Noah Levine:

Sure. First of all, thank you so much for having both of us, Kyle and Dustin. This is super exciting. So I’ve been at Databricks just over a year and I joined most recently from Warner Brothers Discovery where I headed up a data strategy for our US advertising business. Basically, I spent the past decade being my customer, if you will, working at media companies and one measurement company. But before that, I had more than 15 years working at Tech. I worked at Adobe, focused on everything from streaming to analytics and data. That’s where I really began to get into the data side of the equation. And before that I was at double click in Google. I basically started my career at double click in 2000. So I’m OG ad tech I guess, and that makes me old, but there’s a lot of history and perspective. So that’s mate.

Dustin Raney:

Awesome. Thanks Noah. Adrian, why don’t you give us a snapshot.

Adrian Bolosan:

Yeah, just echoing. Noah, thanks for having us on. Love this podcast. So fun fact. So I’m eight months into Databricks Prior, actually spent time at Snowflake. At Snowflake. Fun fact. So Noah was my customer when he was at Warner Brothers Discovery.

Noah Levine:

True.

Adrian Bolosan:

We’ve teamed up here at Databricks. And then prior to that I spent as you guys, because known each other for a while. I spent many years in identity and data with Experian marketing services, and I’m an OG as well. Prior to that, once upon a time I worked for a company called Overture, which became Yahoo Search Marketing in early two thousands. So yeah, trying to stay young.

Kyle Hollaway:

Yeah. Well I love it. You guys bring such a wealth of industry experience both on the brand side as well as on the provider side. So I’m really excited about our conversation today and obviously media is a big topic in terms of both identity and data collaboration and just how the industry is starting to pivot towards a broader focus on first party data as well as the privacy implications of that. Because it sounds great to be like, oh yeah, let’s just pivot the first party data. It’s valuable. We all have our own understanding of our customers or the consumers we’re engaging on our platforms, but then that also carries a lot of implications on how we’re managing that data. And so we just want to dive into that today, hear your perspectives both from the broader just kind of industry level and then also specifically some of the innovative things that Databricks has bring to the table at this point in time. Adrian, why don’t you just kind of give us a quick overview as we’re going into this conversation about collaboration, Databricks collaboration. I know you guys have been doing a lot of announcements to kind bringing that to the forefront and since it’s kind of foundational component to this media conversation, ette

Adrian Bolosan:

For sure. I think foundationally we consider collaboration on one part of the spectrum, just sharing data, not necessarily the big clean room concept. And I think what separates Databricks from a technology standpoint, part of the ethos is open interoperability of data sharing. And so at the core of Databricks, just for those that aren’t familiar, we use an open protocol called Delta sharing, which allows our customers to share data with their customers, partners, et cetera, regardless if they’re on the same platform, cloud region, et cetera, right? You take that a step further, things do quickly get into the clean room territory where we support that as well. And Kyle, to your point, yeah, we went earlier this year in February on Databricks clean rooms, but ultimately our goal at Databricks, Noah, myself and his peers is connectivity between our customers, alleviating friction. At the end of the day, data has to go back and forth all day, probably more so in media than a lot of other areas. And how do you make that easier, more cost effective, less friction at the end of the day?

Noah Levine:

Of course I have to add 2 cents, which is dangerous with Adrian because he knows this stuff better than I do and has been selling it longer. But there’s exactly what Adrian said, but on top of it, it is about empowering the ecosystem to do more with the data than they historically have been able to do in a collaborative cross-party fashion. So when I think about, hey, the media ecosystem and what we’re going through and what does it need, it needs to be moving into predicting everything. And so that’s predicting the wide wonderful world of outcomes. It’s predicting who’s going to click on this button, who’s going to view this content, who’s going to buy this product, who’s going to become an advocate and essentially predicting this. And that’s one of the beautiful things that we’re super excited about, I’m very excited about is that from a data collaboration perspective, the way that you manifest is predicting is through machine learning, which is a form of ai and that’s supported by what we do from a data clean room perspective and a collaboration perspective. And that’s what’s going to allow this market to really navigate what are interesting times.

Dustin Raney:

So Noah, to build on you and Adrian here, thinking about the difference between Delta share and clean room. So I do want to kind of go back to that for just a second. I want our listeners to make sure we’re all on the same page here. So even Acxiom, we are within a holding company and say we have our data sitting somewhere and we want to simply share it over someone else within our ecosystem in a safe manner. So a share, it doesn’t necessarily mean an overlay analysis, it just literally means we’re making our data available in a safe way for someone to access. Am I correct in that assessment?

Adrian Bolosan:

Yeah, a hundred percent. The way I like to think about it, Dustin, is Delta sharing replaces legacy pipelines of APIs and FTP and of data. And in some terrible instances maybe still email, right? So that’s exactly correct. So if you think of things like impression logs and things that are less sensitive where you’re not joining, you had the keyword joining two match keys together, data sharing in general supports 99% of those use cases.

Dustin Raney:

So then the layer on top of that, what Noah just said about this predictive enabling more of a predictive collaborative environment. So I am engaging with brands across the ecosystem in various ways. So really what this does is opens up brands to collaborate in understanding what I’m interested in and what I’m doing, what I’ve recently done across a multitude of brands. And that’s kind of how we see that predictive with AI sitting on top of that kind of collaboration. Am I correct in that assessment?

Adrian Bolosan:

Yeah, I think it’s two or more parties that are joining to get insights across the collaborative group. And I like what Noah said earlier, a lot of the use cases that are happening today, it’s not that these organizations didn’t want to do these things, it’s the technology wasn’t really there to safeguard privacy and also governance and everything is getting more stringent. So it was never a lack of wanting to do these things. It was a lack of technology being able to support them the way that the lawyers and governance and stuff was going to say, Hey, this checks the box, let’s do it. Right?

Dustin Raney:

Yeah, and I’ll say one more thing is obviously there has to be a match at some point when you’re trying to bring together all of these collaborative environments. So I think that’s a good opportunity maybe to talk about a little bit of what we’re dealing with you guys from just an identity perspective. So would one of you guys be open to maybe talking just a little bit about how we’re sitting, our infrastructure is sitting in between some of those collaboration capabilities?

Adrian Bolosan:

For sure. Do you want me to start, Noah? Yeah, I would love that. Yeah, there you go. So I think starting foundationally up in all the services and solutions that you guys provide, it’s the old adage of garbage and garbage out. You have tons of the largest brands and advertisers in the world leveraging Databricks as their platform. And I always like to say this, but Databricks isn’t a seller of data and identity solutions. It requires the ecosystem to do that. So I think it is basic at the foundation layer, you guys are helping our customers run things like hygiene, understand multiple IDs on just using Adrian Boon as the household. My wife, her maiden name is one identity, but it’s actually under the same household. Some of these things that we walk and talk this language, but a lot of folks don’t understand the day in day out maintenance of a consumer file, a consumer data file.

And then on top of that, what going to Noah’s point prediction, what cars are in the driveway? When did I buy my house? Am I a renter? I’m just listing some examples. You guys provide this enrichment layer on top of that. And then at the farthest end, again, sticking with that, my wife’s maiden name, if we’re trying to do a connection between a large retail consumer brand and let’s say a media company, they may have disparate data assets or identities under the same household, and you guys have those great assets to be able to say, Hey, this actually are different people potentially married in the same household to boost match rates. So these are just great examples of the types of things that’s combined. We’re helping support in the ecosystem.

Noah Levine:

Would you go so far as to say, and this is partly a question for Dustin and Kyle, Acxiom’s also playing the role of activation layer activation and partner. And so that from a traditional sense you’re supporting activation between party A and party B, but you also have a vast network of connections to throughout the media ecosystem for being able to activate campaigns for any of your customers or for the IPG footprint.

Kyle Hollaway:

Yeah, absolutely. As you guys were talking, I was kind of thinking, okay, what’s an analogy here? Because we have a really interesting dynamic here because Acxiom is a customer of Databricks. You guys have a great platform which helps to fuel some of our data curation and the development of our data products, which are our third party data that we bring to market. And then at the same time, we’re bringing our capabilities, including that data that we’ve curated to Databricks customers or our joint customers or ones that are in the ecosystem to make that available for the things we’ve just talked about, for enriching insights, for being able to have higher fidelity matching to be able to collaborate at a scale and a accuracy that traditionally hasn’t been really possible when you were just dealing with cookies, which were very obfuscated and hard to really understand exactly the entities that are being brought together. And then at the same time, we’re also collaborating and looking at the future and then how to jointly as partners go forward with certain capabilities, especially in the collaboration ecosystem and such. So it is an interesting dynamic because leveraging each other’s technologies as well as also bringing it jointly together to bring a better solution for customers.

Noah Levine:

Totally, totally. And the way that I think of it is Acxiom is providing a turnkey set of services and solutions and data products and offerings, and they can be executed fully within the Acxiom product and service and solution. But because you’re running on Databricks and we have a wonderful partnership, there’s additional opportunity to take advantage of, let’s introduce a term called composability.

And so Databricks essentially is able to provide a composable layer, if you will. And so the data products that Kyle you’re talking about and the activation services that we we’re talking about, there’s an additional level of optionality that is the better together story, and that is that the data assets can be made available within customer’s Databricks environments. That type of collaboration can occur and Acxiom can provide services out of the box, but customers can also who have their own data science or data engineering teams can leverage Databricks to be able to do machine learning on their own or incorporate proprietary services that they have. And so it adds a layer of extreme flexibility that I think is really the better together story that we are embarking on.

Kyle Hollaway:

Yeah, I love that. And just for our listeners to introduce yet another term, which is that aspect of native capabilities, when we talk in terms, and we’ll also bring up Snowflake because kind of playing in this ecosystem as well, but the ability to take Acxiom core solution capabilities and the services on that, but actually execute it in a brand’s owned environment that’s running within Databricks and have our capabilities running natively within that ecosystem is really the game changer. Historically, it was like, Hey, everybody send all your data to one place. We will curate it, we’ll do stuff, we’ll send it back to you, or we’ll send it to a partner. And there’s all this data movement and now being able to take everything we’ve talked about delivered in a native fashion where it’s sitting in one location controlled by the brand, by data owners, the controllers, so now they’ve got transparency, they’ve got control, they’ve got auditability where they understand what’s going on, and we’re able to then take the richness of what we can do layered on top of the full suite of capabilities that Databricks brings to the table. And now you’ve got this really rich environment, very secure that’s with full transparency, I think it’s a game changer right there.

Adrian Bolosan:

Agreed. Kyle, just add a huge game changer still early though the adoption shift of those willing to go into the new frontier, let’s call it. It’s like I think we’re still in that, right? You have very large organizations that are still early on in their modernization transition and especially those highly regulated type brands. And so you see those ones maybe a little bit slower versus others, but you’re seeing large financial institutions now and stuff really making that transition. So I think it’s going to accelerate, but I wish I could say everybody was on board. We’re not there. We’re not there yet.

Noah Levine:

It’s definitely a journey, and I don’t know if you guys are aware, we just launched or announced our Databricks data intelligence platform for marketing yesterday,

Noah Levine:

And

Noah Levine:

So that’s basically a big motion to speak to the CMO and marketing organizations directly, which is we’re really excited about that new motion.

Dustin Raney:

Yeah, no, I was going to say it’s like, and speaking to Adrian’s point on the slow to adopt aspect, it’s one thing to move to the cloud, but it’s a whole other thing for a marketer to think about leveraging something more like Databricks and this or clean rooms and these capabilities to drive multitouch campaigns and engage with consumers no matter where they are or what platform or media they’re consuming data on. And we are getting to a place where you guys sit in the middle and connecting all the dots even across the hyperscalers. So it doesn’t matter if your data is in AWS or GCP or if it’s in Azure, you’re able to taking the composability further to go across all of those different general clouds, have applications sitting on top of that that’s agnostic, whether the Adobe’s or Salesforce or CDPs.

Kyle Hollaway:

Very

Dustin Raney:

True. So to be able to have that, it’s like you’re wiring all that up, but then it comes to campaign, the tactical, practical matter. And for decades, the onboarder and these onboarding solutions have kind of been the easy button

And throughout and they use cookies as that currency. Well, the regulators didn’t love movement of data. We started seeing shifts in platforms. Moving to decentralization, you guys certainly plays well in your space, but what are your thoughts on marketing campaign management platforms that are easy journey management to be able to go in and say, Hey, I’ve got this audience, I want to send it to Meta, I want to send it to Trade Desk, I want to send it to Disney or whoever all at the same time. Are we mature? I mean, is that where we’re going? Are we already there?

Adrian Bolosan:

I think I can take a stab. Noah, happy to hear your thoughts. So what you pointed on right there, Dustin, I think is part of why adoption in general has been slower than we thought. So if you just think historically, and I say historically, Databricks obviously is an old of the company as Acxiom is, right? But historically Databricks was selling and still primarily sells our champions and customers are in the CTO org, head of engineering, et cetera. You guys traditionally have sold into the CMO org. So we talked about data intelligence for marketing and us expanding to those personas. So a lot of the slow adoption I think is the connecting, like you said, the connecting the dots of these environments. That’s what we’re trying to accelerate and that’s what I think it’ll do because today for activate campaign activation in a meta great example, Dustin, we don’t have a point solution for that today, right? Databricks plus Acxiom is a point solution for that. And your buyer today is a media buyer that doesn’t log into Databricks every day. And so that is kind of where we’re trying to connect the dots faster.

Noah Levine:

I obviously agree with what Adrian’s saying. I think that there’s, I don’t know if dream state is the right way to frame it, but this is a state of mind that we’re aiming for and that is where there’s zero movement of data and first party data stays in the four walls of the party that owns that first party data. And that solves a lot of privacy challenges. It also solves a lot of challenges associated with signal loss and let’s not underestimate the cost of risk management and with moving data leakage and how the trickle down effect that that has. So when you have zero movement of data, you’re able to solve for potentially the iOS ecosystem, which is 90 plus, 95 plus percent opted out of third party tracking. And so you’re really able to expand your footprint when you’re dealing with first party data, which is super cool.

And then where we’re going is essentially lots of different parties are going to make their first party data available for collaboration through clean rooms. And so in a certain sense it’s like decentralized data activation, decentralized imagine audience segmentation where you have a brand that’s marketing, you have an agency, IPG for example. You have their data house Acxiom, you have a multitude of media partners that they’re activating against. Some of those media partners make their own first party data sets available. If you’re a media company, chances are you’re very rich in first party interest based data. So what content interests you, what topics interests you, et cetera. If you’re a commerce site or if you’re a platform or if you’re a specialized media company that has some really interesting properties, then you might have some really compelling first party intent-based data sets, right? And so there’s this dream state that we’re not too far away from where you could see audience segments being built across all those different entities that are playing a role here for the benefit of that brand executed by that agency and the data house data partner across different media companies. And there’s some nuance there, but I mean, I don’t think we’re that far away and there definitely are some companies that are doing this today.

Kyle Hollaway:

So let me just ask maybe some clarifying questions around media in particular. So at least last 24 months, retail media networks have been a big buzzword around and lots of activity in that space. Some honestly have started to feel a little small walled gardens, which tend to kind of drift to a different direction of this kind of sharing collaborative ecosystem like you’re just talking about. And some of that may just be timing, the timing of certain capabilities being available market. So leveraging technology as it sat 24 36 months ago kind of drove a certain strategy. You guys especially are in those conversations. Noah, head of global media at Databricks, yes. Are those conversations opening up? Do you see maybe some of the walled garden ish behavior opening up some in the media space?

Noah Levine:

So I fervently believe it is an eventual, it’s the eventuality that the vast majority of the media networks, retail media networks, commerce, media networks, whatever you want to call them,

They will open up. And the reason why I say that is I basically lived and breathed the evolution of streaming. Basically I started off in streaming land, I want to say 2009. Yeah, I left Adobe in 2009 and I went to a video ad serving startup called the Tude in 2009. And so I’ll spare you the long historical gory details, but they’re just looking at the past, let’s call it five-ish years. What you had is you had the rise of a bunch of streamers, and I think there’s a direct parallel to what’s happening in the retail media network land. And the default posture was to initially take a walled garden approach and see could they pull off the walled garden approach. Because look at how successful that was for Google and Amazon and Meta. I mean, meta is a wall garden. They’re not

Historically video, even though there’s a lot of video on meta nowadays and it’s really done wonders for ’em. But what ended up happening was that the players in that space, and this was predominantly the media companies they were transitioning into becoming streamers, is that they realized that in order to retain the marketing dollars and budgets that they had, they needed to provide some cohesiveness to be able to still create a representation of that open and well lit marketplace that is historically known as the linear ad market. And so they’ve had to break down those walls. The next thing that happened was you had the rise of the platforms or buddy boxes, if you will, apple tv, Roku, Vizio, the LG ads, the list goes on. And so I described that almost as the balkanization of the CTV space to a certain extent because they in themselves were acting as walled gardens.

And the fact is what the agencies and what the marketers want is they want an easy button that makes it easy for them to see the entire inventory pool and they want to see the entire audience pool. Okay? There’s a reason why the digital giants Google Meta and Amazon are thriving from an advertising perspective in the United States. They have captured 60% of the US advertising budget. That’s amazing for them. But it speaks to the challenge that the rest of the market faces, and that is that you have to be of a massive size in order to really capture the budget. And if you’re not of the most massive size, then it means you have to collaborate. And so that’s why I think what you’re going to see with the retail media networks is they are going to open up their gardens in certain ways, and it’s going to be through data, clean room based collaboration.

And because you have to think about, they’re extremely data rich, but when they are serving ads or selling ads, they have their commerce experiences where they can sell ads, but that’s only one category of marketing opportunity. There’s a massive opportunity, and this might be controversial and a lot of them are very cautious about this, is making their data able to be married up with other types of media. And I think that that’s where we’re going to see things really blossom. And we’re going to see all different types of collaboration, but zero movement of first party data. Sorry, that was long-winded.

Kyle Hollaway:

That was awesome. Yeah, that was a little no dissertation. I loved it.

Noah Levine:

It’s

Kyle Hollaway:

One of my things.

Dustin Raney:

I mean, I always like to make a provocative statement or two in every podcast, otherwise if I’m summarizing, and I’ve said this before, it’s like you have the open internet and then you have the big four, the big wall gardens that you mentioned. And it’s like in a way, maybe Databricks is kind of playing a key role in championing the open internet, giving it another a collaborative capability together to power something that otherwise there’s no way that you’re going to climb that mountain and compete against that 60% market share. But by doing, by Databricks sitting in the middle, enabling this collaboration in the right way, maybe it performs even better. And because at the end of the day, the marketer, the CMO is like, where do I spend my next dollar to get the highest return and what media is going to perform? And that’s going to be data closest to the

Noah Levine:

Consumer

Dustin Raney:

Data closest to the edge and the most relevant, not data. Data. And that’s an area that yeah, Amazon and Meta have and Google kind of dominated, but if you can make that accessible when you’re on walmart.com or target.com or even ESP n or Disney or wherever that is across the open internet and you bring all that together, I just care about getting my brand in front of the right consumer wherever they are.

Noah Levine:

Exactly. And they want to do it as efficiently and with as little pain and friction as possible. Yes, I agree, Dustin.

Dustin Raney:

But then at the end of the day, you also have your analytics guys going, well, I’m also spending on meta and I’m also spending on Google and other areas. How do I actually measure the full journey? How do I truly understand what’s driving incremental gain? So I mean, I think that would be the next question maybe is just like how are you guys seeing all of that come together today? How are brands leveraging your platform even from an analytic perspective to understand the full journey?

Noah Levine:

How about Adrian and I collaborate on this? I’ll take a first stab,

But I mean in the most fundamental brass tack way, it’s basically brands use demand side platforms and they have relationships with the media sellers, whether it’s a digital giant or a traditional media co or it’s a non-GI platform, but still extremely meaningful and valuable, or a retail media network. A lot of those campaigns are activated through demand side platforms. Those demand side platforms such as the trade desk, et cetera. They generate log files, those log files, trade desk log files are called red files. And basically they’re ad exposure files with digital signals in them. And there’s a huge amount of value in that because it’s digital signals in them.

Unless it’s your first party data, you’re probably not going to want to just share it out, send it out, et cetera, even though that’s historically what’s been done, but it requires some ones legal paperwork over the top to cover everyone. And so what the most common use cases is these ad server log files, whether it’s the DSP such as the trade desk or if it’s on the sell side, working with media codes, et cetera. A lot of them use FreeWheel, FreeWheel generates log files as well. And so essentially those files are being made available on in data clean rooms and identity resolution services usually, usually layered in on top of that to at a minimum household the events, and then marry up those events, those household events to census files, I first party data that customers have, identity spine, et cetera. And then you’re able to essentially perform basic measurement impression counting, but because you’ve householded it at a minimum or you got super sophisticated with Acxiom and it’s down to the individual level, you’re able to do reach and frequency, which is the holy grail metric that the media ecosystem operates off of.

And then you’re able to take it to the next level and you’re able to take the end marketers like website visitation data or purchase data, or frankly any first party data that they have, or third party data that could be used to describe how consumers are moving through the funnel to marry it up to those exposure files that are now associated with data. And so you can do really power, you can do full funnel attribution and then you layer in machine learning from Databricks and we do more than machine learning. And then you’re able to get to hey, predicting what’s going to happen next and so forth.

Adrian Bolosan:

I think the only thing I’ll add to that I agree with all that, is the ingredients and how to stitch all this together. People know it comes down to the resources. Dustin is like data science teams, marketing, science teams, organizations don’t have endless supply of those personas. So I think ai, ml and agents are going to help accelerate the adoption of that because they’re not going to need as many of those resources. A lot of that’ll be automated. I think that’s going to be the, because we see the adoption of these patterns and it’s not happening as fast. They just don’t have enough resources. And I think again, this is where technologies fill in the void of resources. I don’t think it’s replacing jobs, it’s allowing people to move faster where one marketing Zions person now has multiple agents working on its behalf. That’s really where we’re going. Exactly.

Kyle Hollaway:

So to kind of play off of some of that, I’d like to get some hot take because Microsoft just announced that they were shutting down their demand side platform. Microsoft invests formerly Xandr and AppNexus. So as you noted there, Noah DSPs have been a key component and continue to be a key component to the media strategy and execution. Hot take on Microsoft’s position. In fact, they’re going to shut down their DSP in support of their advertising platform emphasizing AI powered solutions like copilot thoughts.

Noah Levine:

I heard about this earlier today. I got stuck on an airplane for like 12 hours yesterday. I had a pit stop in Syracuse, New York that was unplanned, but ended up making a home. And so I was like, wait, Microsoft has a TSP. And then I remembered, oh yeah, they acquired AppNexus. I kind of forgot about that. And look, ad tech can be a challenging business and there’s a lot of DSPs today. There’s a few independent DSPs that are doing very well. Obviously the trade desk is the biggest independent DSP, but if you look at the walled gardens, Google has a leading DSP, Amazon has a leading DSP and has been extremely aggressive with investing there.

And so I find it interesting for the wall take into account the walled garden DSPs. They have, I don’t know, privilege access to the walled garden inventory and first party data signals and so forth. And if you compare what Google and Amazon have from a audience experience and inventory perspective compared to what Microsoft has in their audience in terms of how audiences interact directly with ’em, it’s very different. Obviously Microsoft has a massive footprint with gaming and so forth, but I think the hot take, you’re like, when are you getting to the hot take? The hot take. I don’t know the answer. I

Adrian Bolosan:

Guess that’s good.

Noah Levine:

Help me

Adrian Bolosan:

Adrian. So if you just look at market cap today, Microsoft’s doing well, right? They’ve topped it. Everybody as of right now, I am always intrigued in their strategy, but I feel like they’re focusing on be the destination and DSP, not necessarily being the vehicles to that destination, be the best destination in the world, and everyone else is going to build destinations to you or build, sorry, conduits to you. Video games is one great example that Noah touched on and we’ll see if that works out for them.

Kyle Hollaway:

Yeah, Dustin thought.

Dustin Raney:

Yeah, I mean my hot take here is that Microsoft is like, Hey, let’s just stay in our game and we will cut our losses or whatever that is. I don’t know what the books were there. I think it’s another hot take would be, you’re right, the Google announcement, antitrust, the breaking up of the giant and what’s that going to look like? Are they going to be required to sell their search or make their search data available to the competitors?

Noah Levine:

Are they going to have to spin out double click,

Dustin Raney:

Right?

Noah Levine:

Or parts of double click.

Dustin Raney:

And then I think there’s even lawsuits in play around the UID right in the state of California. So it’s a highly volatile time that we’re in that I think again, it kind of points to now we’re living in a world where we’re constantly shown, we have more access to information and people are more educated because we’re dealing with our screens and social media and stuff like that every day when things pop up in the newsfeed. So we’re more aware of how our data’s being used. And AI is I think, freaking people out a little bit. So people are like, is AI going to become sentient? And

I think there was a stat that said, I forget how many millions and millions of tens of millions of dollars that open AI is spending just because people are saying please and thank you in the prompt. So that points to, I guess people think that they better be nice to AI’s going to come and hunt ’em down later and lie. It’s a crazy world. And you kind of see, where’s this going? We know that there’s desensitization of data for that very reason. You all grew up out of ai. That was your bellwethers. You’re centered on AI driven before anyone else in the market to a certain degree. Where do you see that landing? I mean you talked about yeah, resources. My wife, she always sees it as an insult when she’s a photographer and when someone sees one of her pictures and they’re like, Hey man, what camera you use to take that amazing picture? And she’s like, what about me? I’m the person using it. So let’s talk AI for just a quick second. I know we’re running out of time, but kind of want to land it maybe there. What are some thoughts on where you guys see it going and where you’re driving it?

Adrian Bolosan:

You want to start Noah?

Noah Levine:

All right. And I’ll try to keep it brief, which is challenging for me apparently today. So AI is going to help us scale. I don’t believe it’s going to replace us at least at this point. And we have an umbrella of use cases that we particularly focus on. It’s called batch inference. Basically it’s summarizing information. AI is going to, and it’s already eliminating the drudgery of work. And that is going to be how it predominantly manifests in our lives. We’re going to see it begin to enter into the creative process as a tool that helps us work with things like Photoshop, but easier Photoshop. I worked at Adobe for six years. I don’t know how to use Photoshop, and it’s a hard to use tool. And when I think about what AI will do, it will automate the process of what power users of Photoshop will do. Yes, that is challenging for graphic designers and so forth. So yes, there’s a category of specialization there that AI is going to begin to get into, but I think really the most of it is about being agents that empower the information workers and we’re all information workers on this gathering to be able to scale and do their jobs better.

Adrian Bolosan:

Just to add, doesn’t scale move faster, but slowly but surely it’s changing the way we do everything in our own personal lives as well as work. You touched on the Google thing and breaking up parts of that. I joke and say that’s eight to 10 years late, especially for search. Search numbers are down. My daughter, she’s 10. She doesn’t say Dad, Google that. She says, will you chat GBT? That for me? That is the fundamental change in the home. So it’s the little things, right? So that is like the aha moment for me.

Kyle Hollaway:

Yeah. Well that’s great, man. I knew this would happen. We have way more to talk about than we have time because I know there’s a lot of stuff that we could pursue because as Dustin noted, this is a very dynamic environment. A lot is going on every day, new innovations, new announcements. But I’ll go back to something that Adrian said at the very beginning, garbage in, garbage out, that still holds true as all of this is taking place, ensuring that it is a captured, effectively managed carefully, privately, but enriched and validated and then shared in a manner that has high scale. But high accuracy is really going to be the foundational to all of this, whether it’s activation, whether it’s feeding in Agentic AI system to do things based off of that data or whether it’s just in generating more insights. We’ve got to get that data and AI layer correctly. Oh, look at that. That’s a Databricks tagline. Data and AI company. Yeah,

Adrian Bolosan:

It’s working.

Kyle Hollaway:

It is working, but it is foundational, right? It is there. And Acxiom, being able to bring our expertise into that ecosystem on behalf of brands, I think really directly addresses that garbage and garbage out question and ensures that all of these activities are done at scale, but with accuracy. So that said, we do have a standard wrap up talking about ai. We’ll just go ahead and go down this path. I’m going to ask each of you to volunteer, if we fed, which I’m sure somebody has probably fed all the data about Noah into an ai, what are three words that you would use to describe yourself? And then Adrian, you can go after that.

Noah Levine:

Well, apparently loquacious,

Dustin Raney:

You actually did it.

Noah Levine:

Yeah, loquacious, opinionated and hopefully curious. Love

Kyle Hollaway:

It. Yeah, love it. Brought the big dictionary today. Look gracious. All right. Adrian, what about you?

Adrian Bolosan:

I think mine pretty basic. I think data would definitely be the top word there. Give my background collaboration given what I’ve done the last few years and probably a joker at the end of the day. So it’s got to in there, given always like to bring a sense of humor to these things.

Dustin Raney:

Fantastic. Well, humor always brings joy to the world, and I really appreciate that. And man, we really enjoyed this conversation and appreciate what you guys do because I do believe that Colin, I agree and Xi agrees that having partners that you can trust in to do incredible things together that actually solve key things, challenges for humanity, the fact that data’s not flying all over the place, that it is secure, that our own data, we’re consumers, it’s got peace of mind there. And I think Databricks has a very solid, builds solid foundation from that standpoint of consent, security, privacy, but also innovation by doing that allows us to innovate together. It opens up doors that weren’t open before, like you said, because of lack of technology. And I think because of a lot of what you guys are doing and how you’re advancing the infrastructure layer, we’re going to see hopefully a better tomorrow as it relates to data-driven marketing and the value exchange between consumers, brands, and publishers. So thank you guys so much for being on the podcast today. Definitely want to get you back at some point soon to update you all on these new incredible things that we’re doing together. Kyle, do you have anything else?

Kyle Hollaway:

Nope. Thank you guys. Appreciate it. Great conversation and exciting times and happy to be partnering with you all. So look forward to our next conversation.

Announcer :

Thanks for listening to Real Talk about marketing and Acxiom podcast. Find all of our podcasts at Acxiom.com/real talk or your favorite podcast platform. Until next time.