The New Frontier of Personalization: Juan Mendoza on Agentic AI, Consumer-Led Journeys & the Data Black Box Ahead

ICYMI: Juan Mendoza on the Rise of Volitional Personalization and the Collapse of Traditional Martech Assumptions

“Consumers using AI agents to do their bidding for them is an entirely new channel we have no idea how to grapple with yet.”

In this episode of 10 Minute Martech, Juan Mendoza, CEO of The Martech Weekly, unpacks one of the most consequential shifts in digital experience: agentic AI—AI agents acting on behalf of consumers.

Forget personalization as we know it. Forget tidy funnels, predictable journeys and clean attribution. Juan argues we’re entering a new era where customer experience may be driven less by marketing automation … and more by the AI intermediaries consumers choose to trust.

This isn’t theoretical. It’s already happening! In product search, in real estate, in ecommerce and increasingly inside tools like ChatGPT and Perplexity.

Juan blends optimism, realism and a dose of existential dread. But more than anything, he argues marketers need to rebuild their assumptions from the ground up before the ground moves again.

Juan’s Big Idea: Volitional vs. Speculative Personalization

Juan introduces two concepts reshaping the future:

Volitional Personalization (Pull)

Consumers initiate the experience through AI agents by asking conversational, nuanced questions they’d never type into Google. This creates rich, unstructured, deeply personal data you’ve never had access to before.

Speculative Personalization (Push)

Marketers build experiences based on known data and planned campaigns—the traditional model.

The Breakthrough?

Not choosing one. But actually integrating both to create more intuitive, more contextual, more human experiences.

10 Memorable Moments with Juan Mendoza

1. Agentic AI Is Reshaping the Martech Stack

AI agents aren’t just automating tasks. They’re redefining workflows, strategy and even job design across marketing organizations.

2. Marketers Must Sort Hype from Harm

Juan sees both extremes:

1. Tech vendors racing to embed AI agents everywhere.
2. Companies seeing real gains from automation.

But not all outcomes are positive, and not all use cases are ethical.

3. A Real-World Example: AI-First Personalization in Real Estate

A property marketplace Juan works with has fully embraced agentic AI for personalization. Consumers ask conversational questions (“I need a backyard this size, a living room like this …”) and the AI delivers precision-matched results. This is personalization marketers could never build manually.

4. Consumers Are Already Using AI to Shop

Juan searched ChatGPT and Perplexity to find a very specific alarm clock that wouldn’t wake his wife. AI delivered. These tools are becoming new discovery channels, one’s brands can’t see into.

5. The Data We’re About to Collect Is Unlike Anything Before

Conversational queries reveal desires, preferences, constraints and contexts consumers have never shared in traditional search boxes. This gives brands unprecedented insight but also unprecedented responsibility.

6. AI Agents Will Soon Talk to Brands on Behalf of Consumers

Google has already demoed AI agents making phone calls for users. Brands must prepare for a world where your customer isn’t the one contacting you, their agent is.

7. The ‘Dehumanization’ Narrative Is Too Simple

Juan believes agentic AI may actually humanize customer experience by giving brands better context, clarity and empathy, if used responsibly.

8. The Biggest Hot Take: Your Data May Not Matter Anymore

If consumers shop through AI agents, your carefully built martech stack like emails, funnels, CDPs, onsite experiences might never be seen.

Only your product listings and structured data may influence AI decisions. Everything else becomes invisible.

9. We Are Entering a Data Black Box

Unlike SEO, where we can inspect rankings and visibility, AI agents reveal nothing.

You don’t know what queries surface you. You don’t know how you’re ranked. You don’t even know if you’re in the conversation. This is the part that “keeps him up at night.”

10. This Is a Moment for Builders, Not Playbooks

There is no textbook for agentic AI. No authority. No standard.

Juan urges marketers to engage with the builders, engineers and communities shaping AI today, because the future is being written in real time.

His go-to hub: Hacker News, where developers share unfiltered insights from the front lines.

Juan’s Martech Hot Take

Agentic AI is creating a customer experience channel that marketers cannot see, track or influence yet may soon dominate buying decisions.

If consumers no longer interact directly with brands, the martech function becomes disempowered. It’s a future few want to imagine, but one we need to prepare for.

Who Juan Follows

Instead of individuals, Juan turns to Hacker News, the Y Combinator–run hub of engineers, founders and AI researchers sharing raw insight on what’s actually happening in Silicon Valley and inside AI labs. It’s where he goes to learn the things no marketing thought leader can teach yet.

Listen to the 10 Minute Martech Episode Now

• Watch on YouTube

  
  

• Listen on Spotify

• Find it on Apple Podcasts

Next Up in the 10 Minute Martech Series!

Sara sits with Hazem El Zayat, Chief Experience Officer (MENA) at Ogilvy, explores why AI’s true potential is still largely untapped—and how a strategy-first mindset is essential for making the most of martech.

Want to keep reading in the meantime?

Full Episode Transcript: Juan Mendoza

Here’s the full transcript to keep you transfixed. Every insight, every quote, unedited and unforgettable.

Sara Faatz: I’m Sara Faatz, and I lead community and awareness at Progress, and this is 10 Minute Martech.

Juan Mendoza: This sort of volitional personalization, consumers actually using AI agents to do their bidding for them, is an entirely new channel that we have no idea how to grapple with yet. And as a marketer and marketing technology leader, it’s worth stepping back for a minute and going, in the future, how much of this customer experience will be reliant on my own data.

Sara Faatz: That’s Juan Mendoza, CEO of the Martech Weekly. Let’s get started. So Juan, can you tell me what is your burning martech idea right now?

Juan Mendoza: Everyone’s talking about it right now, but it is agentic AI and its implications for the marketing technology profession and the marketing technology stack. All of a sudden, consumers and marketers are using these tools to actually automate a very large swath of their work as well. So that’s sort of the big burning topic I have at the moment, is the role of these AI agents and how much of it is actually valuable and how much of it is actually quite harmful as well, both for consumers and brands.

Sara Faatz: Yeah, very interesting. Help me unpack that a little bit. What do you think is, where is the danger or what should people be concerned about?

Juan Mendoza: On one side of the marketplace, you have all this anxiety around, well, these AI agents and a lot of these larger technology businesses are now looking to actively supplant employment and career opportunities in favor of AI agents and the efficiency it creates. On the other side of the marketplace, we have this extreme hype and enthusiasm around this stuff, from Mark Benioff getting on stage and announcing Agent Force and saying that, yep, you can use these AI models to basically drive a lot of marketing and sales use cases to almost every single technology vendor that we’ve been reviewing, and I’m talking hundreds and hundreds of vendors, embedding agentic AI capabilities into their platforms, and then working with some companies and brands actually seeing some amazing results through AI automation, from small stuff such as content automation and decisioning and versioning to big stuff like customer data management, personalization strategy, and actually delivering new consumer tools as well. It’s not like this technology trend is something where it’s all negative. It’s actually positive and negative, but sorting through, and what we’re seeing with a lot of companies and our clients and people that we work with, a lot of them are trying to get through that and think through ethically, morally, but also commercially, practically, creatively about how these tools should actually be working in their business.

Sara Faatz: Yeah, that’s super interesting. Do you have examples that you can share with us about somebody that’s actually doing it really well right now?

Juan Mendoza: So this organization is a large real estate marketplace. So basically, they have a lot of consumer data. They have a lot of financial data, and their product is really about helping consumers find their ideal property, their next property. And we sort of work with them, and we’ve been looking at, okay, well, they want to embed agentic AI capabilities and large language models into their personalization customer experience stream of work. Initially, when we sort of worked on this idea, we looked at it as a dichotomy. So what I mean by that, there was two camps. So you have what we call volitional personalization, which is a better way to sort of explain that would be kind of a pull type personalization, where a consumer may use an AI, for example, a chatbot on a website or a tool on a website, to get a personalized experience through a large language model. They would ask it, hey, where I’m actually just looking for these specific features and this specific criteria from the thing I’m looking to buy. In that example, the AI model would use their probabilistic reasoning, and they would provide those options.

Juan Mendoza: And now that is actually a format of personalization. Can I tell you one more story on this, because I think it’s a really interesting sort of domain. I’ve been looking for a new alarm clock. Like a lot of people these days, I’m sick of waking up and the first thing I see is my phone, because my phone is an alarm clock. And I was just looking for a very simple alarm clock with a specific feature. The one feature was I didn’t want the alarm clock to be so loud that it would annoy my wife. So I actually went on to ChatGPT and I searched, I’m looking for these specific features for alarm clocks, and lo and behold, I found it. There’s other tools like Perplexity. You can go into Perplexity, and they’ve just actually announced their own e-commerce integration with retailers. This is an entirely new channel for customer experience. And that’s the thing that a lot of brands need to grapple with.

Sara Faatz: Where do you see the human element in all of this? You’ve talked about consumer experience and things like that. Talk to me a little bit about how you think humans should be interacting with AI and what the humans can be doing to make sure that we’re optimizing all of our experiences.

Juan Mendoza: Yeah, it’s a great question. I might actually take that back to the second part of the marketplace story, which is this marketplace. So we’re looking at that pool. So consumers pooling experiences from themselves using these AI agents. But then there’s also what we call the speculative personalization, which is marketers sitting in a room and going, okay, we’re going to take this data that we know about our customers, and we’re going to do this campaign, this use case, and actually push something out. So it’s a push motion than actually getting an experience out. And that’s been the default way in which most customer experience strategies have come together is it’s all push. So where this comes together, and I think this helps answer the question about the human element in all of this, is that for this company, the big opportunity is not in this dichotomy at all. It’s integrating them together. So you can collect a deeper level of customer data. People are asking extremely different questions to a large language model than they would ask Google search, a search box on your retail website or your marketplace website. And so the big opportunity there is that in this sort of sector of real estate is that you’ve got this opportunity where people can converse with these AI agents and actually share more information about what they’re really after.

Juan Mendoza: I want this house with like, I want at least this size backyard or I want this sort of living room space and this sort of, that’s what’s possible now. And so all of a sudden all these brands have access to this unstructured, qualitative, but extremely valuable data about their customers. And I actually think that that will help. One of the big blockers between a marketer and a consumer is a lack of knowledge and a lack of actual context to meet them where they’re at. Brands should be able to use that data to actually create a better, personalized, less friction, overall better experience for the customer. And so that’s the thing, like the big narrative out there right now is this story about, this is dehumanizing. People will lose jobs and then you’ll never get another human on the other end of a phone call anymore, right? But guess what? These AI agents, there are tools and products out there where consumers are going to use AI agents to call brands anyway. And Google has been showcasing that. So there’s actually consumers going, no, this is going to be more convenient for me and easier to do.

Juan Mendoza: And then there’s going to be a lot of automation on that side as well. So that’s the big sort of story there around, I guess, human interaction, human experience in relation to brands. I actually am pretty optimistic that it could actually lead to a better, more humanized experience between the consumer and the brand.

Sara Faatz: I love that. And especially if we can normalize, I think my Martech hot take, I usually ask what other people’s is, mine is going to be that we should normalize speculative and volitional personalization because the combination of those two is exactly what you’ve just talked about. And I love both of those terms. That’s awesome. So I guess leading into that, I’m going to ask you, do you have a Martech hot take outside of anything else that we’ve talked about today?

Juan Mendoza: Yeah. Well, we actually recently just did an essay on this, which really tracks the evolution of AI agents right now and what’s hype and what’s not hype. And it’s a fantastic essay from Keanu Taylor. So he’s our global head of research here at the Martech Weekly. And we spent a lot of time talking to large enterprise brands and global companies and how they’re thinking about embedding these technologies. And some of the biggest feedback off that is that this stuff is not just another marketing tool. The paradigm shift, and this is my big hot take, is that in the future, a lot of the data you’re collecting, a lot of the customer experiences you’re creating, a lot of the automation and the tooling and the marketing technology integrations and all of those things may not matter so much. Because if you take that example of me finding a radio and an alarm clock so I can wake up early in the morning, the only thing that matters in that equation with me asking Perplexity or ChatGPT to find products for me, the only thing that matters is the data that’s exposed by the retailers. The listings on Amazon, the listings on all the other e-commerce websites.

Juan Mendoza: Because the experience that I’m getting is personalized and tailored to exactly what specifications I want for a product. And as a marketer and marketing technology leader, it’s worth stepping back for a minute and going, in the future, how much of this customer experience will be reliant on my own data? And I would say my biggest hot take is that this sort of volitional personalization, consumer experience, consumers actually using AI agents to do their bidding for them is an entirely new channel that we have no idea how to grapple with yet. Because unlike SEO, you can actually peer into their search console and see which content is trending and which isn’t, which is getting ranked and which isn’t. Right now on all these mainstream platforms that have now hundreds of millions of users, it’s a black box. You have no idea where you’re visible. You have no idea what people are asking and there’s no marketing visibility at all. And I think that is something that, going back to the top of this conversation really keeps me up at night. Because if all of a sudden there’s no data, there’s no collection and customers are not even buying stuff on your website anymore.

Juan Mendoza: And they’re using AI agents as a gateway to do that, it means that the marketing technology function would be very much disempowered. So it’s not a very happy story, but we needed the hot take because people would normally think like, yeah, we’re always gonna have data, we’re always gonna have marketing and we’re always gonna have email, probably a future where that may not exist anymore. So that’s my hot take, which is a little bit depressing.

Sara Faatz: You know what, I think Juan, we should set a date a year from now. I mean, the way things are changing, we could probably set a date for two weeks from now and have a totally different conversation, but I would love to see if what you’re saying does come to fruition. I think it’s an exciting time. I think it’s a time that we all should be experimenting and trying new things because to your point, we don’t know exactly where things are going, but I do love that hot take. One last question for you and I’ll let you go. Who do you follow for information or inspiration?

Juan Mendoza: You know what I’ve been doing recently, it’s no one individual, but I’ve been on Hacker News almost every day at the moment. So Hacker News is a website that was founded by Y Combinator and Y Combinator is one of the really famous Silicon Valley incubators for startups. But they’ve created this blog and it looks like it’s from the '90s, it’s really ancient. But Hacker News is quite cool because it’s giving me probably the best inside view of what’s going on in Silicon Valley right now, particularly around AI and agenting AI. It’s like my number one spot right now for this content. And it’s a milieu of everything, you know, blog posts from developers, inside scoops to tech companies, you know, it’s a really fantastic resource. And I’ve seen for my own educational learning around this, it’s been kind of my main place where I would actually go and learn about this stuff because it’s happening. There’s no textbook, you know, there’s everyone wants to say they’re a thought leader and they’ve solved it or they’ve figured out the best framework for this, but we are in a constant state of flux here. You know, there is no normal, there is no authority right now around this stuff.

Juan Mendoza: It is a total world of experimentation. So my advice to listeners of your podcast is find those pockets where you have a lot of nerds talking about how this stuff actually works in practice and get around it. And for me right now, Hacker News has been a fantastic resource for that.

Sara Faatz: I love that. That’s great. Thank you so much, Juan. As always, it was amazing talking to you and I really appreciate your time.

Juan Mendoza: No problem. It was an absolute pleasure to talk about such a huge topic. So thanks for having me.

Sara Faatz: Listeners, thanks for tuning in. Remember to like and subscribe wherever you get your podcasts. Until next time, I’m Sara Faatz, and this is 10 Minute Martech.