Michael Hartmann: 0:24

Hello, everyone. Welcome to another episode of OpsCast, brought to you by MarketingOps.com, powered by other Mopros out there. I'm your host, Michael Hartman. Joined today by my co-hosts, Nami Lew and Mike Rizzo. Long time since we've been together. All three of us. All three. We're here. Well, today is a big, big episode. We've got a special guest. The traditional B2B marketing playbook is breaking down at the exact moment when AI is reshaping how buyers research, decide, and purchase. Many of the systems marketing operations teams built over the last decade were optimized for a world that no longer exists. And our guest today is very familiar with that, John Miller. John co-founder Marketo helped define modern marketing operations as we know it. More recently co-founded Engagio, and today he's the co-founder and CEO of some stealth AI startup. Maybe you'll talk about that. Maybe you won't. Maybe a little bit working with the intersection of AI, buying behavior, and revenue systems. So we're going to be talking about what no longer works, what replaces it, and why marketing ops may be uniquely positioned to lead in the next era outgo the market strategy. John, welcome.

Jon Miller: 1:25

Thank you very much. I'm excited to hang out with all of you.

Michael Hartmann: 1:29

Yeah.

Jon Miller: 1:29

Well, we're excited to have you.

Michael Hartmann: 1:31

Yeah. Well, maybe let's start with uh, I mean, you you've got a long history in this space. Walk us through what you're working on today and how that how things have changed, your perspective has changed and shifted over the time since starting and building Marketo and Ageo.

Jon Miller: 1:48

Sure. Well, yeah. So I mean, I don't mean to be too uh coy or stealthy by being in stealth. Um, it's just what I'm building is taking a very long time. So I didn't want to do a big launch and then have everybody sit around for 18 months saying, hey, where is the product? Um, but what what I'm doing is building a new startup that reimagines marketing automation for AI, the age of AI. Uh so it would be an alternative replacement to the legacy vendors, including Marketo. Um, okay. I I think that you know, the I'm proud of what we built at Marketo, and and I think the other platforms have been amazing. And yet I I see that they are really stagnating uh in the face of I think two pretty major disruptions that have happened. Um the the first is that in some ways the the old playbook that they were built to support, the playbook I helped you create, as you just said, I think ultimately was doing more to burn relationships than to build them. You know, the the perspective has, I think, shifted kind of away from hey, how do we generate MQLs to more of one of how do we build relationships with buyers, you know, even before they're ready to contact us? Um, and that's that's a pretty profound difference that you know the the the current generation platforms aren't really built to support. Um and then the second big disruption is pretty obvious, right? Which is AI, which changes both how we go to market and also what is possible. So in the on the former, we're not just marking to humans, we're marking to humans and AI, right? The current platforms aren't built for that. And then, you know, the AI on the on the latter changes what we can do. You know, the legacy platforms are really all rules-based systems, they're they're really fancy if this then that logic automation tools, um, which is useful, but isn't matched the real world. The real world is kind of messy and nonlinear, and there are all these complexities that you end up creating more and more spaghetti rules to handle every possible scenario in every possible case, and those change over time. Um, whereas reasoning AI can handle all that mess uh in in a much more profound way. So new things become possible. Um, and then lastly, I think you know, AI is it machine learning's been around in marketing tech for a long time, but nothing's really nailed it, right? We're still all talking more about doing one-to-one than actually delivering on one-to-one. And I think that the modern advances in AI can finally maybe unlock that as well. So I think it's time for some new platforms, and I'm trying to build one.

Mike Rizzo: 5:00

I'm super excited by it. Um I think the the sort of landscape shifts or the the you know, the two things that you talked about are are obviously like super spot on. You're you're well uh versed in the in the area of all this go-to-market technology, particularly on the B2B side. Um I'm gonna diverge a little from probably what was the next uh question that we had in our minds. But um, one of the things I just I just want to get a little bit of your your take on this, John, and you know, hopefully it doesn't uh force you to feel like you have to answer anything that you're working on by any stretch. But um one of the things I've observed in working with some of the more new startups, um, you know, they'll come inevitably like the community talks about what's going on and they're giving feedback. The market sort of is like in and out of our space all the time, talking about things. And and then I get to occasionally talk to some of the founders of these products or the product marketers. And the feedback that I had from uh a soon-to-be-released sort of product in the Adobe side of things, a couple of products that are already in market, uh, from some of the more modern AI forward go-to-market, uh, you know, marketing automation solutions, is like I I feel like nobody's approaching it from the marketer perspective. And by that I mean there is there isn't an established base of hey, what the heck does your company do? What does it sell? Who does it sell to? What are your price points? So that this AI sort of quote unquote native solution can can do the things that you're talking about. Like John, you're saying, hey, AI can do a really good job of being a lot more personalized. But the way to do that, we've all like fundamentally observed, is to give it the right context. Yet for some reason, every platform seems to miss the layer of like this is the context layer, and everything operates from here. Um, have you been sensing some of that in like maybe the legacy players? They're trying to make a shift in that. Like, what are your thoughts on I don't know, like that perspective? I feel like that's just missing.

Jon Miller: 7:19

No, I mean, at the risk of sounding right, context is critical, right? Like an AI agent without context will make you know very confident errors about your business. Yes. An AI agent with context will actually operate and act like a trusted employee who knows and understands what you do. Um, so I think that having that context slash governance layer, they kind of go hand in hand, uh, is an essential element of any agentic marketing platform. Um, I can't really comment on all the other folks building stuff, you know, because I don't know what they're building, but I know what I'm building, you know. Uh and you know, that context layer needs to cover a lot of things in marketing, right? You mentioned who's my ICP and what's my pricing and some of those types of things, and those are important. There's more tactical things like how frequently can you email somebody? Right? Yes. More importantly than how frequently can you is how frequently should you email somebody? Yeah. Right? Then you get even more tactile. When I build a new campaign, what naming convention do I need to follow to make sure that my downstream Salesforce reporting doesn't break? What exclusion lists do I always need to remember to include based on the geographies that I'm mailing to? You know, et cetera, et cetera, et cetera. Like you, if you don't capture all those things, you know, you haven't really streamlined the operations. Yeah.

Mike Rizzo: 8:58

Yeah. Well, I love I love that you brought that up. The the you brought it up. The context piece is certainly important. Uh and the the governance thing, I think just like gets everybody on the op side, their ears perk up, right? They get real excited about that piece. Because I feel as though that's a huge, huge unplug. And and truthfully, uh again, from what I've seen in what's currently in market, and you've you sort of said already, yeah, most of the products are doing marketering activities a little faster and a bit smarter with some AI uh capabilities behind it, but they're not really fundamentally changing the game, uh, I think at the at the governance layer that you're sort of playing at. Uh and so I and I I am really excited about that potential unlock for AI. So I appreciate you entertaining the question because I, you know, I wanted to get your take on, you know, what you think in that area. Totally. Anyway.

Jon Miller: 9:59

Yeah. And I think that the to your point about rethinking kind of what you know what what can work uh in the age of AI, I think the the most profound change that needs to happen is moving away from a campaign mindset uh and into a mindset that is more about um you know what is each customer's individual journey, um, which might be a mix of campaigns and other touches, and it might you know be mixed and matched across a bunch of different uh places. We we can talk, I think we should talk a lot more about what that is, but but but both the what the reason I bring that up now is because both that and the context and governance thing we just talked about have really, really profound implications on the job of marketing operations. Right. Marketing operations stops being, you know, the folks who just build stuff in Marketo, right? They become the team that is managing the context, managing the governance, you know, managing the business strategies that go into orchestrating these journeys. Um, making sure that the data that the AI is using to orchestrate those journeys is is clean and accurate and right. Like those that's that's still marketing operations. I don't think AI is gonna mean there aren't market operations jobs anymore, but I do think they're gonna change and become more strategic.

Naomi Liu: 11:33

I think the challenge, at least that I see, is that the people who are in these roles, we may become so-called experts in AI and are learning about all this stuff and like understanding all the tools and adopting, you know, this way of marketing. But unless we're also like bringing along the rest of the organization internally and educating them on this stuff, you're kind of operating in the silo, right? Because we may have all these like ideas and things we want to do and ways we want to go to market, but at least in operations, sometimes depending on how the organization is structured, you don't always have that control, right? Like there are people who are pure marketing ops who are, you know, they're they l they skew or lean more heavily to the IT side as opposed to, you know, folks in marketing ops who also do content generation and demand gen and you know, really focus on lead generation. I feel like there's gonna be like a pretty big challenge with how like there's a net, there's a there's a there's a gap, I guess. I don't know if that kind of makes sense to what I'm saying. It's like, how do we bring along the rest of the company to adopt that same mindset too?

Jon Miller: 12:51

I think it's really important. Like I'll give you a very specific example uh that I'm already starting to see. If you move to a world where it's less campaigns and more orchestrating customer journeys, each individual marketer's offer, like I'm running a webinar next week, you know, it becomes not who will get invited to this webinar. It becomes who could get invited to this webinar. Right. And and your built basically individual campaigns become a library of things that you could do and not will do. So that's cool, right? Because now mobs people are don't get into like all these crazy air traffic control disworries of like, well, I can't invite them to this because they got this and I have to handle this, right? That's just nightmare complexity that people shouldn't have to deal with. But what it does mean is now, to your point, the demand gen person who doesn't really care about communication frequency caps and making sure we're not overcommunicating. They just care, did I get enough people for my webinar? Right, are gonna go to that mobs person and say, Why did your AI not invite all these people to my webinar? Right. And the mobs person is gonna have to explain, well, because they got these other touches that were better and more relevant and higher priority for them. Right. That's to me a perfect example of the point you're making, right? Which is like you can do amazing new things with AI, but yeah, that demand gen person is gonna have to kind of change how they think, and mops needs to bring them along.

Mike Rizzo: 14:24

Yeah, and and it's it's no longer why didn't I get all of the people? It's I think that the KPI changes, right? It's to you know, and you said it without saying it. It's like you said you actually kind of said it almost directly, but it was, you know, I got the right people. Oh, I got 20% less to register, but 80% of them were the right people this time. And so, like, you lost the 20% that were maybe not the right fit, or this hopefully that would be a wonderful outcome.

Jon Miller: 14:53

And they'll and that one then got something that was better and more relevant for then.

Michael Hartmann: 14:57

Right, at the time, right? So this all gets back to what people are incented by, what they're measured on. And it it seems like that's a key part of what you talked about, John, that the the current go-to-market playbook is c is broken. Feels like what you described falls into a couple of categories. One is it doesn't recognize how the buying process has changed. Like that's one, right? It's not this very linear, like we can push people through a very stepwise funnel, it's more chaotic than that. But also the because that hasn't recognized, it hasn't changed how people are measured either. Like, what does success look like? And it feels like those two combined, unless you can break that. Is that I mean, is that what you're talking about when you say the playbook is broken, or is there more to it? What's your take on that?

Jon Miller: 17:00

Yeah, well, it's funny. I I got a whole discussion on LinkedIn about this like two weeks ago, you know, when I sort of like say, like, oh, buying has changed, you know, it's complex and nonlinear. And people are like, buying is always complex and nonlinear, right? That hasn't changed. I think what's changed is you know, like we used a linear model to understand buying, right? And and any model is sometimes useful to help you understand the world and may have limitations in how accurately it understands the world, you know, and like we see this in physics all the time, right? You know, like um economics, yeah. Like models can be useful, but they can also be flawed. And the linear model, like we wanted buying to be linear because it helped make us under it helped make it more measurable and more understandable, right? And that was the playbook was based on that model, but we now I think have come to the realization that that model is fundamentally flawed because it's not how people buy, and therefore, everything built from that model, like measuring people on MQLs, I think sort of um starts to fall fall away. There are a couple of other things, like so, so so a big part of it is just the mental model we used, while nice and simple and perhaps useful at the time, was just flawed. There are some real changes too that have happened. One, uh, tragedy of the commons, right? We see this all the time, right? When when one somebody hears a tactic work, everybody does it, and lo and behold, it stops working. And we've seen that time and time again, you know, and then I think buying has changed to the degree that there are generational and uh behavioral shifts, especially that we've seen post-COVID. Um, if it's as simple as you know, Gen Z doing more of the buying and they're gonna do more self-service re you know research. Um you know, AI changing buying behavior. So there are some things that are different. But but yeah, though it's fundamentally it's the model.

Mike Rizzo: 19:15

Yes, uh yes, there are things that are different. I I think I think it's uh minimally the the best word to describe it is the entire thing is complex. Uh but while true that uh buying cycles were complicated before and they've always been complicated, I do think that there was a point in time where the myth the maturity of the market and the way that people sought information and bought, kind of like what we're talking about now, was totally different before. Before it was, I've never done mass email before. I've never done marketing automation before. In fact, I just want that capability. And it was very linear. It was I've never done it to zero to one, right? And so, like, it was like I'm on the hunt for the thing that lets me send emails to everybody, and then it was like, I'm on the hunt for the thing that makes me do that smarter. And then it was, oh, I need to score these people. And so, like, the maturity has progressed over time so that now people understand the art of the possible and they understand that the landscape of opportunity is much wider than it once was previously because the technology evolved. And so, yeah, the buying was complicated, but it was actually more simplified before because it was like I didn't have this capability before, and now I want it. And today it's like, oh, that capability is in like almost every tool that I want to go buy. And so the complexity has been brought on by the fact that it's actually more abundant and it's more accessible, and now I'm more educated and I'm watching what other people are doing and it's working for them. So now I want to replicate, but for my unique business. And so today I think it is more complicated than it was before. Sure, it was complex, but I think it's actually a lot more complicated now than it used to be. That's my perspective.

Jon Miller: 21:02

Yeah, the um tyranny of choice combined with the tragedy of the commons. Yes.

Michael Hartmann: 21:09

Well, and what you what you described, like I like to call the fallacy of best practices. If like if we all do the same best practice, then like the what you said, the tra I think that's the leads to the tragedy of the commons, right? Everyone gets like the least common denominator or least common quality results from that. Um so you said that one of the things you're working on is kind of a new ver new new iteration, new uh kind of marketing automation. I don't even know if you want to call it marketing automation, but like what do you think you know, why do you think that the we've gotten to a point where the marketing automation platforms you know, I think you said is stagnated. I I would tend to agree with some exceptions here and there, right? It feels like it's been spotty where there's there's been evolution on those. What's what do you think has led to that?

Jon Miller: 21:58

And what do you maybe I think it's a pretty classic story, right? You know, one acquisitions, you know. I mean, when Marketo is standalone, you had you know Phil Fernandez, who is incredibly strong and detail-oriented CEO making sure the innovation was happening, right? Then Adobe acquires it, and I think the most senior person at Marketo is a director of product marketing or something. Um, and that's just a symptom of you treat this thing like a cash cow rather than investing in it. And you know, and um, and and and then what happens inside Adobe, right, is that the the incentives for them are to build additional additional ancillary products, you know, so they can sell more. So now you're buying CDP and you're buying a journey optimizer, and you know, those are very expensive and not for every Marketo customer. And and the net is that the incentive is to not really invest a ton in updating the core platform. And you see the same thing at uh Eliqua inside of Oracle. You see the same thing with Pardot, uh, with Salesforce investing in just all the different flavors of marketing cloud instead of improving the core partot platform. So, like if you're on Pardot, you have an interesting challenge.

Naomi Liu: 23:21

Like you have a lot of comments about that, by the way. Or I'm like that's a separate podcast, I think.

Jon Miller: 23:27

All right, fair enough. But if you're on Pardot, you have a challenge, right? Because like you have a migration coming your way, right? You're going to have to do a migration. So should you migrate to Salesforce Marketing Cloud next? Or Salesforce Marketing Cloud Growth Edition? Or should you actually like consider what are the other new AI native alternatives that are really focused on your unique needs? That's an interesting question. So that's one is the is the acquisition. And I think the reason why you see that in reality is what's the one of the four legacy marketing automation platforms that's still innovating? It's HubSpot, right? The one that's sort of still mostly standalone, even though they are arguably inside of a larger CRM company now. Um I think the other factor that's just worth, but two other factors are worth briefly mentioning. You know, one, these are 20-year-old platforms. You know, I like to tell people we started Marketo the same month my son was born, and his birthday is in two weeks, and he turns 20, right? So these are 20-year-old platforms, like architectural legacy makes it hard to inundate. And then there's a more subtle one going on too, which is perverse. But when Marquetto and the other platforms are complicated, it creates a whole industry of marketing operations experts and consultants whose business is I know how to make this complicated thing work. Uh and that creates lock-in in some way, because there's this network of people who who are sort of their livelihood is based on the thing not getting better, you know, almost if you will. Um you know, and so effectively, why make it easier or better to use when the complexity itself is profitable and it makes the switching costs feel higher?

Michael Hartmann: 25:21

Yeah. Yeah, it's I think that's one of the ironies that I've realized recently is that the attacking automation onto the word marketing and call it and thinking it's just gonna do its own magic, right? That that's been one of the fallacies around marketing operation, marketing automation, I think, is that it just it's automated when it in reality it requires, especially like to your point, right? People to maintain it. And as soon as you add on the next thing that integrates and the next thing that integrates, and that you start adding more complexity to all that, then it's a recipe for things to break down.

Jon Miller: 25:59

Agreed.

Michael Hartmann: 26:00

So uh I wanna I'm curious about we've talked a lot about AI and you're too since you're building something that's sort of AI native. And I think you've hinted at one of the examples, right? I think I think a lot of people in this in this new world, if there's some sort of black box, you know, if you will, AI, that makes the decisions about who gets what when, because it's got some insight that maybe is not obvious to the humans that would have otherwise been doing it. Trying to explain that's gonna be hard, right? So whereas it used to be rules-based versus I don't even know what you would call it with the AI-driven one. I think that's one challenge. The other thing that just has hit me in the conversation, I had another one not too long ago with somebody, it kind of hit me. We talk about AI and marketing as if it's just one thing. It feels, but it feels like there's at least two major categories. One is the stuff that is, you know, customer-facing, whether it's evaluating the activities and things that people are doing or better targeting, but there's also internal operational stuff, right? Improving content, you know, accelerating that, use reuse of of content. How do you like how do you think about like the challenge of explaining AI output to people who are used to very you know rules-based, specific outcomes? And then how do you think about that? Like, do you think about those two different ways to major categories of AI?

Jon Miller: 27:19

Yeah, I'll answer that second one first, which because it's easy. Like I think absolutely you need to think about AI and both in terms of how it's going to drive more efficiency as well as how it's gonna drive more effectiveness. You know, how can it, on the efficiency side, how can it streamline campaign building and list uploading and all these other things that take mops people potentially hours currently? And that frankly is pretty tactical, you know, and that mops could should be doing better. Uh, that you know, the effectiveness side is how can I actually deliver, you know, the the proverbial right action at the right time, you know, to each person that's going to you know be the right touch to sort of move that relationship forward. Uh they kind of both come together, the efficiency and effectiveness, around that air traffic control concept that I talked about earlier, right? Where MOPS people are spending an awful lot of time managing the rules and the smart lists for who should get what campaign when there's overlap and different people who want to market to the same buyer, you know, and like that's that's that's both time consuming and annoying and also limits your effectiveness because first come, first serve is usually really the best uh you know strategy there. In terms of your other question though, around um I I think you know I like I like using playlists as an analogy, you know, like because we all understand Spotify. You know, and you know, like well, we most of us have used Spotify or or or some other tool like that, you know. First off, Spotify is not creating a new song for each person, right? It's the same library of songs, right? In some ways, you could think of that as a library of pre-approved songs it is allowed to play. You know, um, what it's doing is it's picking the best playlist of songs for you based upon what you've asked for and what you've engaged with and what you've liked to done you know do in the past. Right. I think that's a powerful metaphor for how we should be thinking about marketing journeys, right? Like, is it useful to write you and me get a different email because I posted something on LinkedIn yesterday and you posted something different? I don't know. That doesn't feel very valuable to me, really, right? But to really know like where I am in my journey and what I've done and what I'm interested in and what I've engaged with, and to know I tend to respond to long emails and not short emails, and to know I tend to respond on Sunday mornings, you know, but not you know, weekday evenings, and to just kind of build, you know, know all that stuff over time and then use that knowledge to put together my playlist, you know, that to me is really where AI effectiveness is going to really, really sing, you know, even more than like, let me write you a customized email, or like, or or what the state of the art seems to be today, let me write a token, a string that's customized, and I'll stick that in for you and call it personalization. I was like, all right, fine. I'm not saying don't do that, but like I don't think you're gonna like win the war with that tactic, right?

Michael Hartmann: 30:41

It's um so it's interesting because I think a lot of people the way I just this is all real time for me unpacking this, but it feels like people are gonna react to it as it feels intuitive when we we all say we want to be data driven, but in fact, it sounds like actually if you've got the engine and it's working correctly, it's gonna be more data-driven than we will be, but it won't feel like it because it's maybe not matching what we believe intuitively is the right next thing to do with that particular prospect or customer.

Jon Miller: 31:12

Yeah, I mean, like, think about what are we doing today? Or we're writing these journeys, like okay, like if somebody is a VP in financial services, this is the journey. But then if they're a director in financial services, they're gonna have this journey. But then if they show interest in my WYSIWANG B product, then I'm gonna take them out of that journey and put them into this journey, right? And you might think you know all these things, but just imagine the spaghetti of actually building all that. Right now, compare that to an AI, right? The problem with traditional machine learning is that to build a model, we had to create independent attributes that we could train a model on. And those independent attributes had to be uh relatively low cardinality. So, you know, you would take all this rich behavior we have about somebody, which what web pages do they go to, what content do they download, what events do they respond to, what time of day do they open things, you know, all the things I just alluded to. We would squash all that down to some number of attributes, like number of emails opened in the last week. Like that's all we could do before. And that's why this stuff wasn't satisfying in the past, and probably why marketers did go and try to create their own custom journeys. But we've all used our LLMs in our day-to-day lives. We know they can take a massive amount of context and make pretty interesting reasoning choices, you know, around how to how to interact with somebody, you know, that that can literally, if it's a big enough context window, it could know your entire behavioral history, right? And and consider all of that when deciding kind of what to do with you. Of course, that's going to be better than you know, a human who by by just by definition will be limited to a manageable number of tracks. You know, like eight, 10, 12, like like there's only so many a human can actually craft themselves, you know, versus effectively an infinite number of possible tracks when you let the eye mix up the different playlists.

Michael Hartmann: 34:42

Yeah, it's um and we all have our own biases. It's right. I I get into I don't want to say arguments, but discussions with people about how I would much rather have a well-done self-driving car than drive myself because I think it'll be far safer because it's not easily, it doesn't have the distractions, it doesn't have my biases, it doesn't, you know, all these other things that I think are more beneficial, yet so many people are hesitant to let go of that.

Jon Miller: 35:12

Yeah. Well, I mean, I think I like the self-driving example a lot, right? And I and I think the issue, you know, there is right that like most self-driving cars don't have the processing, the the inputs, sensory input that we humans have, or the processing capability that we humans have. So it it it actually isn't quite ready to replicate doing it as well as we can. Whereas maybe not yet, but but no, not yet. But I think I but agree with your point. Yeah, some like I'd I would rather have an always-on machine that was constantly you know paying attention than than me looking at my phone alert that I just got, you know. Yeah, so that yes. You know, we're not quite there for self-driving cars. I think we are there for personalized buying journeys.

Michael Hartmann: 35:57

So I want to get dig in a little bit more. You also mentioned this idea of um orchestration versus campaigns. Yeah, right. And it it so it almost is like the AI becomes like a conductor, right? Choosing parts of the symphony to play and all that. What um how are you I mean you mentioned playlists, we talked about that. Like what is that? So from uh maybe a practical standpoint for the listeners here, people are watching, yeah. How do they how should they be thinking about today? I do this very like step-by-step campaign. Yeah, how am I gonna how should I be thinking about transitioning to that orchestration model?

Jon Miller: 36:37

Yeah, I think I think today people spend a lot of time building different tracks, you know, building campaigns that have the right exclusion audiences to make sure that they're managing their communication frequencies. Um, I think we should move instead to thinking about building with the Spotify analogy, building the library of possibilities. Right? What are all the approved things that we could do? And then tagging those things appropriately. Here are the things that we should do for directors and above. Here are the things that we might do for people in the buying stage, here are the things that we cannot do for people that are open to opportunities. It comes back to Mike's question on governance and context, right? Like, like build this library possibility, tell up, tell the AI what is okay to do when, you know, and and then so when you think about it, the the mops person's job shifts from visiodiagrams to strategy and guardrails. Um, you know, if you're not doing if-then logic for every scenario, you're you're doing things like setting business goals by segment, frequency caps, budget constraints, compliance rules, you know, data, and then letting the machine do a much better job than you can of handling all that combinatorial math.

Michael Hartmann: 38:04

Yeah, I like that. So I go ahead, Mike.

Mike Rizzo: 38:09

Sorry. I was just gonna say I think that one of the areas that I feel most excited about is the the I I often use the word like stewardship um between the business and the the technology and sort of the capabilities, right? Or the possible stuff. The translation of what does the business want to do and who do they want to sell to, and where do they want to go to here's the tool sets that I could potentially use to deliver on that desire. Uh, but then continuing on that stewardship of I now need to coach my AI model on all of the things that we're trying to accomplish. Who did we build this product for? Why? Who is it intended for? At what point is this product probably most useful? In fact, let me feed you some of the case studies of some of the early sort of developers that we use that, you know, the customer developers that we worked with that helped us build the product. Now that we've been around a while, let me show you what they've done. Let me tell you about the things that they've had success with and what their initial challenges were. Right. And so this is like what a what a really good salesperson or a customer success person has historically internalized and they have context for because they're talking to you, John, who has now come to them as a pre- or post-sale customer or prospect, and you're saying, Hey, here's my problem. Are you guys able to help me or do I need to look elsewhere? And Naomi and I were kind of just like riffing on the side just a second ago with like more often than not, we don't want to talk to people until we kind of have a problem. And I do think that with AI behind us and we set the systems up, that stewardship opportunity for marketing operations is to act as that translation layer to like, great, I understand this, that, and the other thing about our business. I understand how to apply it to the capabilities of what we're trying to do with the tech stack. But I also now have to teach the AI all of those things. And I need to partner with you, product marketing expert person and salesperson who's had these conversations to make sure I get all the right inputs to continue feeding that system so that hopefully the buyer starts to learn on the other side, because like really it's the buyer behavior that has to change, that they can now come to not just the Chat GPT, uh the cloud, but the website of the products that are out in the market and start having a conversation with those businesses. And yeah, maybe it's an AI. But if it knows more contextually why it exists, why the company exists and who they're good for, hopefully the buyer has now learned because of their use of AI and how they engage with AI over the next decade plus, that it's just like, oh, actually, the right thing for me to do is to just start asking questions because they're really intelligent and they know their business. Right. And I just think that that's we're we're a little ways away from that, but the stewardship opportunity, yeah, I think is the most to me, where my brain goes with that, it raises an interesting question, right?

Jon Miller: 41:26

Which is you know, you talked about the agent that's chatting with the buyer on the website when they come and answering their questions. I'm talking about the agent that is orchestrating, for lack of a better word, your outbound actions, right? Am I sending you an email? Am I pushing an ad to you? Am I inviting you to something? Um, but your core point is both those agents need a context layer. And what I think will be really interesting to see in marketing stacks of 2030 is that a single unified context layer that both those things are talking to? Or, you know, is is you know, qualified or whatever qual the future of qualified is, they've got their own, and then you know, whatever the future of Marketo is, they've got their own, and we have context fragmentation. I don't know. I don't know. It's gonna be interesting to watch.

Naomi Liu: 42:25

Um I'm curious if at some point it'll be like AI marketing to AI. We're just like there's a social network now. It's right, we're just not even paying attention to some of this stuff anymore. And it's just like, let me use AI to like filter out my emails. No, I mean what uh prioritize, and it's just AI talking to each other now, and nobody's actually paying attention.

Jon Miller: 42:46

I think that is gonna happen. I think you know, I I've written about and talked about earlier, just briefly earlier on today, about that the buying committee we're used to buying committees and having different members of personas in the buying committee. I think some number of those buying committee members will be AIs and not humans. Right. And Naomi, your point is, and maybe the members of the the AIs will be the earliest members of the buying committee. Right.

Naomi Liu: 43:12

Think about that. And I think it's yeah, and I I've I've had this conversation with, you know, other colleagues and just friends who are in the space too. And it's interesting because like I think it also is what's interesting about AI is like when you use in that context of work, is I guess even personal too, is that it just tends to agree with you, right? Depending on how you ask it. It's like, oh, that's it just it's very agreeable. So I I don't depending on how you ask things and how you ask it to interpret, interpret or give you the information, it just tends to be very complimentary, you know. That's a great idea. You know, it's a good thing you you thought about that, you know. I can see why you would ask that, right? It's just it, and so I I'm just I guess I I I wonder in three, five, ten years, what is this landscape gonna look like with regards to how we all use it? Yeah, you know, both at work and our personal. Um, how do we use it to make our lives easier? But then are we filling that with other things that are also making our lives easier or just more noise?

Jon Miller: 44:20

You know, like these are all fabulous questions to be asking, you know. One the one minor, minor uh thing I'd build on what you said is that obsequiousness or or agreeableness that we associate with AI. We gotta remember that's that's a side effect of the actual chatbot application that is built on the underlying LLM. I don't think it's necessarily a universal feature of all LLMs. Um and and that you know, that's okay. But as as we see AI in more and more contexts and more and more places, you know, it we might it things might it we might experience different different uses in different or different personalities in different applications, I guess is where I'm getting at. Um but you know, but if I you know the job of an entrepreneur, which is my job, is to have a vision of the future and build to it. And man, it is hard right now to Nyambuy. Like this world is changing so fast, and I don't know if any of us really know what this is gonna look like in 2030.

Michael Hartmann: 45:30

Uh I mean Naomi was saying three to five years. I was like, I'm thinking like one to two years. Like I talked to somebody just a few minutes like earlier today who's done something without any experience in coding in three weeks, something that probably would have had to go out by off the shelf, configure it, spend a lot of money. And this was it, like spending a couple hundred dollars a month in some time.

Jon Miller: 45:53

Yeah, I'm not one of the folks who buy into the oh, it's the end of software. I don't think anybody's vibe coding their own Salesforce for a large enterprise anytime soon. I think you will see fewer ancillary tools all around because that's easier to vibe code. But I think the core platforms aren't going away. Um for for you know, kind of you know, for for that's what for what what's worth there. Um but yeah, you know, the what I do know here, so see, I yeah, we don't know, I don't know what 2030 looks like. What I I do think I know what 2026 and 2027 looks like, right? Which is most bigger companies, especially like maybe let's not startups, but like companies of any decent size, right? They're not gonna like completely transform everything about their go to market to this AI native world overnight. Scott Brinker's Martex law says that you know, or technology changes logarithmic or exponentially, but organizations change logarithmically. Yeah. So so I think 2026, 27. What people are going to be looking for is a better, smarter, easier Marketto. Um, you know, that does a lot of things we talked about today, but but fundamentally fits into that same mental bucket of an application. And then I think by the time you get to 2030 or beyond, it might look very different, where we have this context layer as a service, right, that we just talked about. And that might complement a data layer, like Databricks or something else that sort of finally catches on in B2B and a decision, uh, you know, a broader decision layer. Like it might look very, very different. So the the best advice I have for anybody today is first off, do not sign anything more than a one-year contract. Like it, this stuff is just moving and changing too fast to lock your organization in to anything, including my own product, for two or three years. Like, like that, I think does a real disservice to your organization just to save 20% or something. Um second, make sure, like, you know, if you are choosing something that you're you're you're figuring out, you know, how does this not just help me in 2026 and 27? How is this helping me get the flexibility I need to be able to adapt as the world changes towards 2030 and beyond? You know, that's I mean, again, that that's advice I'd give anybody, regardless of what I'm building, um, just because like because it's changing so fast right now.

Michael Hartmann: 48:38

Feels like that would be the same advice I would have given a few years ago with a little longer time horizon.

Jon Miller: 48:43

Yeah, maybe it's still faster.

Michael Hartmann: 48:44

I mean, it's things are moving so much fast right now. That's the feels like the big difference. Yeah.

Mike Rizzo: 48:50

Maybe I um go ahead. Sorry, Mike, like, Michael. Uh I just wanted to sort of pose this question, I guess, uh, to you, John, and and obviously to the to the crew here. But do you think um I'll I'll be honest, like, I don't I don't think a lot of the data inputs from the market uh in terms of like when you were talking a moment ago, you know, you were addressing, like, hey, here's some of the things that you were talking about, the agent that's interfacing with the person on the website, etc. And here's kind of what I'm talking about, the sort of you know, inbound, outbound, nurturing, all all of those, all of those things. Um do you think that the data inputs fundamentally are like really going to move the needle that significantly from the outside from your own your own data, your your sort of first party stuff, that's different. Once they're in your little you know, walled garden, so to speak, uh that that's probably a little different, right? Orchestrating and and enabling something really special based on things that are unique to maybe your product usage or or whatever, sure. Those are great data inputs. But today, to do something halfway intelligent, you know, the world of signals that that we used to sort of buy data from and we still do today. Um, you know, I think they they come down to a few pretty standardized plays, right? It's like someone changed jobs, uh, the company is hiring a lot or lay people off. Um, a particular role was hired, therefore you kind of know like they might have some initiatives underway. Um, but I think like the the stuff you were talking about, the the real special sauce of like what makes something truly personalized was the stuff you talked about just a little earlier in the show, which is like, oh, you actually know that I like to respond to emails on Sunday. You like longer emails, not shorter emails. There isn't a system anywhere that I'm aware of that I might know about you because I've had you in my system for a little while and I've just engaged with you, and now my first party data has told me that about you, John. But broadly, the market, if you were net new to me as a prospect, I wouldn't know that about you. And so the question is like, is there gonna be a point, do you think, where the market, people like you and me, who say, you know what? Have at it. I want you to understand everything about me. Because I would rather you talk to me the way I want you to talk to me than fill my inbox with something that I really don't care about right now.

Jon Miller: 51:40

All right. Well, there's a lot of things wrapped up in kind of your question, like to your very last specific question. I mean, people have been talking about this for a while, right? Like where, like, you know, you know, I might expose my data in exchange for. I mean, you're saying just to get better marketing. Maybe I expose my data. Like I let companies subscribe to my data, you know, for a small fee so they can market to me better. I mean, yeah, right. There are I feel like we talked about that at one point. There are scenarios. I don't know. And I'll be honest, like, I'm a little skeptical, right? Because because we have been talking about that stuff for 15 years and it's never taken off. And I'm not sure that AI or anything else changes that per se.

Mike Rizzo: 52:18

Well, the blockchain was invented for a reason, right? And nobody adopted like like fundamentally the blockchain technically the the thing that was meant to help do.

Jon Miller: 52:26

That's a great, great point. And like, and and it just didn't go anywhere. I I think you know, a couple other things I want to say based on what you were talking about. Like, first off, like in my full Marketo activity history, the information that I like to open email Sunday night is probably in there, and yet that is not captured in any way in Marketto. Um, and and that is one of the scenarios where I think AI can uniquely help us in a way that has been lost in the past. The other thing is you talked about signals commoditizing. Um Brendan Short had a great article about that, where he referred to what investors call alpha. Right. So, so in you know, when when when everybody has the same access to the same public investing information, right, every fund returns about the same, right? And there's lots of research that kind of you know shows this. And alpha is this whole idea of the extra lift you get from your proprietary research and information, you know, and I think we're seeing the same thing with signals, right? Job change and all these relatively easy, publicly accessible signals, you know, are probably table stakes. You know, and that's what you're gonna need to have access to just to perform as well as the market. You know, and it's gonna be your ability to create additional insight by mining your own first party data, by combining unique signals or just working a little harder than the next guy to come up with some other unique signal. That's where the go-to-market alpha will kick in. Um the last little subtle point is you mentioned, okay, I've never seen this person before. How do I know they like emails on Sunday morning? Right? And I don't. But when you have a rich enough representation of the I feel like you're just saying that so you don't get emails on Sunday morning. But if I have a rich enough representation of a rich enough set of people, you know, you can an AI can look at a new person and say, All right, I don't know if Naomi likes emails on Sunday or not, but I I I I know what what people like Naomi like, so I can at least make some educated guesses. Yeah, you know, and again, none of this could be done with legacy machine learning, and it is exactly why I'm so excited today for what AI can unlock. Yeah.

Michael Hartmann: 55:01

Interesting.

Jon Miller: 55:01

Yeah, love it.

Michael Hartmann: 55:03

John, I know I know we're close to the end of the time. Are you open for one more question?

Jon Miller: 55:08

Hit me with it.

Michael Hartmann: 55:09

All right. So there's one that I we hadn't talked about. It's unrelated to I think what we've talked to about already, but I think your unique perspective will be useful because you've been both a CMO and a founder and run companies. I think one I heard a statistic years ago, I don't know if it's still true, that something like less than 10%, I think it's closer to 5%, maybe even lower, of like board members have any exposure to actually running marketing or what it's like to do it. And so that's at least the theory was that partly why there's challenges between the rest of the C-suite and marketing leadership. I guess what's your take on that? Do you think that's still a real challenge? And then how does all the stuff we just talked about gonna affect that?

Jon Miller: 55:53

All right, well, you yeah, I thought you were gonna have to ask a quick one to the end here. Okay, well, I so I I let me let me try to adjust that at least um kind of at a high level. Yeah, I I think that the back to the early part of the podcast, you know, when we were applying this broken, simplistic machine model to marketing, yeah. I think we we trained CEOs and CFOs to think of marketing like a gumball machine, where they could just put some budget in and get some pipe you know, pipeline out. Um, and that had a it was meant to empower marketing, it was meant to let the CMO be able to say, well, do you need this much pipeline, give me this much budget. And what happened in reality is it, I think, diminished marketing to be really an owner just of one of the four P's promotion, um, and to not and to lessen their impact and influence over product or distribution, which is place, you know, or pricing, you know, for example. And you know, I think it's a real problem. When you basically push the um the decisioning around um, you know, or or ask marketing to drive revenue, you know, when they don't own all those things, you're really just asking them to kind of you know decorate stuff that they don't have control over, you know, effectively. And that and no marketer is going to be successful with that. So all the stuff we've talked about today is like how can AI unlock things that are better or different? Um so, first off, like this fundamental CEO challenge is deeper than AI, right? It is it is it is how executives think of the role of marketing, and that will take effort and time to change, regardless of AI. I think specific around AI, you know, if marketing can become more effective at crafting you know good journeys, you know, then that won't hurt the the the role of marketer. I think more deeply, if marketing operations can be the owner of the context and governance for how do we treat our customers and how do we talk to our prospects, you know, that becomes a fairly strategic position um to kind of help influence the organization. So nothing we talked about today is gonna hurt marketing's credibility, you know. Um and it might it might help, but but it's a deeper, longer journey than just that.

Michael Hartmann: 58:53

It makes sense totally. Well, I'm sorry, I didn't I didn't I didn't uh fairly set the expectation for my my open-ended questionnaire. Uh hey, appreciate it and and indulging us for that with that, John. It's been great. If if folks want to keep up with what you're doing in your stealth mode when it's ready to be launched and all that, what's the best way for them to do that?

Jon Miller: 59:13

Follow me on LinkedIn. You know, that's uh you'll be the first to know there, and also you'll get plenty of hints and trip tips and secrets as I leak them out over time. So yeah, John Miller too on LinkedIn.

Michael Hartmann: 59:24

Fantastic. John, thank you. Mike, Naomi, thank you as always. And to everyone out there watching or listening, we appreciate you. If you have ideas for topics or guests or want to be a guest, you can reach out to Naomi, Mike, or me. We'd be happy to get the ball rolling. Till next time. Bye, everybody.

Mike Rizzo: 59:39

Thanks. All right. Thank you.