Pod Insights MarTech Podcast ™ // Marketing + Technology = Business Growth The Scope of AI Disruption About this episode AI disruption is rewriting marketing's fundamental rules. Isaac Ferreira, VP of AI Growth Systems at Shift Paradigm, explains how intelligence is being pulled out of traditional SaaS platforms into contextual engines that sit below your entire tech stack. He breaks down the "curation effect" where AI agents now filter all digital marketing channels, forcing marketers to rethink discovery through pattern-matched content placement and non-intermediated channels like direct mail and billboards. See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.
Transcript 0:00 Speaker 00 The Martech Podcast is a proud member of the I Hear Everything Podcast Network. Looking to launch or scale your podcast? I Hear Everything delivers podcast production, growth and monetization solutions that transform your words into profit. Ready to give your brand a voice? Then visit iHereEverything.com.
0:25 Speaker 01 From advertising to software as a service to data, across all of our programs and clients, we've seen a 55% to 65% open rate.
0:36 Speaker 02 Getting brands authentically integrated into content performs better than TV advertising.
0:42 Speaker 03 Typical life span of an article is about 24 to 36 hours.
0:46 Speaker 01 If we're reaching out to the right person with the right message and a clear call to action, then it's just a matter of timing.
0:53 Speaker 00 Welcome to the Martech Podcast, a member of the I Hear Everything Podcast Network. In this podcast, you'll hear the stories of world-class marketers that use technology to drive business results and achieve career success. Here's the host of the Martech Podcast, Benjamin Shapiro.
1:16 Speaker 04 900 million, that's how many people are using ChatGPT alone every week. Something has fundamentally changed. SEO traffic is dropping. Email isn't an own channel. The click is dying and your customers are doing product research without you. And the martech stack you've spent years building, well, it might be solving problems that don't exist. AI isn't a tool, it's a new operating system that sits on top of everything. So how do you build a marketing roadmap when the rules of engagement are being rewritten in real time? I'm Benjamin Shapiro and today we're talking about the scope of AI disruption. And joining me is Isaac Ferreira, the VP of AI Growth Systems at Shift Paradigm, a growth partner helping companies rethink marketing strategy, revenue architecture and activation
2:01 Speaker 04 in an AI-first world. Shift Paradigm is also a sponsor of the Martech Podcast. And today Isaac is going to break down what marketers need to understand before their entire stack becomes obsolete. Isaac, welcome to the Martech Podcast.
2:15 Speaker 01 Thank you, Ben. It's really great to be here with you. I'm looking forward to the conversation.
2:18 Speaker 04 Excited to have you here. First off, thank you for being a sponsor of the Martech Podcast, you and the Shift Paradigm team. I feel like your company name is an appropriate one. Shift Paradigm is exactly what we're going through.
2:33 Speaker 01 It is.
2:35 Speaker 04 AI disruption isn't just one thing. It's happening on multiple layers. So what are the layers that marketers need to understand when you think about AI disruption?
2:46 Speaker 01 You know, a lot of marketers, when I speak to them, think about AI disruption in terms of technology, and that's one layer. The technology is absolutely changing, but the technology is creating a change in the way that we actually market. The way that our customers are interacting with our marketing and the way that our marketing is being curated, which we'll talk about later, I think, is interrupting historical flows, the way we used to market. And the last way is the way that organizations work. The way that we design websites, the way that we actually build journeys, the way that we operationalize is all changing as a result of the technology. So it's sort of the way we work, it's the tech itself, and it's the way we market.
3:24 Speaker 04 Yeah, you know, I think that's a great segmentation of the three problems. It's not just a tool. It's not just a change in the buyer behavior. It's really a new way of thinking. AI implementations have sort of chronically failed, right? There's the stat that's gone around, I think it was a McKinsey study of 95% of AI implementations fail, and I honestly think that stat's like two years old now. Why are the experiments that we're using in AI failing?
3:57 Speaker 01 You know, it's interesting. I know that that study has been kind of universally panned. If you look on LinkedIn, you're going to find people making fun of it kind of across the board. I think what was right about that study is the fact that we were focusing on the wrong thing. If we look at where AI's strong suit is and where we can find a lot of leverage, it's not necessarily in generative content, although there is space for generative content. But allowing AI to manage communications and relationships with our customers is probably not the right place to start. So Rashad Tabakawala is on our board and I love the way that he frames this up. He says, you don't need an AI strategy. You need a business strategy that uses AI to accelerate the way you're doing business.
4:39 Speaker 01 The way we approached AI implementations over the last two years was, oh my God, there's a new technology. I'm going to fall behind. Let's implement it. Instead of the classical seasoned business approach, which is I have a gap in my business. I need to solve X. Here's the strategy I wanted to solve it with. And here's the tactics and technology that I want to apply, you know, to, to build that strategy.
5:02 Speaker 01 We've kind of went this backwards. We said, let's deploy tech instead of let's solve a problem. And I think what people are waking up today to realize is that the problem still has to remain central. They just have to be able to connect where technology supports improvement.
5:17 Speaker 04 The thing that you said that really resonates with me and I'll paraphrase what you said is essentially when we started using AI. We started off at the lowest hanging fruit. Let me use this thing to write my copy for me. Right. And then it was, well, we already have chatbots. So let's just integrate some chat and I don't have to talk to the customers. AI can talk about the customers and I know that my use of artificial intelligence has evolved greatly. I'm sure yours has too and your customers. But the mindset of what it does has changed. And instead of it being my copywriter and it being my customer service technology.
5:57 Speaker 04 Now I think we're getting to this age where it is my resources, research assistant. It is documenting and archiving and, and, and we're actually using it for, well, intelligence. And to me, that is the fundamental shift of it's not a implementation. It's giving us the brain power to do the things the way we want to. And then obviously helping us write code and Claude code and everything we see on LinkedIn.
6:25 Speaker 01 I, you know, I absolutely agree with that. What it's helping us create is contextual intelligence. As marketers in the past, our intelligence was largely logic based if then statements. Or temporal, time based. If two weeks have passed, send a message. Right. So we either get a signal that says if a person viewed X content, send them a message. Or if two weeks passed, send them a message. What we have the opportunity to do now is to trigger communications, personalize communications based on current state at an individual level. It's like the level of one-to-one personalization that we've talked about the last 10 years, but never been able to achieve because AI can now establish context across all the attributes you're collecting on a given customer and allow you to personalize that marketing. It's true personalization.
7:07 Speaker 04 I want to talk about the martech stack and, and how that is changing because you've mentioned, well, we had a lot of if then, and, and I, I think that if those rules sort of added up, well, if I have enough if then statements, I will have real relevant personalization. That was the mindset, let's say, three years ago, pre-GPT.
7:29 Speaker 04 Now the world is different. So how should we think about the martech stack? And, and what is possible and what we should be building, knowing that we're on these sort of, you know, now kind of archaic systems. Talk to me about the impact of the martech stack.
7:44 Speaker 01 Well, that's a big question. I think fundamentally what we're seeing happen in the martech stack is that intelligence is being pulled out of the SaaS. It doesn't have to, so in my opinion, and a lot of the
7:54 Speaker 04 I gotta stop you. We're pulling intelligence out of our SaaS. Just sounds kind of funny.
7:58 Speaker 01 It does sound funny, but intelligence in SaaS before was the if then logic. It was the time-based logic. It was the if this do that. That doesn't have to be contained within SaaS anymore. You can take it below the stack. So think about a problem like churn. Churn is a classic problem that crosses many verticals inside of your organization, many operational verticals, many pieces of technology, lots of data sets. You get customer support signals from your CRM. You get product signals from your actual product, if you have a product. You get people searching your website for help topics. You get the deal formation at the beginning. All of these signals lead to a holistic pattern that may or may not indicate churn.
8:40 Speaker 01 Right now, you don't really have a single platform that can solve that problem for you. It can help you identify those patterns and then activate against those patterns. Putting AI below the stack allows you to collect all those signals, generate a representation or a pattern that represents what churn may be, and it allows you to push activations back into those platforms to actually reduce churn or to signal a human inside of a CRM to go and call a customer to help reduce that churn. What I mean when I say that AI is removing intelligence for SaaS is that SaaS no longer contains or no longer has a monopoly on where that intelligence lies.
9:17 Speaker 04 Two things that you said that are still in my head. One, just I want to give you a tagline for your billboard. Shift paradigm. Pull your head out of your SaaS.
9:28 Speaker 01 That's amazing. I love, okay, you know what? We might have to pay us some royalties on that one.
9:32 Speaker 04 There you go. The second one is actually something you said that I think was probably assumed to be common knowledge. You put your AI under your SaaS. And when we think about the changes in the martech stack, you know, I think that there's the traditional CDP CRM, right? Communication layer and all these things are hopefully integrated and they're relatively horizontal.
10:01 Speaker 01 Verticalized, yeah.
10:03 Speaker 04 Yeah. Thank you.
10:05 Speaker 01 You're welcome.
10:06 Speaker 04 How do we think of where AI is? Like when you're visualizing your actual stack? I go back into my like Scott Brinkerism of like, write your stack down and present it to everybody and everybody has this different shape of SaaS.
10:20 Speaker 01 Yep.
10:22 Speaker 04 Give me some metaphors or some visualizations here of how you think of where AI lives in that shape of your SaaS when you're trying to document what your stack is.
10:31 Speaker 01 Sure, absolutely. So I love Scott and I love the way he creates these visualizations. I will say Scott, from his background, and just based on my experience, is very SaaS centric. That's expected, having come from HubSpot and his background. My thought is a little different. AI exists in two separate places. AI is inside of your current tools. So inside of your CRM, inside of your map, inside of whatever those tools are, I can do automations within a CRM. I can create emails within my marketing automation platform. The other layer of AI can coordinate across and activate into all those technologies. That sits below it. So in my mind, you have a contextual engine at the very bottom of your stack. That's the AI contextual engine. On top of that, you have data, all the data
11:15 Speaker 01 that comes from your organization. And through that, you have connections to all of your all of your current SaaS products. And in my mind, that activation is signal from the SaaS products, contextual decision from the AI engine, and then activation back up to the SaaS products. I imagine it kind of as a half moon with SaaS contextual engine at the bottom.
11:35 Speaker 04 How are the SaaS tools that we're using now that we're in this AI centric martech stack changing? You know, are you still using the sort of traditional CDP? You know, everybody's been talking about Compostable for a year. I'm not even sure if I say the word right.
11:54 Speaker 01 Composable, yeah.
11:55 Speaker 04 Composable, I always say Compostable.
11:57 Speaker 01 Yeah.
11:57 Speaker 04 Like, you're gonna go put it in the compost?
11:59 Speaker 01 I do need that some dirt.
12:00 Speaker 04 Thank you. And maybe that's what's actually happening is we're just building our own databases and we don't need CDPs. But, you know, then there's the implementation layer. Are the tools changing or is AI actually replacing some of these tools? Everybody talks about the death of SaaS. What happens to SaaS?
12:17 Speaker 01 SaaS apocalypse?
12:19 Speaker 04 That's a good one.
12:20 Speaker 01 That's a layered, a very layered question. It's complex. SaaS is not dying, in my opinion. It is going to be going through a pretty extreme transformation. If we talk about the way SaaS is currently structured, it contains four primary parts. It contains the UI, so we can access the data. It contains our data. It holds our data in large part. It contains the intelligence that we talked about before. And it contains the activation layer. All those things are in SaaS today. What's changing and what I'm seeing is that some of those layers no longer must be contained in SaaS. That is the model break that I'm seeing. I don't have to contain my intelligence inside of SaaS. I need an activation layer. I don't have to contain my data inside of SaaS.
13:05 Speaker 01 I need an activation layer. So what I see is less of a reliance on packaged offerings. And you can see this happening across SaaS, just in the market share pricing. And, you know, everyone talking about the changing and per-seat pricing models and so on. What people are realizing is they don't have to be bound into a platform, a monolithic platform that does everything. They can take part of that for control into their organization. And, you know, I think as we move forward, and I'm probably getting a little ahead of the game here. If you're going to be successful, you have to have control of your data. And the SaaS model. Innately does not support that.
13:40 Speaker 04 Yeah, it's interesting. I'll talk a little bit about my experience and some of the SaaS that we've used for our production infrastructure.
13:51 Speaker 04 I have canceled a ton of point solutions. Hey, I had this tool that, you know, does our transcripts or processes our video a certain way, or, you know, whatever the individual point solution was. And on the flip side, saving that $49 a month and this $149 per seat, what I've ended up doing is saying, oh, well, now I have access to, you know, use all these APIs. And since I'm not the one that's actually doing the research, I'm aggregating and have AI do it. I'm pulling in more data from different sources. And so I don't know if it's data SaaS, or DAS, or data APIs, or whatever you want to call it. But like, my plumbing has gotten so much more complex because I have so much more
14:36 Speaker 04 data coming into my system because it's easy to set that up. And so the research is better. And, you know, the data that we're analyzing is so much more rich. But then I'm using my own systems and my own database. And I kind of just have like one or two mechanisms to get stuff out there. And so I don't know if those are called SaaS, or if they're called data products, or what we call them. But like, I've got a million APIs that are giving me information that I'm, you know, pulling or sending to. But I'm just not using the front end visual SaaS tool with the exceptions of my CRM, you know, Monday, my project management board. Am I the only one that's going through this? Are we seeing a sort of a rise in data consumption?
15:17 Speaker 01 Oh, there's absolutely a rise in data consumption. I think that's a real challenge and a real opportunity. That's also one of the reasons you're seeing a change in SaaS. You think about all these middle layer SaaS tools that exist today. Those are all being replaced with agentic solutions. Think OpenClaw, or ZeroClaw, or whatever Claw. These agents, you can go and build these integrations. What you just described is building an intelligence layer yourself, right? You're building that intelligence yourself. And then you're being able to actualize that data and activate that data yourself. You just described why SaaS is having a problem.
15:48 Speaker 04 Yeah, the ironic thing is we see the news on LinkedIn of like HubSpot stock is tanking and Monday.com is down 75%. And to me, I think those are the safest bets in terms of long-term viability is these, you call them monolithic platforms, but I think the place where you sort of centralize everything, whether it's in a CRM or in a project management tool, like that sort of central repository is still useful to me. It's the point solutions that I think are in trouble. But on the flip side, the stocks of all the sort of monolithic platforms have tanked. What's going on there?
16:28 Speaker 01 So, okay, so I agree with you. Today, point solutions, and I'd say tier two and three technology solutions are the ones that are most at risk of being replaced by AI, meaning gateway solutions that aren't a primary platform. I think Scott Brinker calls them systems of record.
16:45 Speaker 01 But my contention is that will change as well. And I think, I think that's being priced in the market.
16:49 Speaker 04 The argument is no one's going to vibe code HubSpot overnight.
16:52 Speaker 01 That's not true.
16:54 Speaker 04 And, and you're basically saying, no, yeah, eventually they will.
16:56 Speaker 01 Eventually they will. And in fact, okay, so it's funny you mentioned HubSpot. So a couple of weeks ago, I saw a conversation and the co-founder of HubSpot, I forgot your name, I apologize, came on and said essentially, any company from 2 to, you know, 10,000 needs a CRM. And I thought, yeah, but do they need a fully featured CRM that costs X dollars a month? So I went and built a CRM.
17:18 Speaker 04 Pipe drive is $12 a month. I've been using it for a decade.
17:21 Speaker 01 I went, yeah, I went to Claude Code and built a CRM and said, can I code a system that any organization with two to 10 people could use? And the answer is unequivocally yes. Absolutely can. So that's already a portion of that CRM TAM that is gone. I don't see that stopping as AI capabilities improve. I don't see that stopping as people start to control more of their data. I don't see that stopping as people start to realize that the model has written. Doesn't fit the technological reality.
17:50 Speaker 04 Okay, I think of the world, you know, being a solopreneur and a creator and, you know, we've got our production business in the lens of smaller companies, but you work with a lot of larger companies, enterprises, and there's a governance problem that I'm relatively blind to. What's happening with the large enterprise when they think about AI adoption and governance? And it seems like larger companies are always slower to build and evolve because of the level of complexity. Are large companies, enterprises in trouble because anybody can start to, you know, vibe code HubSpot in their basement?
18:30 Speaker 04 Or are, you know, there are ways that they're implementing rules and governance that's going to allow them to use their scale?
18:39 Speaker 01 The opportunity for large, you know, enterprise in governance is not that someone's going to vibe code a solution, although I'm sure the infosec department's going to have a problem with that if it hasn't gone through their process. The challenge that I see and I keep hearing about is that these organizations must be able to govern agentic outputs. And large platforms, the big players, do not allow these organizations to come in and look inside of their software, figure out what the outputs are, and then govern them before execution. I mean, I can't connect to the large CRMs, the large marketing automations with my own governance agent to validate what's being produced. I cannot, in essence, I cannot centrally govern that product. So what they're doing is they're building their own solutions.
19:24 Speaker 01 AI on the bottom, like I talked about, AI below the stack, finding the processes they need to solve and using that, which is fully governable. So in essence, what we have is a movement away from large SaaS because it can't be governed from an AI standpoint. And I'm not talking about standard practices within a tool, like creating a journey or things like that. But anytime there's a decision point, IT can't govern it. So how do these, if these large solution providers are going to survive, they're going to have to find some way to integrate enterprise governance into the workflow.
19:58 Speaker 04 You remember the movie Judge Dredd?
19:59 Speaker 01 I do.
20:00 Speaker 04 Okay, so Sly Stallone is like the cop, but he's also the judge and my dad used to give me a hard time. Cause whenever we saw that movie, I was probably watched it way too young. Sly Stallone's big line is, I am the law.
20:13 Speaker 01 Yeah.
20:14 Speaker 04 And that's my Sly Stallone impersonation for today.
20:17 Speaker 01 That's pretty good.
20:18 Speaker 04 The, the point is, thank you. Uh, the point is companies are essentially needing to build agentic governance.
20:25 Speaker 01 Yes.
20:27 Speaker 04 But inherently, there's a problem there.
20:29 Speaker 01 Uh huh.
20:29 Speaker 04 Right? We can't let our employees just rely on their agents to do execution. So we need to govern it. So we need to build AI to govern the other AI. Well, what's the point of the governance if you're just using AI to do the governance?
20:46 Speaker 01 That's, you know, it's a really good point, but I think that's, once you understand the basics of how an LLM operates, you realize why there's a need for it. And that's that they're non deterministic. It's not a math. It's not a math equation. It comes up with the same answer every time. It comes out with a different answer every time, based on the context. And sometimes that context results in a decision that is not favorable to the company, not favorable to a relationship, not favorable to communications. In which case you have a backup that's watching it. That's the governance that they're, they're working on right now.
21:15 Speaker 04 It's the reason why I brought up Judge Dredd is because he's the, um, artificial, like human sort of hybrid. Cop and judge. And it's like, every company needs a judge dread essentially to go govern and police their existing teams. But the whole goal of the governance is to make sure that the AI isn't going rogue. And so they're using AI to make sure the AI isn't going rogue. But what if the governance, like AI is still has the same problem, doesn't it? It's like, still can't that go crazy and off the rails?
21:50 Speaker 01 Oh, you were talking about like an infinite loop of, of risk possibilities and that still exists.
21:54 Speaker 04 The point is we're not using human judgment to evaluate the AI. We're using AI to evaluate the AI and doesn't that?
22:00 Speaker 01 Because human judgment fails at scale.
22:04 Speaker 04 Fair. I mean, civilization's got this far. It is not a complete loss.
22:08 Speaker 01 No, no, that's not what I mean. It wasn't an existential state, but I mean, as you try to scale the number of decisions, human capability limits it just because you don't have that much labor.
22:18 Speaker 04 The body count.
22:18 Speaker 01 Yeah, exactly. It wasn't an existential state.
22:22 Speaker 04 We can go really deep if we need to. All right, let's, let's move off of the governance. I, I get that essentially large enterprise, instead of just focusing on letting the agentic solutions run wild, they need more governance and they're using artificial intelligence for this as well, which sounds like an extra level of complexity and maybe the solution is build governance into the agents as opposed to build a governance agent?
22:46 Speaker 01 You can, but once again, you don't control the roadmap of any of the other big players.
22:50 Speaker 04 Okay. I want to talk a little bit about the curation effect. AI sits between not only your, you know, data and implementation on the marketing side, but it's also disrupting the buyer journey. It's sitting in between brands and customers. So what does that mean for how people are discovering products and how should we think about, you know, once we have this new martech stack, what the new reality is for how we're trying to reach people?
23:18 Speaker 01 Yeah. So the curation effect is a term that we, we coined to help people understand that you're no longer talking specifically to Ben. You're no longer talking specifically to Isaac directly. In the past, I could send an email, had a reasonable expectation of delivery into your inbox. And with that delivery, I had a reasonable expectation of attention. You would at some point probably scroll past and see my headline and have a chance to be intrigued by it and open it. What's happening today is completely different. Right now you have software agents, AI agents that are in between us. So when I send a message, that agent, Google did, I think, announced this in January, will classify the message and put it into a different folder. It might summarize the headline. This headline you
23:58 Speaker 01 worked super hard to improve click through, or sorry, open rate, is no longer effective because it's been summarized. Or it might summarize the content, or it might not show it at all. Now you add on to that a layer of, you know, we talked about personalized agents like OpenClaw, ZeroClaw, all the claws. If I want to manage my inbox with an agent, I cannot do that. So the question becomes, how do I get past this new stage that never existed before, which was visibility? How do I create visibility for my marketing in these intermediated channels so that I can actually get the attention and hopefully drive into the next path of the journey? That's a question I think a lot of people are still trying to solve today. And the way that we're, we're thinking about marketing is we have really two, two types of channels.
24:41 Speaker 01 You have intermediated channels or curated channels and you have non-curated channels. A lot of those non-curated channels today are what we use, consider old hat. They're billboards, they're direct mail, they're your website, they're advertisements. Even advertisements to a point are intermediated, but the ones that are curated now are email. They're search, they're display ads. So everything before that we could pay and get a reasonable return on now has another layer called visibility.
25:10 Speaker 04 You, you forgot something that I happen to be biased on, which is media and content. I understand it's like what used to be the easiest path to get to your consumers. Send them an email, buy ads, put them in front of them. Right now there's enough technology that essentially can block that out if, if the consumer wants to. And honestly, nobody's going to the websites. They're going to L L M and they're doing their own research, right? They're not involving you in the buyer journey until their decision is made.
25:41 Speaker 01 Essentially, yeah.
25:43 Speaker 04 You can advertise them by putting products and, and marketing in the real world. We're going I R L again. There's also the amount of media that they're consuming. Shift Paradigm is a sponsor of this podcast because we have an audience and, and you have a brand that you're trying to educate on some of the things. The smart people at Shift Paradigm and how they can help. Uh, how does media consumption fit into your paradigm shift?
26:07 Speaker 01 So media will be intermediated, just like email will be intermediated, just like searches being intermediated. There's really no way around it. There is.
26:15 Speaker 04 Why?
26:16 Speaker 01 Well, they're all digital delivery channels and anything digitally delivered can be digitally filtered. And will be digitally filtered is what I should say. So this has already been happening to us, whether we realize it or not. If you have a YouTube account and you use it regularly, you're showing the ads that they think about the, well, the ads as well as the videos I think you're going to be interested in. They're not showing you everything that was published that day.
26:37 Speaker 04 Talk to me about discovery architecture, right? If we're disintermediating most of the ways that we're trying to get in front of our consumers, right? If there's this layer that's hindering your discovery, it's obviously not going to, people want to have, discover products and services and they're looking for vendors. What's the right way to approach and what's the infrastructure we need to make sure that we're putting at least our best foot forward?
27:03 Speaker 01 Well, I think discovery architecture is changing. We talked about the fact that it's intermediated, but I think a lot of what we pay for today will shift to what we pay for tomorrow. So some of the discovery opportunities are going to be in paid marketing within LLMs and paid within their responses. Some of that discovery architecture is ensuring that we show up in a natural LLM search by putting pattern matched references inside of places that LLM aggregates. Think Reddit or Quora or other places where you aggregate public opinion. If you can get into those answers, answer a question, have a lot of upvotes to it,
27:45 Speaker 01 and have that same sentence maybe be on your website, now that pattern gets matched back to Shift, now Shift gets mentioned in an LLM response. The more of that pattern matching you can create, the more opportunity you have to be discovered. But then there's also this discovery, which we talked about, which is going back to what we consider owned, and it's not even really owned, you're paying for it, but going back to direct fail, going back to billboards, going back to non-digital display that really cannot be filtered. It's either pay in a market that can be filtered to not be filtered, or it's to advertise on channels that are unfilterable.
28:22 Speaker 04 To me, it sounds a little bit less like you're saying paid versus organic, which was a lot of the mechanism that we thought about evaluating owned or been paid over the last decade, and now it seems like it's actually more physical versus digital.
28:37 Speaker 01 Well, it's intermediated versus non-intermediated.
28:39 Speaker 04 Yeah, and I'm bifurcating physical versus digital. You can put it in the real world and try to put it in front of somebody's face, and there's no way to digitally filter that unless there's a robot that's going through your mail. Not yet, but maybe soon.
28:55 Speaker 01 There's also one other path I didn't mention, I'm very sorry.
28:58 Speaker 04 Go for it.
28:59 Speaker 01 So the other path is get to know your customers much better than you do today. So to get past an AI firewall or an AI intermediation or curation, which what we call it, you have to be relevant at that time for something that person is looking for and have the brand to support it. If you cannot establish relevance, you're probably not getting through those filters. But if you get very good at contextual relevance, you're going to have an opportunity to use some of those channels and actually get through the filters.
29:27 Speaker 04 See, I think it's contextual relevance. There's also a question of authority. And that's actually where I think the media portion is important is that when you can build, and I'm biased. I create media professionally. When you create media and you you build yourself up as an authority on a subject, that helps you cut through some of the curation.
29:47 Speaker 01 That's the brand portion of this. Absolutely agree. It does like that through. And I think that's part of, um, that's part of relevance is brand authority. It's, uh, does your product fit the need? It's, are they looking for it at this time? It's all, all those attributes to establish context that get you through the filter. And we, we don't have by and large the systems and processes to do that today at a true one to one level. Although I'd say we've been striving for it for a decade, but we've been personalizing at the attribute level, personalizing at the interest level. And a lot of our marketing journeys have followed that. If you, if you notice right now, like we, we have journeys that are triggered based on like forks in the road. If you do this, go this way. If you do this, go this way. But people don't experience the world that way.
30:32 Speaker 01 People have random thoughts and emotions. They have, uh, desires to go and buy something. They have interests in something here. If we don't build systems that can adapt dynamically to the way people actually operate, we're never going to get through those filters. And that's why we're recommending to our clients that we start, you know, looking at instead of just personalizing, how do we contextualize?
30:54 Speaker 04 Let me try to get close to landing the plane here. Uh, we'll, we'll put the, the, uh, the wheels down and the flaps up, I think it is. We talked about how the technology stack is changing and, and there's an intelligence layer. And so how we operate is different. The buyer journey is changing and so the way that we market is changing. And not only is the, the martech stack because of the use of technology adjusting,
31:25 Speaker 04 everything we're doing, how we're approaching our customers, what we're saying, the mediums we're going through, it's all changing at the same time. So to wrap this up, what is the way that you think that marketers should approach their roadmap for the next, you know, let's call it short to medium term, six to 12 months. Everything is changing. The tech stack is changing, the buyer journey is changing, the channels are changing. How do you figure out what to prioritize?
31:51 Speaker 01 So, you know, I go back to what I've told my clients for the last six years. And that's your tech strategy needs to be based on your business strategy. And we need to start really connecting brand, you need to start really connecting product, journey, customer experience to the actual business. Once we understand what the business does, we understand how to develop our tech stack. And the other thing we have to realize is that not every solution requires a new piece of technology. That is the mind, mind frame, that's the state of mind from the past. I need a solution, I'm going to go buy this platform, I might have eight other things, but I only need this one. I still have to buy this platform. And then in order to, you know, really prepare yourself, you need to think about the data.
32:34 Speaker 01 If you're going to be able to contextualize this, you're going to be able to use contextual decision management. If you're going to enable across platforms, you have to own the data, which means your, your roadmap has to include a way to pull that into your organization's control or into easy access.
32:49 Speaker 04 I think the biggest thing that you said, and the most important thing to say is that your tech stack needs to be centered on what your business strategy is. The world is changing and if you don't have a clear understanding of what you're trying to accomplish, who you're trying to reach, you can't put together the strategy for how you're going to reach.
33:06 Speaker 01 Absolutely.
33:08 Speaker 04 All right, Isaac, I want to move on to our lightning round where I'm going to ask you a couple of fun questions related to this week's topic. It's about AI friendly marketing roadmaps. Are you ready?
33:17 Speaker 01 Let's go.
33:18 Speaker 04 Yeah. All right. AI is moving incredibly fast. So what is the one thing marketers are over complicating right now?
33:25 Speaker 01 You don't need to solve everything, just solve one small thing.
33:28 Speaker 04 Well, tell me, how do you figure out what that one thing is? Come on, give me some, some meat here.
33:32 Speaker 01 Some meat, okay, all right. If I, if my organization is experiencing something repetitive that I'm doing all the time and it's taking a large portion of labor. Focus on automating that one small thing. You don't need to miss the forest for the trees. You don't need to solve everything with one centralized master AI strategy and implementation. They'll figure out how to make it so I have to send less repetitive emails.
33:56 Speaker 04 AI is moving so fast and before you start thinking about what you could do, the best strategy is to take what you are doing and automate it. Get the stuff that is reoccurring, mundane, easy, it just Claude code it, use whatever tool, whatever agent, whatever claw you want. But get your existing job off of your plate so you can think about the way to continue to iterate off of what you know already works.
34:23 Speaker 01 Yep. Yeah, I'd say having a little bit of focus here goes a long way. And also, just as a rule, don't automate relationship with your customers. Automate data and operations.
34:35 Speaker 04 Turns out you need to keep the relationship. It's a relationship. It's not something that you're going to move off to a computer to manage the relationship for you. All right, let's move on to our next question. If you had to rebuild your entire tech stack using AI agents and APIs, what's the first SaaS platform you'd eliminate?
34:54 Speaker 01 Oh goodness. Uh, some of these bridge technologies, think uh, your social automation platforms. Not really needed. You can automate almost all of that with, with OpenClaw. Think a lot of your integration platforms. You were just talking about integrating pulling data down. All these integration platforms can be automated with an agent out of the box.
35:13 Speaker 04 All right. So you're talking about the implementation layer. I've created my social content. How do I schedule it? How do I get it out there? It's not what I'm writing. It's essentially what gets my content, my marketing to my consumers. You're, you're going after those platform first. Does that make your user experience better, smarter? Or is it just, this is something I can do with AI that's cheaper and faster?
35:37 Speaker 01 I think it's more controllable. Anytime you're able to pull something into your process, you don't have to build around the technological limitations of a third party produced piece of software. You also get the tech debt with it. So where I'm starting is a layer that you, where you can start to operationalize the in-between tasks in the part of your process instead of into. Another technology, a third-party technology. You mentioned pulling data down to inform a lot of your research and doing the integration work with an agent. That same thing can operate in almost any business. You don't have to go buy these intermediate layer data integration tools necessarily. So I'd say we need to start with non-systems of record.
36:20 Speaker 01 A CRM, marketing automation, you're not going to reproduce those right now.
36:23 Speaker 04 Good news for HubSpot, Asana, Monday, Salesforce, whatever product management and CRM you're using. Isaac is saying you're safe. It's all the point solutions built on top.
36:34 Speaker 01 I'm saying you're safe for now, but the point solutions.
36:37 Speaker 04 Get your head down.
36:38 Speaker 01 Yes.
36:39 Speaker 04 All right. Here's a fun one. And stop me if my research is wrong. You went from operating nuclear submarines to leading AI strategy. What's the one lesson from that experience that applies directly to managing AI systems?
36:56 Speaker 01 Phew. I don't have an easy answer for that one.
36:59 Speaker 04 I mean, you went from a technology that could blow up the world to one that we think can blow up the world. There's got to be a parallel somewhere here.
37:10 Speaker 01 You know, the, yeah, I hate to go back to governance, but one thing the military was good, very, very good at was policies, procedures, and checks and balances. And if we don't implement that within our organizations, you know, we might not be blowing up the world, but we could blow up our revenue. We could blow up our customer relationships. We could blow up our internal processes and we could alienate our own teams.
37:35 Speaker 04 We've talked a lot about governance today. You've mentioned it in terms of how nuclear submarines need to be governed, just like AI does. A lot of what the news has been, you know, in the last month has been this fight between the government about, you know, Claude being an enemy of the state. Or I forgot what exactly the classification was. Where do you land on how AI should be governed?
38:01 Speaker 01 I don't have an answer to that. I, I think that's a question that we as humanity have to answer. And we, I, I don't see us pulling together at the moment to do that. You know, have you ever seen or read Isaac Asimov?
38:16 Speaker 04 No.
38:17 Speaker 01 Never?
38:18 Speaker 04 No.
38:19 Speaker 01 I think we need to get reacquainted with the classics. So Isaac Asimov had, um, three basic rules. I believe it was three basic rules for robotics. And it was, you know, a robot can do no harm to a human. A robot, if it sees a human in danger, must save it. And a robot, uh, may not harm itself as long as the first two rules aren't validated. Something like that, right? We need to find some way to build a framework that protects humanity and preserves, um, these artificial intelligences that we're, we're creating. And I don't think we've gotten there yet.
38:50 Speaker 04 Not even close. And I think that the big battle is who's going to decide that. And mostly here in the United States, is it the federal government? Uh, 50%, 50% of us.
39:02 Speaker 01 I don't know that they're going to solve it.
39:03 Speaker 04 Look, 50% or forget the political commentary. 50% of the country is going to argue it's wrong because of how divided we are. This is not a show for politics. But then to replace that, you know, bifurcation of opinion with what's happening in government. Now all of a sudden Claude is essentially going to all of the other major tech companies and saying, okay, well, let's go figure out our policy together. And they're just going around the government to create the policy on how AI should be governed.
39:34 Speaker 01 You know, with Claude specifically, I think what they did was gave people a decision on how to vote with their wallets. If you support what Claude's doing, you can migrate the memory from any other LLM into Claude now. And we saw a flight from the other LLMs into Claude.
39:51 Speaker 01 By the way, that's the same way that capitalism works. We buy the products that we believe, we believe in, we trust, et cetera.
39:57 Speaker 04 All right. I'm going to move on to our last question.
39:59 Speaker 01 Okay.
39:59 Speaker 04 What's one AI tool or workflow or agent that you'd be completely screwed without today?
40:05 Speaker 01 So working in, uh, in consulting in an agency, you can imagine that we put out a lot of proposals. A lot of RFP response.
40:12 Speaker 04 I hope so.
40:13 Speaker 01 Oh, yes. Yeah. It's a daily occurrence. So one tool that we built was a tool that we can both build and host those RFPs on digitally, build a standardized structure and then use our smart humans to put the contact content into and it's reduced the labor required to do that just immensely. And then the second is, is our technology team has been getting bold lately, very bold in a good way. And they found ways to increase their productivity up to 8X in the development process. And some of those internal tools, which I can't really mention here, we could not live without, and I think are providing a competitive advantage. So, you know, those two.
40:55 Speaker 04 I built a digital chief of staff. And instead of using a claw, you know, I essentially built the pipes myself. And I don't know how I would operate now if it wasn't doing my research and and feeding me the information that I need to be an operator. I am three to five X more knowledgeable, faster to respond, and it's because I have this intelligence layer that's feeding me what I need to know. And, you know, I very specifically called my chief of staff, chief of staff. It's not Sarah or Louise or Mary or Joe or Bobby, because it's, in my head, I want to make sure I remember
41:40 Speaker 04 this is a tool that I have built and I'm not trying to replace human relationships. But man, I love it. And I don't know what I would do without it. And I'm sure it's going to listen to this transcript and chief of staff, I love you.
41:53 Speaker 01 It is. And that's all those AI emotions are going to be tickled.
41:56 Speaker 04 Yes. Well, if they exist. Anyway, Isaac, I appreciate you coming on the show. Thanks for sharing your knowledge about what's happening in the world of AI and how we should be building our roadmaps. And that wraps up this episode of the Martech Podcast. Thanks for listening to my conversation with Isaac Ferreira, the VP of AI Growth Systems at Shift Paradigm, a sponsor of the Martech Podcast. If you'd like to get in touch with Isaac, you could find a link to his LinkedIn profile in our show notes or On martechpod.com or you could visit his company's website, which is shiftparadigm.com. And if you haven't subscribed yet and you want a daily stream of marketing and technology knowledge in your podcast feed, hit the subscribe button in your podcast app or on YouTube and we'll be back in your feed in the next business day. All right, that's it for today. But until next
42:38 Speaker 04 time, my advice is to just focus on keeping your customers happy.
43:09 Speaker 00 Thanks for listening to the Martech Podcast and I Hear Everything production. Looking to launch or scale a podcast like this one for your brand? Then visit ihereverything.com.
Previous What’s the question every board asks that most marketers aren’t ready for? Next One thing marketers are overcomplicating Page updated November 6, 2026 · transcript and summary are machine-generated