Sponsored by: Lovelytics | Rebuilding Data & Identity at Scale: Acxiom's Platform Transformation

Summary

Acxiom partnered with Lovelytics to rebuild their data and identity services as a native application layer on Databricks, standardizing identity resolution across first-, second-, and third-party data.
The architecture uses Unity Catalog, Delta Sharing, and Databricks clean rooms to enable zero-data-movement collaboration, allowing Acxiom to make its assets available in client environments without data copies.
The transformation reduced maintenance overhead, accelerated innovation for Acxiom's clients, and enabled secure campaign activation workflows for media planning, as shown in a live demonstration.

Chapters

FAQs

What did Acxiom transform and why did they need to change?

Acxiom, a 50-plus-year leader in customer data and identity, needed to make its data assets available directly within client environments without requiring data movement or migration. The transformation rebuilt their data and identity services as a native application layer on the Databricks Data and AI platform to meet modern interoperability and scale requirements.

How does the zero-data-movement architecture work for Acxiom clients?

Acxiom uses Databricks clean rooms and Delta Sharing to allow clients to access and collaborate on identity-resolved data without Acxiom copying data into client systems. This preserves data security and governance while giving clients the ability to activate Acxiom's assets in their own Databricks environments.

What is Lovelytics and what role did they play in the transformation?

Lovelytics is the largest pure-play Databricks consulting partner, with over 500 completed Databricks projects and eight-time Databricks Partner of the Year recognition. In this engagement, Lovelytics conducted the architecture assessment and designed the future-state infrastructure that standardized Acxiom's product infrastructure on Databricks.

What business outcomes did the Acxiom platform transformation deliver?

The transformation reduced maintenance overhead by standardizing product infrastructure and accelerated Acxiom's ability to deliver innovation to its clients. This video includes a live media planning workflow demonstrating campaign activation where client teams can execute against Acxiom's identity-resolved data without data movement.

Full transcript

[00:07] Thank you for joining us for this session. So, this will be the session sponsored by Lovelytics for rebuilding data and identity at scale, Acxiom's platform transformation. So, we have two speakers here with us, Ankur
[00:19] Jain from Acxiom and Jay Gable from Lovelytics. Can get started. >> Thank you.
[00:32] >> All right. Okay, well, thanks everybody for joining us this morning. Thanks for squeezing us in between the keynote and between lunch. So, starting off with just a couple of the obligatory statements here. So, everything that we mentioned here is uh um Databricks is not responsible for or
[00:46] accountable for. And remember afterwards to take your survey. So, good to go ahead and get started, Ankur? >> Yes. >> Okay, so, talking about rebuilding data and identity at scale, going through
[00:59] this journey uh with Acxiom. I'll start off with a a couple introductions about myself. So, so I'm Jay Gable. I'm with Lovelytics. If you're not familiar with Lovelytics, uh who are we? We're the largest pure play Databricks
[01:11] consulting partner. Uh by the numbers, we are 100% focused on data and AI. There's about 550 plus of us uh worldwide. We've done over 500 Databricks projects
[01:24] throughout our lifetime, so well versed in this, obviously. And we've been a Databricks partner of the year eight times now. And lastly, we are the first consulting firm first consulting investment by Databricks Ventures.
[01:41] >> Thank you, Jay. Um hello, everyone, and welcome. My name is Ankur Jain. I am chief cloud and data modernization officer at Acxiom. Um Acxiom, think of about Acxiom as um you know, a company marketing company which basically helps you
[01:54] deliver marketing outcomes. So, we work across various different clients across various different sizes. You know, these are clients from financial services, health care, retail, CPG. And um
[02:06] what Acxiom is actually known for for last 50 plus years is in the space of data and identity. When I say data and identity, this is about customer data and customer identity. How do you resolve the customer identity across various
[02:18] different channels? And how do you enrich the customer profiles? How do you build a solid marketing foundation so that brands can actually deliver the outcomes to their end customers, right? So, um Acxiom is
[02:30] actually part of a bigger larger group called Omnicom. Um that integration is still in the works. We got acquired earlier this year. And uh what I'm going to talk about today is the transformation journey. You
[02:42] know, both both G and I going to talk about the transformation journey uh that Acxiom went through for last couple of years in terms of redefining how data and identity should be built. So, uh before we kind of go into the
[02:55] transformation journey, um let me just uh How many of you are actually from marketing side? Okay. Cool. All right, we have some presence for marketing. And uh how many
[03:08] of you are on from IT IT background? The remaining ones, I think uh a lot of IT folks as well. So, hopefully this session might resonate to both sides of the world because we going to be touching a little bit about technology and also a little bit about marketing as
[03:22] well. So, when we talk about marketing, um this is basically what is on every CMO's mind these days. You know, chief marketing officers are thinking about how can they make their brand, you know,
[03:35] be visible, be heard uh in an always-on fashion. Meaning, it's not like about sending those traditional campaigns once in a while, whether it's monthly, weekly, quarterly, things like those. But how can that brand be, you know,
[03:47] visible to their end customers in an always-on fashion? Depending on which geography you are coming from, whether you are in Asia or Europe, US, you know, the brand should be visible across various different
[03:59] channels. How can you deliver those messages, those offers to your customers in a personalized fashion? Personalization is another big thing. What appeals to me may not appeal to Jay. So, how can
[04:12] brands actually deliver a personalized experience to their customers? And then as I said, you know, it's because of always-on, near real-time is another big thing that they want to reach the customers in the moment of time rather than, you know, sending an offer next
[04:25] day or 15 minutes early 15 minutes later. That message needs to be contextualized in that particular moment of time. And then as you can imagine, AI is everywhere, you know, AI is disrupting
[04:37] every industry and so is marketing being disrupted. So, these are like some core themes that we hear from our clients almost on a daily basis. And the underpinning to all of these
[04:50] initiatives is actually the customer 360, right? Like this is basically what every CMO, every marketer is interested in is how much do they know about their customer. How can they build a 360 view which
[05:03] gives them not only digital view of what's your cookie ID, what's your device ID, things like those, but also terrestrial information as well that your name, phone number, address, how can we reach you,
[05:16] what kind of you know, preferences you have, things like those. So, that's the underpinning, right? And then also along with the 360 view, there are two major things that come along, right? How do you resolve
[05:28] the identity of customer across various different channels and how do you enrich the profile of the customer with additional data sets, you know, so that's where we work with lot of our partners to bring in different type of data sets to kind of help brands
[05:40] build the 360 view of their customers. Now, as Acxiom is working with with our customers, you know, what our clients are asking is um the information to be supplied in near
[05:52] real time, you know, because as I said, you know, always-on marketing, near real-time campaigns, you have to have this data in real in near real-time fashion. And then also, as we work with our clients, the traditional way is that
[06:05] brands send the data to agencies, it comes through SFTP, manual pipelines, it's a nightly batch. Um by the time the data goes, you know, that whole you know, interaction with
[06:17] the client is actually all almost done. So, they are requesting, okay, can can Acxiom provide the access to data and identity assets in their own platforms? Such as if they are on Databricks, you
[06:30] know, how can Acxiom assets be available in client's own Databricks instances well? So, those were like uh some of the key requirements that our clients were asking. And then, as we look at our platform,
[06:42] and this is I'm talking about what existed before, um we were um very heavy on on-prem, right? A lot of our products were actually being built on-prem, and providing that promise of near real-time
[06:55] access, meeting the clients where they are, just wasn't possible, you know, being on-prem. So, these were some of the challenges that we were running into, which is like we were constrained by scale, um we had limited capacity
[07:07] being on um on-prem in a data center, um our platforms were constrained in terms of scale. As you can imagine, efficiency was also um limited, you know, how much can we scale and how much we can grow and how
[07:20] much we can perform. And then, the time to value was also a big challenge, you know, clients were asking real-time access, here we are providing them, you know, um nightly batch or a um monthly batch kind of uh
[07:32] refreshes to the data. And then, of course, uh being on-prem also limited our ability to be, you know, AI-ready. So, those were some key challenges, and that basically became the genesis of okay, Acxiom needs to transform. And
[07:45] that's when we partnered with Databricks as well as with Lovelatics to help us out in what that modernization journey. So Jay, would you like us to uh just walk us through what that modernization journey is?
[07:57] >> Yeah. Yeah, absolutely. Thanks, Ankur. So So Acxiom realized early on that addressing the challenges that Ankur just spoke of and and meeting the new demands of marketers was going to be a transformation that
[08:09] didn't occur overnight. It was actually going to be a journey. It was going to be a journey that required detailed planning and very thoughtful execution. So we were delighted when they invited uh Lovelatics to partner with them on our on that journey.
[08:22] As I mentioned, we've done this for many clients before, so we were well positioned to help them not only build the strategy, but also to to build that um to make that vision a reality. So it all started about a year ago now um
[08:35] where we we kind of joined Acxiom to start with a an assessment of their data and identity platform, build out a a recommendation for them, kind of get into the the future state architecture and things and build out a road map.
[08:47] But before we got into any of that, it really started on aligning on the future state vision. Or as Ankur often referred to it, our our North Star. So that future state vision was really rooted in five key principles here.
[08:59] One is that it had to be cloud native. So as as Ankur mentioned, many of the legacy processes were heavily bound to their on-prem infrastructure. In the future state, we had to liberate ourselves from this this this on-prem infrastructure so that we
[09:12] could take advantage of some of the cloud native capabilities. Things like centralized governance, zero copy data sharing, and clean rooms, which we'll get into in a little bit a little bit later in later in the presentation.
[09:24] It had to be designed for scale. So this this future state had to be able to automatically expand automatically scale to meet the needs of any of their enterprise customers regardless of size.
[09:36] Self-service. So this one is key in that this is where we meet the clients directly where they are. Clients need the ability to take Acxiom's products and incorporate them directly with the within their own environment so that they can use them
[09:48] when and where and how they want. It has to be AI-enabled. So, marketers need the ability to use AI and agentic capabilities to better engage their customers, understand their audiences, and activate them in an in an
[10:01] efficient way. And lastly, it has to be democratized. So, this means that um we have to kind of move beyond the um specialized skill sets and and pockets of information to it to really
[10:13] democratize that cloud-native uh mentality and understanding cloud-first best practices across the organization. So, rooted in that future-state vision, we started our assessment across the
[10:25] three key pillars of people, process, and technology. And I want to emphasize the importance of going through all three of those pillars because all too often we see companies jump right into that third one thinking that they can just kind of
[10:37] reskin or revamp that technology and it'll solve all their problems. So, of course we did the you know the the assessment of their data and identity platform, but to get that holistic look, we also had to evaluate
[10:49] their their people and their process. Who are the people that were building and executing these workflows? How are they organized? What are their skill sets? You know, what process do they use to onboard new customers and provide ongoing quality care to existing
[11:01] customers? All of that is critical to really getting a a well like I said, a well-rounded uh vision to inform the future state. After going through that assessment, we came up with several recommendations,
[11:13] four of which we listed here. One is to move towards towards decoupled workflows. So, what do I mean by that? Uh in the current state, a lot of Acxiom's operational processes were kind of tied to this
[11:27] uh legacy orchestration tool that was built to meet their needs at the time. While very functional and honestly well-regarded by a lot of the folks in the within the organization,
[11:39] that legacy tool um was heavily it was heavily coupled with the between the between the architecture and the the business logic. It also lacked some observability and
[11:52] was just kind of limited in overall flexibility. So, the the recommendation here was to move towards more of a decoupled workflow where we're moving towards metadata-driven workflows that are kind of set that modular
[12:05] foundation that can be used to build more cloud-native uh use cases. Second one, standard product offerings. So, another observation that we had as we went through our assessment is that a
[12:17] lot of the existing products the same product could have multiple different designs and interfaces, mostly driven by the the technology that was going to be consumed that was going to consume these downstream.
[12:30] So, having all of that out there, you know, multiple versions out there, that obviously created additional burdens such as, you know, maintenance, uh duplication, redundancy, governance challenges, and things like that. So, part of the recommendation for the
[12:44] future state is to standardize all of that to one common prod- product interface. So, it's the same inputs, the same outputs regardless of the consuming technology that was going to use them downstream. Third one, um cloud-native app layers.
[12:57] This one is is key obviously and one that we're going to uh get into more detail later in the the presentation, but this is where we actually recommend the use of Databricks clean rooms to bring these products directly to where the uh where the clients are. It's
[13:11] actually meeting them where they are through incorporating these products directly in their own environment. And lastly, team
[13:24] in the previous slide where it's really critical to make sure that the entire organization is moving their their skill sets towards more of the cloud native mentality. Maybe things like, you know, shifting towards from C++ background
[13:37] over to to PySpark or moving from uh, legacy uh, coding techniques to to modularized uh, Databricks notebooks. And maybe using how to use AI to accelerate this entire process.
[13:51] So, with that in mind, the the the the road map kind of looked like starting with reimagining the data and identity foundation. So, um, this is kind of what Ankur alluded to earlier where before you can address that long-term vision of
[14:04] uh, meeting the evolving marketers' demands, you first have to set that uh, address your technical challenges. So, that's what we're talking about here, doing that data and identity foundation using some of the recommendations that I just talked about.
[14:16] Second is moving towards a product implementation. So, this is where you get into taking some of the, you know, the existing inventory of products, but kind of uh, reimagining how they're built to be more standardized and again make these available um, in the
[14:29] environments directly within our clients. And lastly is migration. So, once you have your foundation in place and you have um, products ready that can be consumed by the clients where they are, we need to start migrating clients over to adopt this new
[14:42] new capability. And this is going to happen on a client-by-client basis based on the modernization um, maturity of the clients themselves. So, kind of going through this in a systematic way.
[14:56] So, what does this look like in terms of the Databricks architecture? So, Axiom processes first, second, and third-party um, data from a from a variety of of clients and partners.
[15:09] But the intent here is being able to ingest this data in a way that's much more transparent and scalable than they have in the past. Moving into the the right, moving to the
[15:23] right here, the the red box. So, that's your your main data processing layer, and that's where uh standardization is brought to the to the primary data and data and identity workflows. So, some of the core functions that Axon
[15:36] provides like hygiene, identity resolution, and enrichment, these now become reusable platform capabilities rather than highly customized implementations. That's underlaid by a medallion
[15:48] architecture to provide kind of that um structure and trust to that whole process, being able to see that progression from raw to refined to to business ready output. And that also provides the lineage and governance
[16:01] along the way. And and just makes it the data easier to to use and consume. Along the right-hand side of the um processing layer here, those Databricks capabilities. So,
[16:13] so Spark, SQL, Photon, serverless workflows, all this. This is what provides that um the scalable execution foundation, Databricks scalable execution foundation, that really allows them to
[16:26] process these large-scale data and identity workflows in more of an elastic cloud-native way than they were in the past. And also, it allows them to kind of move beyond some of that rigid orchestration patterns like I talked about before to a
[16:38] little bit more of a flexible operating model. Underlying the the processing layer here is shared libraries. So, this again contributes to the reusability of the platform overall. So, this is where we're going to uh kind of
[16:52] integrate um common logic, rules, and services into packages that can be reused across clients and workflows. By doing so, this obviously um minimizes
[17:04] um maintenance, redundancy, and it allows you to accelerate new capabilities faster. Continuing to move to the right here. So, within this Databricks architecture, governance is a platform-level capability.
[17:17] So, things like Unity Catalog and Delta Sharing, this provides that consistent foundation into access, permissions, and governance, and and secure data collaboration.
[17:34] I Oh, and obviously with with that, whenever you you know, whenever you're thinking about data and identity and and audience-based use cases, that's critical, as we all know, because, you know, privacy and uh compliance are are table stakes in
[17:46] this industry. Moving towards the ML layer, some more of the advanced intelligence side of things. So, this is where NL Flow, notebooks, feature engineering, Databricks APIs,
[17:58] all of this kind of provides that foundation to building out uh audience propensity score audience propensities um scoring and and and other um data science use cases.
[18:10] But, the key point to call out on here is that all of this ML is grounded in governed, repeatable data, rather than just um you know, one-off model development.
[18:22] Moving towards AI, this honestly just kind of brings more intelligence and and ease of use to the platform. So, through things like Genie MCP and agentic capabilities, users are able to interact with the platform just using
[18:35] conversational text. They don't need to have that technical ability to understand end-to-end how everything works. They can engage uh business intent directly into the workflows, into the into the models, and
[18:47] and activation just through conversational text. And lastly, the endpoints. So, the goal of this entire architecture is to be able to activate directly into the places where where where customers work.
[19:00] So, this goes through a variety of endpoints. One key one here to call out is Omni. Akshay, do you want to talk about Omni a little >> Yeah, no, definitely. Um Omni is definitely a platform from Omnicom, um
[19:13] as the name suggests and this basically helps marketers plan the media where which channels do they want to activate the media in, how much do they want to spend, what the budget looks like, so things
[19:25] like those the activations across various different channels. So that's the platform that Omnicom has and Acxiom data as you can imagine is embedded is integrated very tightly into that platform. So anything that Omnicom
[19:38] processes basically Acxiom becomes the data and identity layer underneath, right? So those are Omni is definitely an integration point but also as we work with clients they
[19:50] may be on various different martech stacks could be Adobe, could be Salesforce or any other platform so we Acxiom has built native integrations with those platforms as well. And then of course as marketers run their campaigns they want to activate
[20:03] this in paid media, social media, owned media like all those integration points you know via our connectors are also available. But the key point I would highlight here is that without going too deep into the technology as you can see
[20:16] this is the modern foundation that Acxiom was looking for, right? Through Unity Catalog it provides you the governance, through Delta Sharing it gives you a secure way of sharing the information across various different
[20:28] platforms. And then of course the sheer size and the scale being cloud native you can process any amounts of data sets whether structured, semi-structured, unstructured in a scalable fashion, right? So those
[20:41] were like the key underpinnings of this architecture. >> Yeah, 100% yeah the only thing you know I would I would add on to that by saying this architecture basically enables Acxiom to meet clients where they are. It enables data and identity
[20:53] capabilities that are that are more scalable, more governed, more reusable, AI ready, enabled to activate in ways that were never seen before. >> So
[21:05] So you know like uh the story doesn't end here, right? So, Jay kind of touched upon the clean room aspect, um, and that's exactly what we um, actually partnered with LiveRamp and
[21:17] Databricks as an extension to this whole transformation journey is that, okay, how do we bring uh Acxiom products, you know, integrated with our clients' environment, right? Um, how can we
[21:29] enable um the identity resolution capability? How can we provide the hygiene uh performance, um, and also data enrichment capabilities in clients' own Databricks environment. So, imagine a client, um, setting up,
[21:43] you probably heard today there was a big announcement around customer lake, um, you know, how marketers can now bring in their marketing data in natively within Databricks using customer lake. So, imagine marketers have, um, customer
[21:56] lake with which houses a lot of their marketing data, first-party, second-party, third-party data, and now they want to resolve the identity. They want to build that customer 360, um, this is where this clean room basically
[22:08] will come in handy, right? Because it's integrated with Acxiom products behind the scenes, so our, um, governed and, um, the previous architecture showed you, all that goodness is basically
[22:20] powered powering up this this clean room application. And then clients can actually spin this up in their own Databricks instance. So, without sending PII across, um, internet over to Acxiom, they can
[22:32] actually implement this clean room within their own environment and actually start to resolve uh customer identity and start enriching their customer profiles as well.
[22:45] >>
[22:57] across, and it's fully governed, built for scale as it is uh native to cloud. So, those are like some key features, you know, through this clean room. And Jay's going to walk you through in detail as to what goes behind, you know,
[23:09] this application. But before we do that, I want to show you one of the pilot that we did with one of our client who happens to be from the automotive side. And they were actually in a process of
[23:22] launching a campaign for one of the new models that they are about to roll out in the market. And as you can imagine, you know, being a big brand, global brand, the number of sheer dealers that they are working with, you know, in US, and
[23:36] the amount of data that they process, you know, it's not only the customer information, prospect information, but also the subscription that every customer has, you know, in terms of add-ons that they buy when they purchase
[23:49] the vehicle, the finance data around lease and car loans. So, just the list just keeps growing, right? Like social media data, different campaigns, the performance, and everything. So, they were running into
[24:03] key three key challenges, like, okay, who should we target? Can we trust the data? Is the data trustable? And then finally, you know, okay, how much do we know and can we can
[24:15] we know more about the customer information that we have? So, these were like three main key targets that they actually wanted to um test with, and that's where we kind of implemented this Data Bricks
[24:27] application. And work very closely with them with all those aspects that I talked about for for the clean room. But Jay, you want to walk through the the details of what the clean room application actually looks like?
[24:40] >> Yeah. Absolutely. So, happy to announce that this IDGen clean room application is available directly within the Data Bricks Marketplace. So, just go out to the Data Bricks Marketplace and search for Axiom. You'll find many tiles that
[24:52] pop up. Just click on the the identity resolution one. That's going to be the one that we're walking through here. So, this this app allows users to basically do everything from
[25:05] onboarding the clean room, configuring it, to actually executing um ID gen workflows directly within the app. And there's a couple different ways that we can do that, which I'm going to go through here. One is is kind of using step-by-step um user guided steps. Um
[25:20] and the other is through MCP. Obviously, you can use conversational text to interact directly with an AI assistant that will help you along the way. So, we're going to see both of those.
[25:33] Okay, so just kind of go in step-by-step um just as if you were putting a new app on your iPhone, first step is to go to the marketplace, find the Axiom app, where it gives you a little bit of information about the app itself. What is uh a little background information about the app, what Real ID is, the
[25:45] benefits of Real ID, how you would use it. And this is also where you'll find uh the option to go ahead and request access to the app. So, you would click on uh getting access there.
[25:58] Now, once you have access to the the app, the next step is to go through and then the next it brings you to kind of this this user guided step-by-step process to uh onboard your app, get it configured. So, there's a few there's a few things that are happening
[26:10] here in parallel. Some is on the the client side, the responsibilities for the users themselves. Things like um kind of defining the permissions or making sure that the permissions are correct, as well as configuring or
[26:23] provisioning the the clean room itself. Let's keep in mind, this gets installed in the client's environment. And then there's also some steps that are taking place in the the background automatically on the Axiom side to do things like provisioning resources to
[26:35] that clean room that are needed to to generate the Real ID, uh setting up the credentialing process, things like that.
[26:47] Once the clean room's up and running, next step is to go through your input data set, do your map, do your uh input data mapping. So, one of the cool things about this is that this leverages auto mapping technology. So, it will do its
[27:00] it'll do its best to um automatically map these these fields based on what it's using your input data sets, and it will learn as it goes. So, as you go through this a few times after you run a few of these jobs workflows,
[27:12] it'll continue to kind of learn your specific installation and and map the the fields directly to you, kind of easing the user experience there.
[27:29] Next up, once you've mapped your input fields, you're ready to go. Go ahead and click the um Go ahead and click the run the job. So, as the job is running, as you can imagine, you know, it's it gives you some of the the the usual statistics
[27:41] that you would look whenever any job is running, you know, when it starts, when it stops, a lot of the volume metrics, what you know, how many what what data sets went into it, the volume, how many matches you got, things like that. You can watch that in real time here as the job is processing.
[27:59] And then when you complete, it takes you to a nice kind of visual representation here at the end, which will give you a lot of great statistics around you know, what was your what was your match rate? What was your score
[28:12] distribution? So, you can get an understanding of the confidence in your matches. So, you can kind of kind of see you know, I'm matching on these fields, which is giving me you know, this this level of confidence,
[28:24] where you know, maybe these other fields are are not you know, providing as as great of a match. And also, not shown on this this screenshot, but you can also get a sneak peek of the actual raw data itself with the with the real IDs
[28:37] attached directly in the data itself. So, everything that I just walked through there was kind of the the UI self-guided process. I mentioned earlier that um you know, an alternative here is
[28:50] to engage with the AI assistant to navigate through much of what we just saw. So, I know it's probably hard to uh to see on the the phone there, but this just an example of how how we're engaging with the the console here.
[29:03] Um in this in this example, the user's saying, "Hey, I don't I don't have any columns to map to the IP address in the currency currently selected table. Is that going to be a problem?" So, the the AI assistant is then able to
[29:15] say, "Okay, no. It's not a problem unless you're using IP only matching, which which you're not." And then it goes into providing an explanation of how the matching process works and how you can get by with the other fields
[29:27] that are that are not specifically for the IP address. So, uh that's kind of an overview of the the app. Very intuitive um way to navigate this clean room process and
[29:39] execute the ID gen capabilities directly in your own environment. Okay, Ona, close it out. >> Yeah, absolutely. Thank you, Jay. Um how many of you are by the way familiar with the identity resolution concept?
[29:51] You know, I know Okay, good. At least um we have quite a few. Uh but those of you just to kind of uh elaborate the example that Jay brought up earlier, IP address matching, basically
[30:05] you know, ID matching is like you have some identifier, which could be name, phone number, email address, IP address, device ID. You basically um run it through Axion graph to see, "Okay, can we get a match and assign a global ID to that particular customer?" So, that's
[30:17] basically the concept of ID. But through this presentation, hopefully you get the idea of the transformation, right? I mean, of course, it's a it's a big transformation, you know, from people, process, technology from that perspective, and it's it's not a easy effort, you know, as you all can appreciate. I'm sure you guys
[30:39] are all transforming within your organizations as well. But some key learnings, you know, that came out from our transformation, you know, uh is definitely time. What used to take weeks, you know, for us to process this data and build our
[30:52] assets, build our products, you know, that we take to market, we were able to shrink some of those pipelines in a matter of minutes, right? Like what the workloads which used to take weeks, hours, days are now shrunk to
[31:06] minutes and, you know, even sub minutes as well. A lot of specialized skills, you know, Jay mentioned C, C++, Java, like you name it, we had it. All these different
[31:18] technologies were there and how we kind of democratized that, how we kind of streamlined it to be forward-facing, you know, with AI-powered. So definitely that was another
[31:30] big transformation for Axium was to transform the skills of the people as well. Quality definitely improves, you know, the performance definitely improves. How fast can we take these products to our
[31:42] customers? We heard, you know, very clearly through pilots like these that I just showed you that how we are able to kind of help our clients, you know, to be present in their environment whenever they need it,
[31:55] wherever they need it, you know, so it kind of helped us, you know, to be in close partnership with our clients. A lot of data movement which used to happen through SFTP or various different
[32:07] means, we are able to kind of minimize that, make it more privacy and security compliant as well. It helps with our Axium definitely leads with privacy first. So it it actually helped us to kind of
[32:20] bolster even further the privacy aspect as well. And of course, speed to value, how fast can we take our products to our clients? That's that's very important for us. So it
[32:32] definitely helped us, you know, quite a bit. So hopefully was valuable um information for you all. With that, we'll just open up for questions.
[32:44] >>

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.