What's Up with Tech?

How An Agentic Platform Turns First-Party Data Into Loyalty

• Evan Kirstel

Share episode

Interested in being a guest? Email us at admin@evankirstel.com

Most brands pay again and again to win back customers they already won. We wanted to fix that. In this candid conversation with Sandeep Menon Co-Founder Auxia we unpack how an agentic AI platform can sit on top of rich first-party data and finally make it work: autonomous decisions about timing, creative, channel, and offer that meet people at the right moment across web, app, and owned media. No more brittle, rules-based journeys that spam everyone. Think continuous learning, clear objectives, and experiences that build trust.

We talk through the reality of today’s martech sprawl and why so many marketers feel like systems integrators. Sandeep shares a practical model for role compression, where a suite of agents handles the repetitive work and analysts’ chores, while the human team sets goals, brand guardrails, and strategy. He breaks down the “reacquisition treadmill” and why acquisition gets the spotlight while retention gets sidelined, even though the biggest gains often come from deepening relationships with people who already converted.

You’ll hear a tangible story: a large C2C marketplace used cross-category nudges to move fashion buyers into electronics, driving an 84% lift in customer lifetime value. We dig into how the decision agent and analyst agent work together, how marketers can query outcomes in plain language, and why the next two years of AI—especially improved reasoning—will reward teams that measure success by revenue impact, not pilot buzz. If you’re ready to turn first-party data into loyalty instead of noise, this one’s for you.

If you enjoyed this conversation, follow the show, share it with a colleague who lives in lifecycle or growth, and leave a quick review so others can find it.

Support the show

More at https://linktr.ee/EvanKirstel

SPEAKER_00:

Hey everybody, fascinating chat today in the world of marketing as we dive into an agentic AI platform that helps brands keep customers before they drift away in incredibly challenging problem-facing marketeers. Sandeep, how are you?

SPEAKER_01:

Evan, first of all, thank you for having me on your show. Great to talk to you today about this exciting topic and about Oxia.

SPEAKER_00:

Yes, indeed. And congratulations on the fundraising and growth at Oxia. Before that, maybe introduce yourself, your journey to Oxia. And what was the big idea, the big problem you're solving for marketeers?

SPEAKER_01:

Yeah, um, so um at a very high level, Oxia is an agentic uh journey orchestration platform. So what we do is we have the ability to kind of sit on top of the rich first-party data that most large enterprises have, but unfortunately not have the unlimited data scientists or data engineers to really use that data effectively. And then on top of the signals that we have from this data, have a suite of AI agents which work hand in hand with marketers and empower them to kind of ensure that you're able to deliver the right experience to each customer in a hyper-personalized manner throughout their journey on your product so that they are happy, satisfied customers who stick with you. And um, to your question on um what um prompted us to do this, I have been a marketer for um over a decade. I spent um um I spent over a decade at Google in a variety of marketing roles, used to lead small business marketing for Google, um, and then was CM of Google India for a few years and moved to the US to lead all of our platform marketing. That's Android, Chrome, Chrome OS, and the foray into payments. And it was very clear to us that we are about to endure this decade of transformation driven by AI, and it will definitely change how the best marketers and the best growth leaders um leverage AI to drive growth on their products.

SPEAKER_00:

Fantastic. Well, what a what a journey. And in plain English, you know, what does an agentic AI marketing platform do beyond the buzzwords? What does it mean in practically speaking, maybe on a day-to-day even uh basis?

SPEAKER_01:

Yes, um, I I think that's a very, very valid question, especially nowadays where there's so much hype and very often it's very difficult to separate um um hype from um reality. In a very simplistic manner, the big challenge that we are solving for is that most enterprises, most companies now sit on a treasure trove of first-party um data. Like uh I've seen um spoken to many CMOs. Um, I think there have been a lot of published studies on this. Uh, most large enterprises use less than 30 to 40 percent of all of the data that they have effectively. And um what we do is number one, we enable marketers to kind of leverage and harness the power of that data. And um, where the agentic aspect of it comes in is that um we then have, once we have this um rich signal repository that we create out of this data, the ability for our agents to autonomously decide what is the right experience that you show to a customer at any point in time in their experience with your product on your website, on your app, or through all of your marketing channels to determine what is the right action, right creative that you need to promote at the right time to get them to the objective function. So that is the foundational aspect of our product. Um, there are, of course, um a few more agents that we build on top of it, which I'm happy to kind of talk about.

SPEAKER_00:

Oh, yeah, and we'll dive in. So the marketing uh tech landscape is incredibly complex and diverse. You've seen these Martech uh landscape pictures with thousands of companies and even a little small business like mine. I have a dozen tools in my stack. How do you fit into an enterprises or you know any business marketing stack today?

SPEAKER_01:

So um I think it's important that we take a step back because you raise a very valid point. And I have experienced this as a marketer, um, as a CMO. And when I talk to other marketing leaders, um, whether it's in a small business or in a large business, the challenges still remain the same. I do think that there has been a proliferation of point solutions, and very often marketers today um are um uh system integrators um and not spending enough time doing the core of what they do, which is marketing. So uh I think um that is a real concern. Um what this current way of AI really enables you to do is what I call role compression, and you're seeing this happening in um in multiple domains. So our belief is that uh in the coming years, um you would move from the current state where you have marketers using point solutions to do discrete manual tasks, right? And um rules-based tasks. Like I'll give a simple example. Let's say you're responsible for doing a lifecycle campaign and in a bank, right? So you have a set of marketers who um work with um I don't with a set of people to determine what are the rules. Uh like Evan has come in today, uh, he's a new user. What is the email I give him? Three days later, what is the next thing? It's all based on rules, right? Like, and what we believe is that going forward, um, um with agentix solutions, you see two things happening. Number one, as what I said, role compression. A marketer is able to do so much more because um, and and the work that they do is so much more strategic. And the reasons for that is that we believe in the future the marketer will be able to work with a suite of agents, and these agents will replicate a lot of the manual tasks and the tasks which they currently do not have support for.

SPEAKER_00:

Brilliant insight. And you talk a lot about something called the reacquisition treadmill. Uh, sounds intriguing, uh, a great phrase. I'm not sure if you coined that, but how did that idea come about and what what does it mean practically speaking?

SPEAKER_01:

Yeah, so um if you uh if you think about where the innovation has been in marketing in the last, I would say, two decades, I would say the biggest innovation has been in programmatic ads, right? So um, wherein as a marketer, you can assign a budget and you can pretty much say, uh, hey, for this budget, Google Meta, bring me the traffic that I need to my website, to my product. That is an amazing innovation. That is what um helped Google and Meta become what they are. Um but what happens after that is equally, if not more, important. What I mean by the reacquisition treadmill is that in many businesses that I speak to, unfortunately, even after you have acquired a customer, a prospect, even after they have even like let's say a typical retailer, even after have they made a transaction with you, bought a product from you, very often because you have not cultivated a relationship with them, a deep relationship with them, you have to actually go back now and then spend money on ads again to reacquire the same user, which I think in many ways is a failure of marketing. Yeah, really, really well said. And that is what we aim to also solve is like once you have a relationship with a marketer, how can you move beyond it being transactional? How do you kind of really understand that customer so that Evan doesn't get spammed with like three more emails, promotions and products he doesn't want, but the brand has a deep understanding of you because you bought from them, you've interacted with them, and they are able to craft a message at the right time if there is a need to share with Evan. And I'm pretty sure Evan will appreciate that message. He would um have a much more deeper connection with that brand.

SPEAKER_00:

Yeah, that's a great point. And you know, acquisition, customer acquisition gets all the limelight. And of course, it's very sexy to show new customers, new wins, new logos. Why do you think retention gets less attention? And why is it sort of the uh redheaded stepchild of marketing?

SPEAKER_01:

Um the the it goes back to what I said, right? Like it's not that acquisition. Um acquisition, Google and Meta, wonderful platforms, have made acquisitions truly easy, right? So um my first job at Google was to convince small businesses about the power of the internet and get them to advertise on Google, right? Like, and like I remember talking to small businesses in the UK and saying, hey, you can get customers from anywhere in the world, right? And they were like, really? Uh how does that happen? And it might like when you talk to marketers now, they're like, what's the big deal? But that's a true revolution. And in many ways, the reason it gets all of the limelight is very effective. Uh and to be honest, most marketers um are going with the EC way. It's like as there used to be a saying that nobody uh fired you for buying IBM. Nobody's going to kind of fire you for um assigning a budget to big ad platforms. It is effective, and I do think it has a very important part to play. Um but I think it's um uh like in the future, I do believe with the current advances in technology and AI, especially with the rich data that most large customers, large enterprises have, you should be able to do a better job in in understanding your customers and building a relationship with them.

SPEAKER_00:

Fantastic. Looking at your website here, you you've seen an 84% list in customer lifetime value through your platform. That that's pretty impressive. Um, how did you get a result like that?

SPEAKER_01:

So um it it you you raised an important point at the beginning, right? Like there's so much um so much hype about AI, et cetera. One of the difficult decisions we took is that if you think about the AI um spectrum, right? So so um generally um there is a lot of focus on um on um let's say cost reduction um and um and I would say um uh resource uh replacement. Um what I feel is um uh equally important um and where the real value is if we can uh impact the top line. So a lot of what we focused on with with the work that we did with our agents was to drive um business impact. Right. So so in this particular scenario, we are talking about a very large consumer-to-consumer marketplace um where um we work with them to drive cross-category purchase. If you're a large marketplace, it's a consumer-to-consumer marketplace. So typically, if if I am a if I come there and I buy fashion, um uh then typically I continue to continue to buy fashion. If I can get uh a user who is um repeat user on fashion to then do electronics, there's a huge upsurge in the uh lifetime value. So what we did is um across their product experience at the right time um at the right moment, um, show um and determine what is the right subcategories or categories that you can cross-sell to and prompt those products along with the targeted promotion, um, so that the likelihood of them buying that at that point in time increases massively. And that's how they were able to get that uplift.

SPEAKER_00:

Fantastic. What a great story. You're also helping customers make better day-to-day decisions based on all that juicy data that you have. Um, how does that look work? How does it look like on a day-to-day basis?

SPEAKER_01:

Um I I talked about the core agent that we have, which is a decision agent. One of the things that we realized, and it also is borne out by my frustration as a marketer, is that you would often um you would often put out campaigns, you would often put out ideas out there, and then you would run it, and then you need to spend weeks, if not months, um either working with a data scientist or analyst um or spending um endless hours on Excel. And wouldn't be great, we thought wouldn't it be great if um if uh there is an agent that can understand all of the data, understand all of the decisions that the decision agent took. Um, and then um if the marketer can use natural language to query that um and give him further insights or her further insights, wouldn't be great. And that was the genesis. That's another agent which is very effective and which a lot of our customers use. Um, and the idea there, as I mentioned at the beginning, is our belief is that the marketers of the future will have this marketer agent interaction paradigm, and it needs to be a uh virtuous circle, it needs to be a uh loop wherein the agent suggests something, marketer takes a strategic decision based on that, uh, takes an action, then you have a new inside the marketer actions, and that's what the analyst agent does on a continuous basis.

SPEAKER_00:

Fantastic. What a groundbreaking concept. Um, so you just closed a$23.5 million round. Congratulations. Thank you very much. Um what are you focused on next with that funding?

SPEAKER_01:

Um I think um the um the innovation in our space um is um is truly um uh mind-boggling. Um I was lucky enough to be part of the um, I would say, the internet revolution during the early days of my career, um, the mobile and then the cloud um um um related transitions that happened as well. But what we are seeing unfolding before us is happening at a pace that is unprecedented. Um so a lot of our focus um is is remains the same. How do you kind of truly develop innovative products that can help marketers achieve their goals? Um, most of um our team's time is uh is focused on that aspect, which is helping our customers succeed and developing more products um in that direction.

SPEAKER_00:

Fantastic. And the amazing thing is we're just getting started, early days, I think, for a Gentec AI. Uh, how do you see things unfolding over the next year or two with all this investment and innovation? What will the world look like?

SPEAKER_01:

The um uh you're absolutely right. Early days, people often talk about um about cloud. And if you think about the transition to cloud started somewhere in the early 2010s, and even now, I think less than 30% of the enterprise workloads have moved to cloud, right? So I often tell my team that um for for those of us who can remember, we are still in the 1995 Yahoo uh web pages age of AI, right? Like so still a long way to go. Um, I think the uh the aspects that um um we are uh we are excited by uh the improvements in the reasoning capabilities of um models and um and the pace at which that is improving and how application layer companies that build on top of that um can uh can truly leverage uh that. Um I'm also equally um excited about um marketers and in general all businesses using um leveraging um AI products, moving beyond what I would call this tendency to just do pilots to really focus down AI products that more um revenue and have business impact. Right. So I think we are only starting and um we are already seeing many enterprises um uh now evaluating products with that dimension. I think it's a welcome change, right? So which is like what is the real business impact? How will it change my revenue, my top line, right? So and how can I measure that rather than hey, it'll make this process slightly easier, slightly better.

SPEAKER_00:

Yeah, well said well, thanks for sharing the insights and uh congratulations on the success so far. Much more to do, onwards and upwards.

SPEAKER_01:

Thank you. Thank you, Evan. Um, it was a pleasure talking to you. Uh I come uh I believe that um that we are just getting started, and I'm I'm truly excited about how the work we do will help marketers move from the manual uh repetitive tasks they do to become supermarketers um and do more strategic work.

SPEAKER_00:

Well, that's uh here here. I would I would second that and thanks so much for sharing uh the vision. And thanks everyone for listening and watching, sharing this episode. Also check out our companion TV show on Bloomberg and Fox Business at techimpact.tv. Thanks, Sandy. Thanks, everyone. Thank you, Evan. Bye bye. Bye.