Good morning, everyone.
How's everyone doing today?
Good. Good. Thank you all for coming.
And I mean that to you sincerely because when we came here yesterday to rehearse, this was a very big room, and I thought how sad I would be when my four closest friends appeared, and they were all sitting in various spots around. And so sincerely thank you all for coming today.
I also wanna just acknowledge for a moment how patently unfair it is that Zack Cass got to go before me.
It's just not fair. That's all I have to say about that. So for the world of abundance that that AI offers, AI should have scheduled him after May.
Yeah, yeah, or my assistantship.
So what we're going to talk about a little bit today is, I'll give you a quick introduction of myself and then Satish, who is my chief technology officer and, and partner in this will introduce himself. And then I think what we're going to talk about a little bit is some secular trends that we have observed in the market.
I'm gonna share a bunch of data with you. Rapid fire, probably data that you've seen before, some may be surprising, some not.
And then we're gonna talk a little bit about the challenges that our clients have had really taking advantage of those secular trends or or pivoting to, to, take maximum advantage of those.
And then how we used those as our north star in building our D3 technology. So D3 stands for deluxe data discovery that is our flexible Nimble cloud based architecture built on the AWS platform to unlock some of these capabilities. So today is probably maybe less of sharing new knowledge with you, although I do hope we share some new knowledge with you and more about telling you about our journey as we went down the path of migrating to and building a purpose built technology platform to serve the needs of our clients.
Alright. So See, this also happened to Zach. So shall I use the other clicker?
Did I not turn it on?
The next slide, you'll see a page of me and my face.
The great news is my face looks very similar, somewhat more aged today.
So I'll just start with my direction. So my name is Chris Lazzaretti. I'm president of Data Solutions at Delox. I came to Delox about seven years ago via position of FMCG Direct, which is the, was the marketing agency of First Manhattan Consulting Group.
Over the last twenty years, I've had the opportunity to advise most of the largest banks on marketing strategy, and marketing execution, and also the opportunity to contribute to a number of publications on data and some technology.
Which is a long way of saying, please believe what I have to tell you, but the great news is most of the statistics are Satish, do you wanna introduce yourself?
Sure. Thanks a lot, Chris.
Everyone. Good morning. This is Satish Palace of Roman.
Tenth month in Delux, but it feels like it's been a long time.
I've been in the financial services industry for more than twenty years now.
Worked for American Express before this hitting, the enterprise architecture in some of the areas. And then I'm also involved in multiple, large digital transformation. So pretty excited to share the stage with, Chris.
Fantastic. Are you going to advance? Oh, the clicker's working. You go. Fantastic.
Thank you.
Most be the might be the most excited I get all day.
Oh, there it is. Okay. Four secular trends I'm going to share with you today.
We think these are long term trends. Some of them have accelerated a little bit.
For trends, placing data at the center of all marketing decisions.
Consumer expectations for personalization always on real time trigger marketing and AI.
Some of these have accelerated. Some of them are newer.
But the challenges that we often see as our clients try to take advantage of these is a proliferation of disparate data, complex analytics and the ability to develop actionable audiences, flexible and nimble marketing automation technology, and understanding and applying the new technology that is AI. And so as we thought about building our D3 platform, these were all things that we looked to solve, and we put as the North Star of our, of our build process. Now, I'm gonna share some data with you just to make these very real.
So This data comes from the financial brand, or the digital banking report, which is the publication of the financial brand. And I think it's not that financial services markers don't appreciate the importance of data. In fact, seventy one percent of, say, data is critical to what they do. Twenty nine percent say it's critical for all projects. Forty two say it's important for most. But if you look at the chart on the right, only nineteen percent say they use it all about half say they use it always or for the majority of their projects, but still a big gap between those who appreciate that it's important and those who actually use it all the time. And that kind of makes sense because if you ask marketers what their biggest challenges are, they're happy to tell you.
And it's access to an ability to use the data are the top challenges. So those two red bars are significant or overwhelming challenge or slight or modest challenge, the only the, the black or the gray there represents no challenge. And so ability to leverage data and access to the right data are the biggest challenges preventing the seventy one percent of those who appreciate it's importance to being all hundred percent of those seventy one percent using it regularly.
And it sort of makes sense, if you look at the data ecosystem, if you look at B2B data specifically, it is a incredibly fragmented ecosystem of data.
So we think that to have a really relevant conversation with a small business for financial services, you need to understand their cash the money in and the money out. You need to understand their demographics, who the right person is to talk to, kind of how they sit as a business today. The problem is all of that data is not in a single place. It's all over. It is everywhere. You can get firmographics from a couple of providers. You can get really good contact data from another set of providers.
Public filing data comes from other places. Credit comes from different, the business bureaus, trade credit behavior, specialty data, if you're marketing to agriculture businesses or, or, or medical businesses, but none of it really comes together. And the challenge there is that business names can be represented in almost a million different ways. And so if you want to reconcile the data and use it in a way that is sensitive to and relevant to the business, you have a challenge.
But it's not just business data. Consumer data is the same too. So if you look at consumer credit data, which is some of the most well regulated data in the marketing ecosystem, and you want to deploy campaigns that are sensitive to consumers inquiry behavior. So consumer credit triggers, you really have to go to all three bureaus to get full coverage of the credit trigger audience.
So that's the first secular trend, an appreciation of putting the data at the center of everything, but a challenge to bring that data together in a way that's easy to excess and easy to consume. Second challenge or second secular trend is personalization. Now, this has been a long term secular trend over the last half decade to decade. But has really accelerated after the pandemic.
And if you look at this is some McKinsey data and some Accenture data seventy one percent of consumers expect personalization, all of their marketing touch points, and seventy six percent get frustrated when they don't get But it's not just consumers.
B2B owners and decision makers say the same thing. Seventy three percent of B2B buyers expect the same consistent and exceptional experiences as consumers.
Not just these folks. Bruce Beagle is here who, who runs Winterberry, and I think of the trends that he might share today is the consumerization of B2B decision making. This is kind of a well known fact.
But what does personalization really mean? About two thirds of folks say that they want highly relevant touch points, sixty seven percent can be relevant product and service recommendations, sixty six percent, tailor messaging to my needs, sixty five percent offer me targeted promotions, things that you have to have a lot of really good data, to do well and at scale.
It's good business if you get it right. So McKinsey again found seventy six percent of folks who, get personalized advertising are more likely to purchase. They're seventy eight percent more likely to repurchase or for financial services, adopt more products and services.
And they're highly likely to recommend their friends and their colleagues to your brand as well. And you get this virtuous cycle of purchase with personalization.
But a fundamental tension has has opened here, and we see this amongst our clients, where the buyer expectation is they expect highly personalized journeys, which has to be powered by data. But at the same time financial services marketers are reporting that targeting personalization at scale are amongst their top challenges. There's a little bit of data. Again, where marketers rank their top challenges, it feels a little bit like family feud here. So if anyone wants to guess the top ten answers that are up on the board here they are.
So personalization at scale and targeting the right audience are, are high on the list.
And it sort of makes sense as you ask marketers to assess their self rated proficiency. So this is where a survey was done to marketers that said, how proficient do you feel at these various levers or various categories of, of functional marketing decision making?
And data and analytics actually was the lowest ranked in terms of self reported proficiency and, and marketing technology was not much above that. Again, all things that went into our consideration set as we thought about building our d three platform.
The cousin to personalization or the sister is the rise of real time marketing. So as consumers have expected more personalized marketing, The industry has migrated from episodic marketing, batch based campaigns, which are still important for certain elements of the marketing ecosystem system to always on near real time activation. And consumers expect that. Again, McKinsey Data, again, sixty one percent of consumers expect marketers to celebrate their milestones, fifty nine percent expect them to communicate based on key moments and fifty three percent expect triggers based on their behavior.
So, again, things that are powered by data and infrastructure and if you get it right, it is like fundamentally good for business. So Oliver Wyman found that twenty three percent of consumers consider changing primary financial services institutions each year. Only eleven percent actually switch.
Of the eleven percent that switch two thirds do so after having a major life event. And it's a material population if you get it right. So new business formation, there are about six million of them. Over the last twelve months, they're five times more likely to respond to bank marketing.
On the consumer side, about one in ten households move at every year, they're two times more likely to change primary banks. But it's not just on the deposit side. If you look at the credit activity of these households as well, pre movers and new movers. We actually did this analysis.
I've I've not seen, I'm pretty proud of this. I think it's a really interesting way to look at the data.
We fundamentally, what we did is we took a population of consumers who had life events. I'm only showing you two of them here and compared them to a randomly selected baseline control population. And then monitored their credit activity over the subsequent ten months.
About twenty percent of pre movers and new movers opened a new credit card over the subsequent ten months after moving compared twelve percent baseline.
And unsecured loans and lines, about eight to ten percent versus a four percent baseline. So it works for credit too.
But, again, back to the biggest marketing challenges to what extent did the following issues challenge your marketing department. Number one, data infrastructure and accessibility, which is fundamental to having near real time always on trigger marketing campaigns is good data infrastructure and good data technology to power So data infrastructure and marketing automation prevent many from deploying real time trigger marketing as they'd like.
And then finally, AI, thank goodness. I don't have very many slides on this because that cast, like, took it all.
I also feel kind of bad because mine are all about the fine impacts versus like the empathy and humanity of AI.
The good news is is it's, empathy and humanity could be big business get it right. So there's four trillion dollars of global productivity in total AI from McKinsey. Seventy five percent of that is predicted to accrue to just four areas, which is operations, marketing and sales, engineering, and r and d.
McKinsey thinks that you can have up to a fifteen percent increase in your marketing productivity based on GenAI alone, which is well over half a billion dollars for in in certain areas.
And so the excitement and that the opportunity is significant in certain areas like content generation, copy development, customer service, and search, for gen AI, machine learning, paid media placement, optimization, personalization, research, predictive analytics.
But again, some thoughts from Bruce here. Basically, what we are seeing is an obsession with AI today. Limited consumer impacts, really to date, but some really significant governance concerns. And that's why you see the usage pattern here. Only nine percent of financial services marketers say they are I use it today.
And so again, these became the North Stars as we chose to build the D3 platform. This was the journey. So we saw that there is a fundamental desire to place data at the center of all marketing decisions.
But the challenge is that that data is dispersed. It's everywhere.
And so we built D3 to easily consume hundreds of different data sets quickly and easily, bring them together into a singular unified data fabric for the use case and make it easy to consume for, for marketers.
Consumer expectations for personalization, the challenge there being complex analytics and the ability to develop actionable audiences, we fundamentally built the D3 platform around developing actionable audiences. We built in what we call our audience decision engines, which are our sets of targeting analytics to build, the highest, profit, highest response audience as possible, always on real time marketing.
That the challenge there being the infrastructure, we have built the platform to natively support near real time deployment based off of triggered behaviors. And finally AI.
And here we have really embraced AI in the platform, but fundamentally put responsible AI at the center of that. So that means respecting consumer preferences, in all use of the data, that means privacy, that means fairness, that means explainability, that means inclusivity.
We think that that's the right long term play for AI, which is putting those first and wrapping the capabilities around them. And we've built that as the native capability in the D3 platform.
And so, while I've talked to you a little bit about the secular trends, the challenges, and how we how we built the platform to serve those. Now I'm gonna turn it over to Satish to, to talk a bit more about the actual technical build of the platform.
Thanks. Thanks, Chris. I think now comes the tough part. Going next to Chris is always going to be challenging. I'll see how I could keep the tempo up wherever Chris lifted.
So just reiterating what Chris called out as the trends or the needs that we see from a marketing standpoint, easy to consume marketing data, having data in a way that we could centralize it, we could use it for driving all the insights.
Converting data into an actionable insight, supporting near real time triggers, life, life events handling, and then how do you infuse intelligence into everything that we do in marketing? So what I'm going to talk about right now is how technology enables some of these needs, some of these trends.
It's going to be more of an one zero one scratching the surface of what D3 platform is. There are multiple sessions. You could stop by to understand more on what we do in this platform.
So introducing, DX's data discovery. Which is our, data business go forward ecosystem.
So in data, we believe technology is the product. So what we have done with D3 is trying to take technology, trying to apply what we call as platform principles, and build an ecosystem, which is going to help with our data driven marketing business.
Just before going into what D3 is, let's talk about What we what did we see as some of the opportunities or challenges that we had in our legacy ecosystem?
So we had multiple ecosystem.
But the data that we're actually using from a marketing standpoint was growing exponentially.
So that we needed a system which could handle the large volume of data. We needed a system which could handle real time events.
So traditionally, we had everything which was coming in more on a periodic basis, more of a batch processing. But as we move now with all the personalization, all, the things that are happening around us, we needed a platform which could support those real time events. We needed a platform which can really create actionable insights, which can be much more democratized, which can be easily accessible for our internal analyst for our data scientist, everyone. And last but not the least, we definitely wanted to embed AI into everything that we do So mix of all these is what the D3 platform is. Now this is again a very, very high level view of what our D3 is made of. So d three is nothing but, centralized, democratized, governed self-serviceable data ecosystem.
As is the case with any, data ecosystem that you would have seen. So what we have is there are data flowing in from multiple of our sources There are, third party data, first party data, multiple external publicly available data. So these data has to be ingested into a centralized ecosystem. So that's where the data ingestion comes in. Now the challenge that we had was there are multiple ways in which this data could flow in multiple formats in which this data could flow in, and we also had data, which could be batch, which could be real time streaming, so name of technology we had to support. So the data ingestion layer was built in a way that it could support all of this data formatting.
It could support all of the periodic or the events in which the data could flow in. And also, we needed to integrate the ingestion layer, which could be more configurable. So we didn't really want to spend a lot of time when we get a new source. We want it really to be plug and play.
Then the data processing layer. So the data processing layer had to really compose a lot of, processing together to make the data available from a marketing standpoint. The data processing layer was built with again a composable architecture, which is going to help us plug and play multiple of the components to come together and create the needed insights or the needed data in a centralized fashion. The data was centrally persistent, which you see as the data leak house, which created the platform that was needed by our team to use the data, create actionable insights on top of The the layer that you see at the bottom, which is, which I feel is one of the key layer, is the AI and modeling layer.
So you You've got data from multiple sources. You centralize the data, but how do you really make this data meaningful? How do you really drive insights out of this data? So that's where the AI and the modeling came into play.
So we created a layer on top of the data where we are going to create those actionable insights using all the latest and greatest modeling techniques that are available. All the all the advancement that you heard from Zach is all embedded here. And then we have the distribution engine, which is going to take the insights, distribute it to our customers in the needed format. Again, the objective here is how can I support it across various formats that the customers are looking for, be it a file, be it an API, be it a push, all the mechanisms are supported from our data distribution data standpoint?
So just to again, kind of, give a overall view of it, So this entire deal ex data discovery platform was built using the platform principles with reusable framework completely a plug and play composable architecture, which can support multiple egress and egress from a data sourcing standpoint. And also it's completely infused with intelligence.
Just taking an example, just wanted to walk through on how the entire processing works, So just take an example of NAVZ mortgage, trying to find a population for HELOC or mortgage loan. So what happens? So the data comes from the bureaus. The data could come in multiple format. It could be a file based process. It could be an API or a push based mechanism. The data flows into our ingestion layer.
We get data from multiple public, records, multiple third party. So all the data flows into our ecosystem.
In in some cases, we also get data from our clients. So the client data flows into the ecosystem. So there are data flowing in from multiple sources that goes into our ingestion layer. Then we take it through our data processing.
So we standardize the data. We do the entity resolution. That's where we much more meaningful information to the data, creating the mastering of the data, and then try to match it with the records that we already have. And then create meaningful data in our data lake house.
Once the data is available in the data lake house, our data scientists pick it up, the modeling team picks it up They take it through the various modeling. So there are few which are dealer specific models. There are few which are client specific models. We take the data through these models.
Trying to create a meaningful insight which gets returned back into our data legos. So you have the raw data available You have all the insights available. Everything is available within a central data location, which powers our entire marketing engine. Then we use the data distribution layer to pass the data out to the clients in their needed format.
So we support varied sources. We support varied source types.
The entire ingestion is more configuration driven. It's not static. It's flexible depending on the sources. We could, adapt to it.
The entire processing engine is more compostable. It handles large volume of data.
ML and generative AI converts data into an actionable insights. That's our secret sauce, and then ECT consume insights and marketing data available in a democratized easily accessible fashion.
The entire data is secured. Security is always at the forefront. So the entire data is secured. Access is controlled, and we definitely apply and responsibility into the AI that we use, and then the effective marketing data is delivered to our client needs.
Quickly moving on. What do we really do with machine learning? I think, we have spoken a lot about machine learning. So what do we really do in machine learning?
So we have always had models modeling in the marketing world. So we started with what we call a statistical model. And then the statistical model could take us to some extent. So we have been using statistical model, but the volume of the data that is currently available is much more than what a statistical could handle.
So that's where we use the part of mission learning and deep learning. So the mission learning and deep learning helps us take the additional volume of data it can go across multiple data points, enormous data points, like some of the files that we get have, like, seven thousand attributes. Now, if you want to do a statistical modeling on that, it's going to have its own challenges. That's where the AI or the mission learning and the deep learning comes in.
And then sorry. Obviously, we have the generative AI, which is going to power it much more with all the other LMSs and everything that is available.
What we see, if you see on the left side, So the entire mission learning and deep learning has increased our efficiency in target generation by forty percent. And we have seen that the entire marketing performance has improved by 2x with all the models that we have.
Why do we really use AI? There are three places where we use AI, one in our data processing because the data that we get is so complex There are multiple sources that all needs to be done to the target. So we use AI.
We we actually use an homegrown LLM there to do the entity resolution.
Then we use AI in our, AI and ML in our complete model generation, which helps us create the target. And then we are also exploring a lot in generative AI, which is all more futuristic around content generation, contract management, audience generation.
So that's the part of AI. So the entire, D three, the D. Deluxe data discovery is powered using the AI and ML technologies.
Just wanted to make sure we have multiple data. We have credit data from our customer, credit data of the customers.
Entire data that we have in the platform is completely So data at rest is secured. Data at rest is encrypted as the data flows in from our solar system as we take the data out to the various, customers the data is always encrypted and secured throughout the ecosystem and in transit.
Finally, I think just wanted to cover it up to say So we have the data ecosystem. We heard what the trends are. We heard what the needs are from Chris. So the data, the DX data discovery platform helps us in bringing data from multiple sources in a more config configurable way democratizing the data and insights helping our users to consume the data much more smoothly.
The entire environment is secured and access controlled. And the the entire platform was built for handling large volume of data that we helping us to grow at scale. And then intelligence, ML genetic AI is at is at its core. And we have just made sure that this entire ecosystem is repeatable by itself. As we go across multiple of our sources, our as our needs increases, the platform is built in a way that it could adapt itself. And you could grow with it.
Fantastic. Great work, Satish. Thank you for telling the story here.
Before we go on and maybe take some questions, I just wanna say thank you to all of our partners who have helped to make this happen. Some of whom are in the audience, our AWS folks are out here are Experian folks are out here. So thank you to everybody who really helped us bring this together.
While you think of questions, what I, what I think I would say is Satish did a really good job of making that sound reasonably straightforward and simple, but we really do think and we are so excited because we really think it's a step change in what is possible in pulling together truly massively, multi sourced data, applying AI and ML analytics responsibly and quickly at scale and then delivering those back in a flexible way that allows clients to them, quickly.
So with that, any questions we can answer for you about the journey?
Yes. Oh, Clarissa.
It's an insider, by the way. So if I choose to I love that question. Thank you. I'm gonna let you answer the technical part which is, but I'm gonna answer the business part first, which is, if I were to do it again, I wouldn't, which is the business part. So what I would say is if I were, and that's kind of a joke. On the, I would do it again because I think it's the right thing for our customers and for our business. But if I were on the other side, If I were on a bank side or a marketer side, I think going through this journey has really reinforced to me.
The importance of really understanding where it makes sense to insource and where it makes sense to outsource, there is, fundamentally, scale play here where it really doesn't even make sense for the largest financial institutions to kind of invest in hundreds of sets of data billions and billions of records of, of data pulled together quickly, and instead using it truly as a platform to build on top of.
And so I think that was a key learning for me. I'm proud of what we've built, but I think it really makes sense as you think about that in source, outsource continuum. And now I'll let you talk about textures today.
I think anybody hear me? Okay.
The biggest learning for us is I think we have used all, I would say, the latest and greatest technology is available at this point in time, but we have also pressure tested lot of services, lot of technology over the course of the people. So I think the the learning is how do you really make a true platform? So there are multiple producers of data. There are multiple consumers of the data.
How do you really make a platform, which is going to help us go grow at scale from a producer standpoint. Like, I should be able to handle any kind of sources that I want to plug in in future. Similarly send data to anyone, on the other side from a consumption standpoint. Building the true platform, I think it was a channel.
And I think we have learned it over multiple iteration.
But now I feel I think we are at a point where we have built a platform we have, use right set of services, which helps us to process the data that we are, really handling. And I think That's a learning which we really want to apply across other areas within DX. At the same time, I think even from our customer's standpoint, I think probably, we could definitely help them to see. I think, especially with deal express where we say the customers could come and easily plug with some of the products, some of the solutions that we have. I think the learning that we have had in building a true platform centric, model, I think will definitely be beneficial. That's that I believe is the biggest learning of this entire exercise.
While you think of more questions, I've forgotten a few other. Thank you. So Chris actually started this journey a while ago. And so I wanna say thank you to him as well, and Chip writes the check.
Yeah. So, fundamentally, I sort of think about the ROI in the platform.
Maybe even less about us and more about the ROI are customers derived from it.
And so, I think what the ROI is really are the turn is for our customers or the better experience for customers is, the ability to access a more comprehensive data set than perhaps is available anywhere else. I think we've actually quietly become the largest data company maybe you've never heard of.
If you, if you took our data assets and our, the depth and breadth of our data lake and compared it to some of the largest providers, I think you would find that it it's perhaps larger than than almost any any other.
And so one is the ability to make that available to customers immediately.
Is one, I think, better customer experience.
The second is the ability to create actual meaningful, relevant, personalized audiences out of that in a way that's that meets the specific business objective.
And then the third is the connectivity that makes sense the clients. You can access it kind of in any way that makes sense for you. And I'll give you a example of at least the speed and the scale. Barry talked about it earlier today. I'll double click on it a little bit.
So a top ten bank, basically came to us and said we need to raise some deposits.
At scale, and we need to do that quickly. That hadn't been in our plan.
We weren't really staffed up for that. In the old model, that would have meant that we would have had to scramble teams, that meant that we would have had to think about how we prioritize work differently.
And it meant that we may have had to negotiate schedules with them. But instead, we could actually turn to this technology, that had already ingested and stitched together the data that was necessary, that had already been essentially trained with the appropriate models to produce audiences almost instantly My team would disagree with me that it was instantly. There were some late nights on that. And the outcome of that was like three billion dollars of deposits, which is like a huge huge quantity of deposits quickly. And that is just not something that I feel like we would have been able to do or the client would have been able to do without this technology seeing anything you'd add to that?
I think that's that's perfect. I think the, one was the the the bank example that, Chris gave, and also some of the real time triggers that we are currently working on with the older platform I don't know how, EC would have been to taking a new requirement like that. I think with a newer platform, typically, I think, with with some of the technology advancement that we have done, even handling the real time data that is coming in from the varied sources and creating an event that we could push to multiple of our customer, I think is much, much easier. That I see as the biggest benefit.
I I, like, that is a great one. I forgot to talk about that. This is actually so cool, so you'll have to I know this is a long answer. Your short question.
It's kinda how I am. It's just the thing.
Don't get me going. So, we are on the precipice of launching like near real time credit triggers. Thanks to some of our partners who are in the room today. And this product is so cool because if you're a bank and one of your mortgage clients goes out and applies for a mortgage somewhere We can let you know that within minutes that that's happened. So you can have that conversation with your customer.
And that clearly makes sense for the bank. Like, I totally get that. But going back to like Zack Cass and like the humanity and goodness of the whole thing, That is a fundamentally consumer friendly thing to do. So a mortgage is probably one of the largest financial decisions you will ever make in your life. One of the biggest purchases unless it's Chip, and the biggest purchase was actually the D3 platform. Thank you, Chip.
So for the mortgage, If it is the largest financial purchase that you're gonna make in your life, having access to additional offers to make sure you're getting the best rate, the best term the best structure is fundamentally a consumer friendly thing. And so I think the platform not only delivers, you know, real goodness for the actual bank client. I think it also delivers real goodness for the consumer at the other end of that as well to make sure that they're making a really smart decision, and that is something that is enabled by this platform.
Bruce, no no hard questions or I will walk off the That's an easy question.
Okay. Good.
When we think about doing personalization on our own websites or Yeah. Getting this data over to our platforms, like if you're using Adobe or Salesforce, like many others, how do you wanna how do you how do you envision connecting for real time activating that, you know, how do you move that data over?
Yeah.
Yeah. Great question. So on the roadmap, is fundamentally creating a democratized platform that supports all of the largest integration and activations. I think that is really the only way to deliver on our promise of allowing customers to access the insights in a way that makes sense for them.
The timeline on that and the technicality, well, Satish's mic has stopped working. Yeah.
So I think the way we are thinking through it is, like, if you had seen the data ingestion layer, we do have those plumbing already done.
So we can support file ingestion. We can support event ingestion. And many of these SaaS products that you spoke about be it Salesforce, be it Adobe, probably exposes multiple of their events on what happens on the websites, what happens in the mobile apps, So the plumbing is done to pick them up, push it into the ecosystem. The moment the data comes into the ecosystem, it's completely not controlled, whichever way we want to use the personalization. Use it to the models, push it to the clients in real time. All those becomes much more easier.
So to to, short answer your question, the necessary plumbing that is needed for getting all of this information in near real time. In processing the data in near real time is something that we are building as part of this entire platforms, enter cup proof. I think that's the real benefit that platform provides.
Do you wanna say that?
I'll dig it again.
So, our AWS folks are here too. So if we if one asked that, we can get up here.
So, yeah.
What what we do right now is more of traditional machine learning.
And then, some deep learning. But I think every time we build a model, we give importance to get the interpretability of the model and explainability of the model into account. So We will be easily able to explain what comes in and what is that the model has produced as an output. So we have not truly gone into a complete, we are not using our public LMS or anything like that yet. Meaning we definitely want to get into generative AI in future, and we are definitely working on how to explainability much more easier with generative AI. But with the model that we have built, with the model that we have control on, we could exactly explain what comes in and what goes out, and we have enough data points to see how the entire model has performed, how it's performing, and we also keep track of everything that happens inside the model.
I would maybe say one other piece to that is so, you know, fundamentally like a responsible AI framework is critical to again how we're thinking about it, which means explainability and exclusivity and fairness, but The other piece of that is effective challenge. And so we are always going to, if we have the choice between building an AI model or non AI model, we're going to build both of them and identify if we can actually have the same outcome with a more explainable, more transparent model, if possible. And only choose to use AI technology where it really actually drives additional incremental value. And then if we do use that explainability has to be a key piece of that.
Alright. I think I'm standing between you and lunch.
And we're right on time. Thank you everyone for coming. Thank you for all of our partners who supported the journey. Thanks to our customers who are here today. Enjoy the rest of the conference.
Let’s face it: Current secular marketing trends aren't only here for the long term, but they're also accelerating in application – specifically the focus on data-driven decision making, personalization at scale and real-time marketing. Recent advancements in technology, such as AI, are generating even more excitement for the potential of what lies ahead and the possibilities for transforming the way marketing is done at its very core. As the opportunity for marketers to drive value has grown, so has the complexity of navigating the intersection between data, analytics, and technology, creating demand for systems and tools built to handle a growing multitude of variables quickly. Enter: Deluxe’s proprietary D3 platform, Deluxe Data Discovery. D3 is a cloud-based framework that was purpose-built to meet real business objectives through data unification, process automation and advanced data analytics. Discover how D3, paired with Deluxe’s deep and long-standing marketing expertise, can deliver on even your most ambitious marketing goals.