Is the hype around AI in marketing justified, or are we setting ourselves up for another “tech bubble” disappointment?
Agility requires not only embracing new technologies like AI, but also a fundamental shift in mindset, processes, and even organizational structure. It demands a willingness to experiment, learn, and adapt quickly to the ever-changing marketing landscape.
Today, we’re going to talk about how AI is poised to revolutionize marketing, from personalization and customer engagement to the very structure of the SaaS market itself. To help me discuss this topic, I’d like to welcome, Rafael “Rafa” Flores, Chief Product Officer at Treasure Data.
About Rafael Flores
As an accomplished technology executive and proud immigrant from Honduras, I specialize in scaling SaaS companies from startup to high-growth enterprises. My career is built on my family’s deep-rooted principles: valuing education, treating others with equal respect regardless of background, and uplifting younger talent—because I was once that little boy with big dreams.
Throughout my career, I have led transformative initiatives at some of the most recognized names in the technology landscape:
- Meltwater: Played a pivotal role in the company’s successful IPO, showcasing expertise in product innovation and market readiness.
- Datanyze: Led strategic initiatives that culminated in a successful acquisition by ZoomInfo, enhancing data intelligence capabilities.
- ARM Holdings: Spearheaded innovation in Retail SDK and IoT solutions, advancing the company’s technology ecosystem and driving new business opportunities.
- 6sense: Led all automation, data, and AI-products, fostering a culture of collaboration and inclusion, while delivering data-driven solutions that empower GTM team(s) to sell effectively.
- Treasure Data: Orchestrated a landmark $600M acquisition by ARM and secured record-breaking Customer Data Platform (CDP) funding. Today, I am back leading Treasure Data through a transformative era of intelligence and automation fit for scale, while returning to an organization that feels like home—rich with talent, poise, and a passion for progress.
I am also a devoted father of three beautiful children and grateful for the unwavering support of my wife—a registered nurse who embodies strength and compassion. My core expertise lies in defining and executing product strategies, roadmaps, and key performance indicators (KPIs). I possess deep knowledge of CDPs, data management, privacy frameworks, and SaaS go-to-market (GTM) applications, scaling solutions for businesses ranging from agile SMBs to Global 2000 enterprises.
Rafael Flores on LinkedIn: https://www.linkedin.com/in/ref2019/
Resources
Treasure Data: https://www.treasuredata.com
The Agile Brand podcast is brought to you by TEKsystems. Learn more here: https://www.teksystems.com/versionnextnow
Don’t Miss MAICON 2025, October 14-16 in Cleveland – the event bringing together the brights minds and leading voices in AI. Use Code AGILE150 for $150 off registration. Go here to register: https://bit.ly/agile150
Connect with Greg on LinkedIn: https://www.linkedin.com/in/gregkihlstrom
Don’t miss a thing: get the latest episodes, sign up for our newsletter and more: https://www.theagilebrand.show
Check out The Agile Brand Guide website with articles, insights, and Martechipedia, the wiki for marketing technology: https://www.agilebrandguide.com
The Agile Brand is produced by Missing Link—a Latina-owned strategy-driven, creatively fueled production co-op. From ideation to creation, they craft human connections through intelligent, engaging and informative content. https://www.missinglink.company
Transcript
Greg Kihlstrom (00:42)
Looking forward to talking about this with you before we dive in. Got a few things to talk about here, but before we do, why don’t you give a little background on yourself and your role at Treasure Data?
Rafael Flores (00:53)
Yeah, no, definitely. I always like to introduce myself as a people first leader. So I am a father of three, Mia, Noah, and Liam. I am a husband of a nurse and I grew up my whole life and spent my whole career in California. So that’s a little bit about myself as a person. As a leader, I’ve been doing product management for over 15 years. I’m a builder. I love to grow companies from the ground up, take them public. And then I stay for the post public commotion and then I go build on the next one. Right. And so I was with Treasure Data for seven years. I left for about three and then I just came back as a boomerang. Very excited to be here, especially at the time where AI is a lot of noise in the market. And I think we’re poised to hopefully do some great things with it. So thank you for having me, bud.
Greg Kihlstrom (01:35)
Yeah, absolutely. So yeah, let’s let’s dive in here. We’re to talk about a few things and really all centered around AI as ⁓ you know, as as we both mentioned sort of in our in our interest. I want to start with personalization. And this is an area where I’m personally excited about ⁓ personally excited about personalization. But just because again, we’ve been talking about personalization for years and it feels like decades. We’ve been talking about it and yet.
I feel like generative AI and some of these other tools that are a little more recent are actually enabling this. And you’ve spoken about how AI can save physical retail. Can you elaborate on that and perhaps share an anecdote of how, you know, things like analyzing shopper behavior can translate into tangible real-time improvements to the customer experience?
Rafael Flores (02:27)
Yeah, no, it’s a great topic. Personalization, I am with you. I’ve been hearing around how can you master the art of personalization, right, for the last decade or so. Right. And I think before I speak to retail specifically, I think it’s important to define AI, right? And so my definition and our definition of AI here at Trosur Data is it’s a combination of two elements, right? It’s gen AI technology, which it’s obviously hot in the market today, as well as predictive and machine learning.
Right. And so that to me is credible AI, right. In terms of how it can save physical retail and many other industries in my opinion. Right. I think it all starts with in the moment experiences, right. I recently wrote an article where I spoke about this concept, but when you think of physical retail, everyone focuses around foot traffic, right. Getting somebody to the storefront, but you also have to think about that foot traffic in the terms of them navigating the store. Right. And then that foot traffic of them.
leaving your store. Right. And so from a from a physical retail standpoint, you want to make sure that you think about end to end what foot traffic means to you and staying in that moment. And the way I can allow you to do that is two things. The gen AI piece, right. It’s fed a lot of data around your shopper or consumer behavior. It allows you now to go and make sense of data that perhaps before require many years and many, many headcount on the IT and data analytics side.
The predictive and machine learning angle of AI helps you define brand affinity. It can give you an accelerated score with AI decisioning to ensure that you understand who is actually the best shopper for you, right? Across those moments. And so when those two meet, I do believe that it can empower brands to decide in that storefront what the next best action may be. And I think one of the best examples I could give you.
I travel fortunately and unfortunately quite a bit for my job. And we have ⁓ one of our second headquarters in Tokyo in Japan. And I recently had to buy a suit. I ran out of suits. I was traveling so much I was on the road. I ran out of suits. So I had to go because dry clean couldn’t get it to me on time. And I had to walk into a Nordstrom. And what I love about Nordstrom is that they they have figured out right that hey, if you really look at consumer behavior, when somebody comes in, through an entrance, especially those looking for suits. They’re looking for potentially quick hits because a lot of the demographic that’s buying suits right now are executives like me who are on the go, right? They need to go and find a suit. Yes, you will still have your typical shoppers with weddings, et cetera, but that is who’s purchasing right now. And so the moment I walked into the Nordstrom store, suits were right there to the right. They identified that consumer pattern, right? And they put it at the storefront. And so that’s an example of, hey, they have actually kind of mastered the art of personalizing that if I come in knowing the type of shopper I am and what’s hot in the market today with some AI in the mix, they can navigate me in that moment experience of when foot traffic is at your storefront, which is very critical.
Greg Kihlstrom (05:38)
Yeah, yeah. mean, and I love that you reinforce that idea that I mean, I think most people listening to this show probably know that, you know, there’s more than just generative AI out there. But I think, you know, generative AI has been getting kind of all the oxygen in the room, so to speak. And, you know, AI has been around for decades. And I think some of the like most meaningful use cases are still like old school, like RPA and stuff like that of like just getting some immediate results. I think what you’re talking about, though, is really powerful in combining different types of AI together. And also, you know, an example, you know, how a store is laid out, you know, it’s not the typical use case. You know, you’re thinking of like, OK, well, how do I automate something on a website or a mobile app or something? But like that that combo of how do we take predictive data or, you know, analytics in the real world and actually map a physical location?
I wonder, you know, are there are there other examples even beyond retail, like other maybe less obvious applications of of this that you’re seeing across different industries as well?
Rafael Flores (06:49)
Yeah, no, I mean, there’s there’s so many, right? But I can tell you, too, that our top of mind that I think people often overlook that they can have a tremendous impact in the sectors. Number one is financial services. Right. And the way I think it is, if you can actually feed AI the right data, right, and you have all these financial institutions leveraging it properly, right, with the right governance in place, obviously, you name it. Right. You can launch campaigns based on this type of behavior.
Right. So a great example is if I think of financial services, I was having a conversation the other day with a major bank and they said, hey, look, we want to target people based on where they are in their life. Right. And I said, OK, let’s think about a use cases. And one of the ones that popped to mind right away for me was you have steady savers, right? You have folks who who like to say, right. Now, I know we tell everyone, hey, you should be saving, but not everyone say. And so that’s your segment. It’s your audience. And so what I said to them is.
Greg Kihlstrom (07:39)
Right.
Rafael Flores (07:44)
What if you can identify who your steady savers are, so you can then go power a customer experience campaign on something very simple as an email goes out and says, hey, want to beat last month’s saving streak? Here’s where you saved. Here’s where you’re forecasted to save this month. Now for the bank, it’s a win. We all know the more liquidity they have, the more they can loan out, the more stability, right? For you as a consumer, it’s meeting you where you need the bank to advise you.
And so that’s a great example of one. The other one that also comes to mind always for me is healthcare, right? And that in the sense of AI is going to give you your own plan of how to treat things, right? Definitely not. But if we, for example, gave you a use case where, your watch data can tell you when stress spikes, right? And you know that at this time of day, it spikes the most. Well, I would like to know that because then I can send you meditation tips and tricks via SNS push because we know that you’re probably spiking and you’re likely on your phone. And so those are some of the cool use cases that I kind of think of as I hear some of those examples that I think those industries, if they use it to their advantage, there’s so much that they could do.
Greg Kihlstrom (08:54)
Yeah, I love that. Another thing that you’ve talked about and written about is just the AI’s impact on the SAS landscape. So as anyone that goes to any conference or reads any publication, every platform is saying AI this, AI that. so everybody’s got their AI features or whatever. But I think it can go much deeper than that when applied.
you know, when applied well, you you’ve predicted that AI will drive market consolidation and arise in market agnostic software. Can you talk a little bit about that? You know, what specific advantages to does a market agnostic solution provide over specialized platforms?
Rafael Flores (09:39)
Yeah, good question. And I’m going to give you a very atypical answer. I think when I speak of market consolidation and I think of me as just, mean, you 16 years ago as a product manager, just building products, right? think the reason AI is allowing folks like myself to build fast is it has really reshaped how you can build software, right? It’s that simple. I mean, to you, a product may look like regular SaaS, right? It’s on the cloud, it has a UI, but it may very much be all agents underneath the hood. Right. Right. We have an agent to give you an example that now actually will give you the react code for the front end. Right. so market consolidation is there for two reasons, right? It’s there because yes, people want to mitigate costs in the current macroeconomics, which we know we’ve been dealing with, but it’s also now possible because of my earlier comment.
Right. That we can just build so much faster. And I think the advantages of to your second question there, what are the advantages of, of this market agnostic solutions over more specialized platforms? I think specialty comes with the ability to move fast. And if you’re willing to innovate with this companies, they’re going to go and try to specialize and tailor even agents to your need. Right. And that makes it market agnostic. And so I think we’re in a new era where even composable, right? The idea of composability, it’s something that Trusted Data has dealt with for years, Best breed versus composable. Composable to me is being able to use your data anywhere and using the best pieces of any product. You could do that with us. You could do it with many, right? And so the question is who’s going to win that race.
Greg Kihlstrom (11:26)
Right. Yeah. So how do you, you know, as, a SaaS company, how do you, how do you enable your, your, your teams and the platform itself to move at the speed that it, you know, cause being able to move quickly is one thing, but actually doing it and doing it well is, ⁓ that’s the, that, that’s, that’s the magic, right? How do you, how do you do that? And, and, and kind of set, set a, platform and a team up to, be able to move quickly.
Rafael Flores (11:55)
I think you have to rethink across the organization first how you gear up structurally and functionally for it. And I’ll give you a tangible example. We reshaped our traditional professional services team here, and we renamed it to be AI and personalization services. And their focus is to deploy agents that are tailored to you. But we’re gearing up for a world where our product is going to be powered by agents. So we need professional services.
that really knows how to even build those agents themselves. Right. And so I think to move fast, you have to first look at within and set up your organization that way. Now, how can I move fast with quality in mind, right? With my teams, all my development teams. think the, I do bad product management principle and I’ll tell you why. Typically they tell you, Hey, do a feature that impacts many. think it’s different now. think because you can move so fast and leverage AI for it, you can do one thing for one customer really fast. And there’s 365 days in the year. If you shrink that in half and you do just that one thing for that one customer, you touch most of your customer base, which is better than sending them to a community support channel. And so those are all steps that you could take, right, that I personally take here and that we’ve taken as an organization to just get ready for this new world.
Greg Kihlstrom (13:06)
Right, right. Totally.
And so I want to talk a little bit more. know you’ve touched on it a bit throughout, but I want to talk about the really treasured data’s vision for AI and how that’s going to evolve. I always love, I think it definitely speaks to a company and a platform when someone comes back like yourself. you were there, you came back and sort of with this renewed vision. So, can you talk a little bit about how do you see AI being a key part of treasure data and kind of your place in the market.
Rafael Flores (13:50)
Yeah, very good question. mean, one of the reasons I came back, I’ll tell you why I came back. Obviously I trust the team here, I’m impressed with the founders and the board, right? But I was very excited to come back because Trusted Data actually has real AI that you could touch, right? And what I mean by that is I said to them, show me the product, right? Like what has happened? What has changed in the product from when I left it? And they walk me through something new, which we have, it’s called AI agent founder, right? You can actually go build and harmonize your own agents.
So it’s not just AI on the website. I actually trusted data before I joined. They didn’t even have it on the website, right? Which blew my mind. said, well, wait a second, let’s get this on the website. So they had it, right? So we have credible AI, right? And I think when I think about what credible AI means, right? It’s you can actually power the right use cases, right? And I think Trusted Data historically has been, I mean, the leader in many ways when it comes to managing very high scale loads of data.
And so AI is as good as what you feed it, right? If you go on chat, GPT, it’s not very good unless you give it a good prompt. so data is the prompt. And so when you tie those two concepts together, if we kind of own this first party data here, that’s your powerful prompt. Plus we have this foundry here. You can build any agent on top, right? That’s tailored to your business. And so to me, that has to be our focus. We have to be AI first. We can be AI adjacent. I think there’s a lot of AI adjacent companies. There’s a lot of website AI companies, right? There’s not many that actually have AI and we have it. Let’s use it. Let’s make it more powerful. Let’s speak of it, right? And let’s not be scared. You have to take risks and I’ll tell you this, I’m a risk taker. So we’re pushing through.
Greg Kihlstrom (15:18)
Hopefully. Love it, love it. One other concept that, you know, as I was prepping for this, I came across was the Diamond record. So do you mind just, you know, explaining what that is and how that kind of fits in?
Rafael Flores (15:46)
Yeah. So this, this one is all around identity unification, right? And in today’s day and age, we have been in this golden record era where you can go set up your, how you merge those records, right? Where an email here matches a cookie here. It matches a mobile device one here. And therefore that is Raphael at treasuredata. But a lot of marketing dollars are wasted in that approach because yes, it’s deterministic and probabilistic. However, if that email and that device match,
but the device is actually my wife looking at online shopping. Now they’re all this company’s targeting me in ways that it’s a waste of their money. And it lacks the knowledge it needs in real time. And so the concept of the diamond record, it’s different from the golden record in three ways. It allows you to really set up the right merger, how you merge those records in place to safeguard that it’s actually the right profile. Number two, it’s highly embeddable.
Greg Kihlstrom (16:18)
Right.
Rafael Flores (16:44)
We want to embed with a lot of different ID graphs, right? The trade desk, the live ramps, you name it, to make sure that we can power everything across and they can also power things from within, right? The more we share ID, the more we go across. so highly embeddability is important. And then in real time, right? Most companies, I hate to say all, but if not all, right? Have some pipeline delays and typically it’s 24 hours. So you have a lot of credible information and it’s good enough.
But what if it can give you every behavior in real time? That going back to the physical retail store example, if you have somebody walk in and they just got out of their car in between the car and the entrance to that store, they did a quick search of a competitor, right? Like are those shoes cheaper somewhere else, right? If you’re Northland, is it cheaper at Macy’s? And that person knows the moment that person walks into the store when they put their phone number that they just did that search, you can offer them a discount.
Right? But you can’t power that unless you have a diamond record. And so it’s hard to do. That’s why we historically also had not done it. But if we do it and we do it well, we put ourselves in a very different sphere.
Greg Kihlstrom (17:54)
Yeah, I love it. So now, ⁓ you know, look, looking from the you know, there’s a lot of marketing leaders and execs listening to this. lot of people at enterprise orgs that are, you know, any company, I would say probably falls into this. But, know, those those large orgs, they’re being asked to do more with less. They’re being asked to consolidate. They’re being asked to do all of these things. You know, what what advice would you give to those marketing executives that are
You know, they may be excited about AI, also, you know, kind of overwhelmed by the rapid pace of change. And, you know, what, what, what should they be focused on?
Rafael Flores (18:32)
Yeah, good question. And I’ll tell you what I often hear, right? You have the CMOs, the CXOs, even CDOs, where the CEO is telling them, we need AI first vendors, right? Because the board is telling the CEO, you need to be AI to drive valuation, right? So we have to think about this unit of economics that happens here, which even happens to me in my role. But a lot of them don’t even know what AI is, right? Or how to actually use AI to drive profits and revenue, right? Which is ultimately what you’re trying to do. And this is why so many companies are actually failing because
They’re trying to deploy AI and it’s not repeatable. It’s not credible. It kind of reminds me of the internet of things, which I was at the forefront of that when I let teams that aren’t holdings for it. Right. And so my advice will be honestly two things. Number one, immediately set up an AI use case review committee. There’s a lot of focus around AI review committees when it comes to governance and is the AI safe? Can I trust it? And that’s extremely important. That’s a piece of the iceberg.
Why are companies not also focusing on how they can actually use AI in one or two ways that will drive revenue? And that involves a lot of functions coming together, right? So do that now. So when you go evaluate vendors, there’s enough meat to the bone of understanding what AI will do for your business. So that will be the first step I will say. The second step is solve for one, don’t solve for many. Focus on fixing your data and feed in the right data for that one use case. If you try what I’ve seen, and we’ve seen also at Treasure Data with our customer base, when they come with all this data and they just want to fix it all and then deploy AI, it’s too late. Somebody has already beaten you to it. And so you have to make sure you get your data right, but get it right for the credible use cases that are top of mind for you and that are going to drive business impact. That will be my advice. Don’t boil the ocean doesn’t work, there’s a reason that phrase exists, right? It’s the number one advice you get all the time.
