the information below that is being referenced or controlled through our front-end application. And so through this user interface in this application, and I apologize I'm not looking at you because I'm looking down here. Make it work. Um, so I've got a set of connections.
I've got a set of data. I can go ahead and manage the set of data. this. Let's say I'm looking at some credit data and I want to define learn a little bit more about this. And so I've gone through and this has detected the layout of the data that I've loaded in. I've gone ahead and determined how I want to manage my identity resolution. Maybe I want to clean up that data a little bit with Axium's best-in-class data hygiene products.
Maybe I want to go through and this is maybe for a direct mail use case, not in this scenario. And then maybe there's some suppressions that I need to enforce at the time of activation. So in this particular use case, we're not going to use any of those, but those are certainly options that you can configure. So we come back out and the first way that we manage data quality. Yeah, we just threw in a lot of marketing jargon here.
So let me just make sure to the jargon. Let me make sure that everybody is on the same. What do you think was the most confusing? Suppression. Can you please help them understand? That is a great great question. Let's go back to suppression. So when you're marketing to people, uh let's go over here down to this list. Okay. So there are certain types of data or certain lists of data um about individuals that uh must be excluded or should be excluded depending on the use case.
Um and so what AXI maintains is a set of proprietary suppression lists and those lists are people that might be bankrupt, have tax leans, judgments. Um there might be opt out preferences that people GDPR, CCPA, those are like very common use cases where customer doesn't want their information to be used by the brand. So these are basically what we call as suppressions. The people who need to be filtered out before a campaign is launched.
Exactly. So we have a bucket here like non-marketable suppressions. So these could be people who, you know, have submitted a request through one of those regulatory bodies to not have their information. This could be a list that Axiom has curated through a whole set of optout preferences and requests directly to us or to one of our uh agencies. And so this allows our clients and services partners to be compliant when they're actually sending messages out.
So, um, if you've ever submitted a do not, you know, contact me or an unsubscribe request in an email, um, that would certainly be honored, um, through the different suppression products that we have at some point. Um, so there are other things in here as well. I talked about cleaning up data. I talked a little bit about hygiene. Was that another jargon term, you think? Probably. Probably. In the interest of time, maybe just real quick 30 seconds if you want to explain what hygiene is.
Sure. So, um, hygiene. So, let's imagine you have a bunch of data that your customers have given you. Um, but maybe the information that they entered into a website or the customer service agent entered into your CRM system. Um, maybe they miskeyed something or maybe there's some attributes that are missing from your customer's profile. So, the hygiene products here are a way to fill in the blanks and also to correct and standardize the formats of the data that you might have.
Very common example I'll give you. People move all the time, right? when you when you move from one address to another that's another type of hygiene to make sure that we are using the latest and the greatest address right so um postal department you know they they they have a process called national change of address and that's what basically Axim uses to provide the latest and the greatest address of a particular customer or a prospect and that all basically is part of the hygiene package that we do standardizing the data so that before we actually do any operation it's in a global you you know um standard uh look and feel right so that's basically what the hygiene process is all about so this is all like how we treat raw data that we acquire from various different sources whether axim acquires or whether client gives us before we do any processing on them how do we clean it up it's all happening in front of you here as part of axim cloud platform built by built on data bricks right so that's that's basically the picture that we are trying to show you yeah and so like I said earlier we have some very special things coming with our interact platform later on this year.
What I'm showing you here is really kind of our internal proof of concept and prototype that we built out for some of those use cases that you'll see later on uh this year. So we talk a little bit about dirty data and we want to talk about making sure that data doesn't get into your marketing environment and get consumed downstream. And so one of the things we have here with a partner of data bricks of ours which I can tell you more about after the session um is that we leverage their AI powered anomaly detection.
And so here when I come down I'm able to see you know information around uh unique values. I'm able to see if there are you know data profiling statistics that might be interesting. And then up on the top here uh just in the interest of time we're not going to let it process but I have the ability to define custom rules. So let's say that there are certain values that we know are garbage data you know me atme. com or you know all the different generic values of data that might be showing up as repeat values that you want to exclude from your identity resolution rules or your customer profiles and so on.
So you have the ability to set that here so that that data gets removed or handled as the way your business rules should be handled downstream. Um and so that's really the next piece here. And then as far as the uh information, I'm just going to do this really quick. So all the roles and permissions and the data classification pieces that I was showing you earlier in that architecture are all controlled here within this UI and all that gets fed back into the data bricks back end.
So that when users are accessing data, all that is done in a privacy compliant way um based on the client's requirements. So uh I think we're about four minutes out. I'm gonna pull this off of the screen and slide back over to you the presentation for a minute because we have one very very important slide to show here at the end right before we do questions and that is

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.