2025 Finale: “Everyone’s Insane”, AI, Burnout & What Actually Matters Going Forward

Speaker A: AI is the only thing that people seem to care about these days. For better or worse. I would say for worse.

Speaker B: There's a Venn diagram of OCD and

Speaker A: optimizers and it's one circle.

Speaker B: The superheroes, the superheroes Infect me with

Speaker C: the new Peptor me till the next step.

Speaker B: The superheroes, the superheroes.

Speaker C: Welcome to the show, Sonny.

Speaker A: Thank you, Andrej. Thank you for having me back and it is always great to be on this podcast.

Speaker C: And, uh, welcome, Shiva. Nice to having you.

Speaker B: It's amazing to be here. And we definitely didn't record this part after we ended the podcast. Definitely didn't do that. That did not happen. Okay. This is the start of the podcast.

Speaker C: You hecked it. You hecked it. I thought we can just have like a wrap up conversation of the year and maybe get to pivot about what we expect expect from 2026. Um, I mean, we all talk to many really clever people out there, so we might see something. I don't know, maybe not. We will see. So this is why I did not prepare any kind of questions or.

Speaker A: I think that's enough. I think that's good enough.

Speaker B: Uh, so the topic is what we are going to see in 2026.

Speaker A: But also 25, right?

Speaker C: Yeah. But also what we saw in 2025. Triangulating the pattern. If what we see could continues. What about 2026? But let's have fun in, in the first place, right?

Speaker B: So if you have to say, let's have fun. Yeah, we're just going to do it. You don't have to say it.

Speaker C: Yeah. So maybe it's the wrong topic for having fun.

Speaker B: It's like the, uh, it's like your boss who's like, all right everyone, let's go do a fun thing with drinks across the street. And everyone's like, I don't want to do this anymore.

Speaker A: Let's use AI.

Speaker B: This is stupid. This will be the best podcast recording ever. Except for from A to B.

Speaker C: Let's just say I'm sure. I'm sure. So we just started in the warm up, like what were our most favorite.

Speaker A: What's your flex, Shiva? Uh, speaking of best podcast I have,

Speaker B: I have second place. I have a chameleon second place award downstairs.

Speaker A: You self printed a T shirt and stickers.

Speaker C: So you were just saying, uh, Sunny, that your best episode was with Olga from Semrush.

Speaker A: The most eye opening one. Yes. Because this was really the first time I've had a world class SEO person on the podcast. And I think for future of CRO SEO past matters more than anything else because they are the people who have learned how the machines work and how to optimize for machines. And CRO as an industry should be looking at the past of SEO and the present of SEO and learn from those people.

Speaker C: My favorite guest, like from the depth of the conversation with Anoosh. Do you know Anoosh? He's a growth hacker, Anush Adhiya. And it was such a fantastic conversation about customer value and what blocks people and companies. Very inspiring. When I look at the number, what was not the most successful one? Based on the numbers I posted on LinkedIn, the most successful one was tonight talking about the Optimizers DCs. Everybody liked this.

Speaker B: I mean obviously Thanh just gets eyeballs. He just is that allure type of person.

Speaker C: Yeah, yeah, maybe. I don't know. What about you, Shiva? Who was your most favorite guest?

Speaker B: I'll answer that. I want it. I was curious if you guys. I'm going to take a random 5 second detour. Do you guys hate the term hacking in what we do in CRO? Digital marketing?

Speaker A: Do you know what my podcast is called?

Speaker B: No Hacks.

Speaker C: So you hate hacking?

Speaker A: I despise hacks in any form. Yes. People, techniques, anything.

Speaker B: I almost hate that. And there's like, there's some people who are posting in the CRO space who keep on talking about CRO hacks. I hate that because it almost feels like both of those. You can't say CRO and hacks at the same time because CRO should be a process, it should be following a specific methodology. And hacks don't. Hacks purposely break methodology. Right. So if you think about like what is a CRO hack, it's like saying, uh, like dry water.

Speaker C: So I say I love hacks. I love hacking because it's actually a, uh, personality trait I'm looking for. For example, if I'm hiring a developer, I don't want somebody who's just following rules. I, I want somebody that solves a problem. And to me, hacking means digging deep into a problem and solving it maybe with an unconventional way. And that's what I appreciate, uh, as a trait for many people. But what you mean with hacks is superficial.

Speaker B: I'm with Sani.

Speaker A: I think we disagree on the definition.

Speaker C: Yeah, by what you mean with CRO hacks is like superficially copying tactics.

Speaker B: No. So I think the what is communicated with hacking and CRO is something that's very like, to your point, it's like let's break the standard, let's do something to shortcut you to getting money quickly. But if you like, look up the definition of hacking, it's literally identifying and exploiting weaknesses. So I think in theory that's not the worst thing in the world. But I think in actuality how it's used in our space is like, oh, instead of following a proper process or thinking about redoing the process, which I think that's what you're looking for. Hacks is fuck the process, get to the end result by any means. And that, uh, by any means can lead to perverse incentives. It could lead to other bullshit. So that's why I'm like, I don't love hacking being associated with what we do because I think it downplays what we actually do.

Speaker C: I know. And this is maybe why growth hacking had a hype as a name of a discipline, but kind of diminished over the time because people cannot. I know we do hackathons with our clients, uh, also with AI. We put people from different teams together, let them solve something where they're lacking some skills, uh, under pressure, short time, um, great results. This is hacking. So this is a hackathon and people love it. But we have big issues selling this because companies say, well, ah, I'm not sure if you want a hackathon. This is something with developing and hacking. I said no, no, no, no. It's just a mindset how you approach a problem. And yeah, so the term is. Is there for being misunderstood. I think.

Speaker A: Yeah. This was Sheila's 5 second detour, by the way.

Speaker B: Yeah, no, whenever I say five second detour, that's a whole episode.

Speaker C: But why did you call your podcast no Hacks Show?

Speaker A: Because I think anything that definition of, uh, well, no, it wasn't because you started yours two years after mine. Uh, uh, a hack is really doing something where you know, someone's getting the short end of the stick. Eventually it's going to be your future self is going to be the person you're doing it with. It's going to be the system. Someone is going to get screwed in the end. That's what a hack is. A hackathon is like a performative activity where you're trying to see if something will work. It's not something that you will apply in that same way. Long term, that's what a hackathon is.

Speaker C: But if you hack the system for better results and you improve the system,

Speaker A: but that's not hacking. Hacking the system is cheating the system. Really.

Speaker B: Yeah. I think there's a negative connotation with hacking. That generally persists everywhere. In my opinion, it's exploiting things that

Speaker A: should not be exploitable. It's just because they haven't caught that way of exploiting the system yet. And in five years it's not going to be possible.

Speaker B: So I was curious about this. I think there's a secondary definition of hacking where like, if you have an axe and you have a tree, you can hack away at the tree. Right. And I think, I think that's where hackathon actually comes from, where it's like, it's not exploitation, it's the physical act of like, you take a big project and you hack it down into something that you just do quickly. I am not a linguist, but I do think there's something to that.

Speaker A: I like that. I'll just quickly go to stand up comedy. Ah, a hacking stand up comedy is like the worst kind of insult. You can, you can give anyone a hacking Stand up comedy is basically the vibe coding version of a comedian. They're just pretending they're trying to do something that they see others are doing. They don't understand what they're doing. But this flies the audience. TikTok likes this. I'll do this. I don't care about it, but I'll do it. Hacks are bad. Uh, we can all agree hacks are bad long term. You cannot live off of hacks. You can have short term solutions and that's fine.

Speaker C: So there's a difference between the definition and the ethicity of it and, and the mindset how you approach a problem. And what I like about hacking is the mindset of how to approach a problem. And, and I see so many of our clients trapped in processes and following the rules. I, I don't think following rules and processes is not the answer to great success. It's maybe improving processes may, maybe you have to hack them to get better processes. I don't know.

Speaker B: I don't think that's hacking. I think that's just being problem focused.

Speaker A: Me neither. But, but we can, we can disagree on, on the definitions. But uh, I agree with you 100% on what you're saying, Andre. Absolutely.

Speaker B: Well, and it all depends on the rules. Like if you're saying the rules are like, you can't, uh, like legal rules or uh, like otherwise there are some rules that have to be in place and that's where, like when you think about the Matrix and hacking. Right. They broke the rules of a system. Although I guess that maybe that kind of goes to your analogy, Andre, where you want to Break out of the system.

Speaker A: But this is still a 5 second detour. Just a reminder.

Speaker C: Yeah, yeah, yeah. We are still in the detour. So now back to my question. What was your favorite episode? Shiva.

Speaker B: So my favorite episode was with Kelly Hopping, who's the CMO at Demandbase. And she's like one of my. I worked with her at Gartner. Um, and she was such a smart person and I think what she gave to me was a lot of leadership level. Like, what does leadership truly care about? And KPIs and dumbing things down. And like, when I might be frustrated at something someone says, like, I don't want to run tests because I only want to do the things that will win. I was like, that's a really stupid thing. And she's like, well, you have to like, think about where they're coming from too. And there's things like, understand that they're not. Maybe they're saying the wrong things. But really what they mean, if you had to translate it, is I want you to prioritize the things that will drive impact. And they didn't say that, but that's what they're communicating. So she was almost like a decoder for some of like, leadership things. So I appreciated that. And I think that got me. I'm curious what you guys think about this, but we're Sonny. Because I think this is more AI specific, but I think part of that conversation was AI. I think now in 2025 is getting us away in CRO. It's getting us away from thinking about business metrics and more into hacks. Right. Like the AI audit of your tool has no idea what your business impacts are or your data. And even if it did, I think it'd do a pretty shit job of analyzing that data from some of the stuff that I've seen. So I think what AI is doing is getting us away from the most important part of CRO, which is understanding business metrics, understanding business goals, understanding business stakeholders, and optimizing for those things and understanding the user and optimizing for user metrics.

Speaker C: But I didn't get it. It's good that you emphasize that. Sunny is your AI expert here. Me, I'm not, but I get it. And I have to confess, I can't follow this. Why should AI act like this?

Speaker B: I think there's a false confidence that AI is doing that, but it's not. And I think we're cheapening CRO by using tools like this because there's a false confidence. Dunning Kruger, something like that. In the tools, providing an output, thinking it's doing CRO properly. But the people think that they're doing it properly. They have no idea what they're doing. But AI gives them that false confidence. Hmm.

Speaker C: Hm. Yeah, I get it. I see these kind of people who are completely hyped by AI thinking that everything that comes out of an AI tool is the truth, which is a lack of critical thinking. And there's the other part of people who are kind of AI neglectors say AI is evil. I don't want to use it, whatever. I hate it. It takes my job. So somewhere in the middle, that's where the truth is, right?

Speaker A: Always. Absolutely. Also, when we talk about AI, we're talking about generative AI. This was not supposed to be the AI five years ago. Generative AI is like saying CrossFit is all fitness. It's just the loudest, most cultish kind of bunch in fitness. It's Gen AI. If we'll ever have AGI or anything super intelligent or whatever they call it these days, those billionaires, it's not going to be Gen AI. This is a prediction algorithm and nothing more than that. It can be applied in many different ways. That is absolutely true. But this is. What do you expect the next letter or token should be? That's all it does. I think the way we'll see this era, if we're lucky, is it was a detour and distraction from more important things. Focusing so much on generative AI. Well, what you said, working on, working with the customers, working all those, even that, going back to that, focusing 80% of our attention on gen AI and nano bananas and VOs and all that, uh, who cares? Nobody should care about that. Absolutely not. If you didn't care about generating fake photos of you five years ago, you don't care about it now. It's just something that someone is showing you and you're just, oh, shiny toy.

Speaker C: I hear you, but I have to disagree.

Speaker A: Mhm. I want to hear it. I want to hear it. Yep.

Speaker C: Because I just asked one of our guys who is really good, uh, in AI things and automation, I asked him to rebuild our service completely with AI tools. I said I want conversionscraft AI because I think if I don't build it, somebody else will do. So now I'm using these parts to accelerate my internal processes and the result is I could accelerate it by four times.

Speaker B: What parts?

Speaker C: Nearly everything. Nearly everything. From discovery to test report everything. Building a database that finds patterns, uh, so my discovery is much quicker. I don't have to Reinvent anything it may be, finds CRO hacks, shiva.

Speaker A: Because I'm not saying it's completely useless. Just to be clear.

Speaker C: I give you one example. I have different teams working for different clients. So one team has the task to uh, increase newsletter subscriptions for a client. So they're coming up with different ideas, testing their way forward. And yeah, client's happy. Like a couple of months later, another team has another client. They also want to improve, uh, this. We all have these results in our database. Nobody's looking at it. I don't know why I looked at it and said, I see from like three different clients that want to improve newsletter opt ins. Nobody used the data. So now we changed all the data, made it accessible with AI and now we can chat with our experimentation database, uh, which is organized like a growth action tree. Uh, so now if I ask them, we have this and that client, this and that business model, this Audience please look at the data. Uh, what are ah, the tactics that delivered proven results with data? What are the pitfalls? Whatever, I get it in seconds. And this is just one example or test reports, people sitting like half a day, three people sitting, sitting together over all the data from a test report, trying to find the pattern, analyzing it. I gave it to AI. Five minutes done. So these are just two examples, um, how we accelerated our internal work with AI.

Speaker A: I'm sure that works for you. I'm sure it can work for a lot of people. This is not something the humanity and all tech should be obsessing over. That's just the way I see it. It's not that important and uh, it's not that game changing just yet. It might be in the future, but this is generative AI. This is pattern matching.

Speaker B: I'll echo that all. Yes. And I'm without being able to like see some of the details. I'm critical. I'm cautiously optimistic, but critical because I'm pretty pro. Like you can't just reuse or copy patterns from like one company into another company. Maybe there's some evidence to say, like there's some similarities and if you really go deep into it, it's like, all right, it's a newsletter for an E comm, B2C, uh, apparel. And it's like you niche down and then you see that the newsletters are the same. Then it's like, all right, maybe you can rip it. But I think the assumption here is that the newsletters are the same. Like everything's same, same, same. And then I'd also counter, like man, if Your newsletter is just the same as someone else's. Like, yeah, the pattern might work, but like, you're not. Why would someone subscribe to your newsletter over someone else's? So I think like what AI is doing is optimizing for the average. And what you're gonna get is average. You're not gonna get something that is differentiated, that people are like, that's bad. Like, who is doing my meme shit better than me?

Speaker A: Oh, a lot of people.

Speaker B: That's where I'm like, I'm trying to be different in the market and I'm different my content. I'm trying to be different with everything I do so that I stand out versus the other people. You look at a bunch of LinkedIn content, I'm not gonna name names, but there's a lot of people just using AI to write stuff. And you read it and you're like, this is the same fucking post, just Gemini'd seven times. It's like a deep fried meme. It's a deep fried content. It's the same shit. These people don't have original thoughts. So even if like I write an original thought, but I code it differently, cause it's my brain changing it, people will gravitate more towards that. And I think people don't understand that. People are smart enough, can and can detect AI. I, uh, talked to Sonny about this in our like holiday episode, or retro, whatever the fuck we want to call that. But like, have you seen that, Andre? Have you seen the ads where like the whole thing is AI generated? It's like this tai chi ad. Have you seen that?

Speaker A: Don't YouTube do you look at that

Speaker B: and you're like, wow, this is amazing. I got to buy this? Are you like, dude, this is so ass. This is so bad.

Speaker C: I know. I just saw an ad like this for cats. How to kind of, uh, unsharpen their claws. I don't know the English word for this. I was immediately hooked anyway, but I know what you're meaning also about humanity. But for my business. And it doesn't mean if it's my business or anybody else doing, uh, CRO or exploitation in an in house team. It's about being competitive, you know, and like, can I deliver an output for a certain price? So basically, what's my ri? And this is defined by the speed of my processes and I can accelerate it. And I disagree with what you say, Shiva, about the uniqueness. To me, I think there are basic levels where you need to provide a solution. And this is based on principles and I see people trying to come up with ideas to solve these principles, to solve these very rural problems. Like, am I attracted by a newsletter? Is the value displayed? How can I display a value for something nobody asked for, having it, whatever. I have to interrupt people. What are the patterns that work and that doesn't? So, of course, this is like a knowledge base where I first built the foundation. This is not about uniqueness and positioning so ever. Uh, many times we solve simple problems and they're already solutions. We already solve this problem somewhere else. And with AI finding the pattern better than a human mind, I just can accelerate it. I completely agree.

Speaker A: Let me just first say I agree with that. If you are willing to work on your process, on your workflow, if you're doing your job well, and then you also want to use AI to do some parts of it faster, that is exactly what you should be doing. I'm just wondering about the people. I get dozens of pitches, uh, for my podcast daily. It's always the same pattern. It's always the same lazy. I listened to a recent episode with this person, and this thing caught my eye. I have a guess who would be great. Those lazy people can go to hell. And those lazy people, uh, think they're doing important and good work, and they're just damaging everything they're touching. And that is a huge majority.

Speaker B: That's the problem. Uh, pun intended. Which is what I wanted to talk to you about, Andre. But I think that's the problem, is just because you throw AI into a process and you get an output, it doesn't necessarily mean it's good. It might come off as good. And I'm not denying what you're doing with your agency might actually be seeing positive results because you have access to

Speaker A: the data, because you had the process before AI. That's the whole thing. That's what matters.

Speaker B: And you, Andre, and I'm sure you have trained your team to be able to suss out bullshit when you see it. And that's super important. And. And I think what we were talking about earlier is, like, people who. I have a really interesting story. So basically, you guys remember Wolfram Alpha?

Speaker C: Yeah, yeah, yeah, yeah. Okay, we have five seconds for this Shiva.

Speaker B: Okay? So I have 20 minutes for it. All right, good to know.

Speaker A: Let me go get my coffee.

Speaker B: So, Sonny, Wolfram Alpha is like AI before AI. It's basically what, like a bunch of college kids, myself included, use to, uh, like, solve complex math problems for, like, online math things when we didn't want to do it. You'd Copy and paste a math problem, it would just solve it. And then our math teacher found out we were using it because the answer was technically correct. But the way that we got to the answer, it's like, it'd be like 1.333 rather than one and a third. And he's like, you obviously didn't listen to anything I said. You just used Wolfram. And he yelled at us. And I think that's like. Because you don't know. He was unhinged. He yelled at us. It was kind of, kind of scary. But, um, the ultimate thing is that if you can't understand and suss out good versus bad output, then you're in a big problem. And I think, Andre, because you understand what good and bad is, you're in a really great spot to be able to use AI. Because when you see bullshit, you're like, that's bullshit. But I think that's. We're transitioning into a place where people are blindly trusting AI's output. And then you're going to have agencies that are full AI from people who've never run a fucking test in their life. And they're like, well, gave an output. It's like, here are 10 ideas. And it's like, oh, seven of them are button color tests. And they're like, well, I read that blog post about button color testing. So it. And it looks good to me, but they don't understand that that's not a problem someone needs to solve.

Speaker C: By the way, we just did a button color test. It was, uh, nearly 2% uplift.

Speaker B: I test button colors.

Speaker C: I know, I know.

Speaker B: But if you have 7 out of 10 or button color testing, then you have a problem. That's a different conversation.

Speaker A: Hey, there's a lot of buttons on that. Uh, look, the problem is, you know, vibe coding and why. My programming background really despises the concept of being enthusiastic about not understanding what you're doing. That's. To me, that's what vibe coding is. Everybody is willing to vibe any other profession but theirs. When it comes to them, and they see the output, they say, AI is shit at this. It doesn't do a good job. Well, doesn't do a good job, a great job at anything else unless you drive it and into a good output. And that's how it works. And most people will, uh, look at those tools like Apollo and instantly and all those cold email outreach, which I really don't like. It gives them an email template, AI writes an email for them. This looks like a professional email. I'LL send this to 5,000 people.

Speaker B: It's quantity over quality.

Speaker A: It's really, really terrible quality and quantity combined. And that is one of the biggest problems.

Speaker B: So, Andrej, I wanted to ask you this question about the way you're building some of the stuff in the agency side. When we think about like, uh, I'm very problem focused when it comes to solving things. So it's like, what's the research? Tell me the problem and then solution for that. Right. I think that's how I generally approach things. When you have the way you were describing, the way you run this stuff is like you have solutions from problems that were solved in, let's say, for the newsletter in one company, are you then taking a similar methodology of like, okay, now let's look at problems and then let's look at other solutions. Are you still taking that approach? Are you just copying and pasting solutions across? No. And then see, and I think that's such an important distinction because that is really good. Like, that's a, that's a fantastic use of AI where it's like, you've identified the problem, you're still following the process. But there's not, not everyone does that. There's people who will just be like, oh, here's a thing we've tested. Let's, oh, other people have tested something. Let's just, let's just do what works. Let's just do what other people are doing. Right. Like that stuff does not work and there's a false confidence. Uh, I'll get off my rocker. I'll shut up.

Speaker C: Yeah, yeah, yeah. No, no, no. You're absolutely right. You have to understand the problem, but then AI helps you to see the patterns, how to solve it. So it's like an inspirational thing that accelerates beautiful something between understanding a problem and finding solutions. You don't have to recreate a solution from scratch, start from nothing. You can analyze existing solutions to the same problem and how to apply them. And that you can accelerate a lot.

Speaker B: Yeah. Not to glaze you, andrej, but Sani, I think Andre might be one of like four agencies or companies thinking about it the way that it should be thought about. Because all the other stuff I've heard and seen is no one actually following that. And I think that's such a great use case.

Speaker C: Um, thank you. But I just realized that's maybe, um, I just had a conversation with a client saying, well, but we expect lower returns from our CRO program next year and I say, why you discovered m much more problems. Um, this year than the year before. And I said, yeah, but that's how it goes usually, right? So we earned, we yielded all the quick wins. So what's left? I said, no, no, no. We just started to understand a problem and see a pattern. So if you exploit that now, you will get much higher returns because we just started to understand the problems. Then I realized in his mind, a B testing means, like spaghetti testing. We are throwing like our 70 tactics against the wall. See what sticks. We are out of tactics, so we're finished and returns will, uh, be lower. But that's not true.

Speaker B: Huh?

Speaker A: Huh.

Speaker B: I want to add to that. Yes. And the point about focusing on like, oh, AB testing, once you've optimized for the local maximum, then you're good. Once you've identified a theme or a pattern or something strategic, then that's where the unlock for AB testing is. You take that concept and it's not just a B testing. You go work with your creative team, your legal team, your marketing team, your paid search team, and then you take the concept and expand it. And, uh, that has to be reinvested in CRO. But you invest in that concept that A B testing found. I think that's where maybe your stakeholders like, oh, we optimized for this one thing. We're done. It's like, no, no, open the Pandora's box and get everyone else involved in this because there's so much more we need to build to support this concept that's been kicking butt.

Speaker C: Yeah.

Speaker A: Let me ask you guys a question. So you mentioned a client who thinks it's only about spaghetti testing.

Speaker C: I just realized that this is in his mind. Like this is his mindset about a B testing.

Speaker A: And I'm guessing this is a, uh, this is not some. Someone who just started doing business they've been in. Exactly. Isn't that a problem? Like, 20 years in CRO is still seen as a spaghetti testing, throw it on the wall, see what sticks Kind of industry reputational problem. That is unbelievable and mind blowing. That this is still sunny.

Speaker B: Don't you think that's a problem? That, like, war happens a lot?

Speaker A: No, no, no, no, no. But, but SEO people know what to expect and what SEO is. For example, CRO people still think this is just throwing shit on.

Speaker B: I don't think serious people necessarily. I mean, uh, to be fair, not all of them.

Speaker A: Not zero people. Not. Not CRO people. People, agencies, the company.

Speaker B: I don't know. I can't speak for.

Speaker C: Maybe it's the CRO agencies who are throwing ideas against the LinkedIn audience, uh, all of the time, and they get hundreds of commands to get the secret list of 122 proven CRO tactics that work.

Speaker A: But you have to comment to get the deck.

Speaker B: Yeah, that's what I'm getting. See, that is the same shit as, like, vibe coding for you, Sonny, where it's like. It's a promise of not having to do. At the end of the day, no one wants to do fucking work. Like, no one wants to do anything to actually make something good. They just want to be presented with whatever it is and just do it and have money.

Speaker A: I hope this is the soundbite for the episode. People fucking hate the thing you just said.

Speaker B: It's true.

Speaker C: No, people hate doing our work. People hate problems. And I think this is the big part of our work, discovering problems. And people always want to go to the solution space because it works.

Speaker B: It's work to do that people hate. People are allergic to work. I think it's. I think it all just comes down to just. People are lazy as fuck. Which is why, Andre, you're like. That's why they love hacks.

Speaker C: Of course, M. I'm lazy as well.

Speaker A: So, no, I don't think it's everyone that's like that. I don't disagree with what you said, Shiva.

Speaker B: Uh, maybe. Hold on, Sonny. There's not a single thing that you're lazy about.

Speaker A: No. You said hate everything, hate work.

Speaker B: No, uh, people hate working generally. There are some things where people will be, like, more motivated to work through.

Speaker A: And that doesn't explain those lazy comment to get my slide deck thing.

Speaker B: I think people are lazy with certain things. Fine.

Speaker A: That's just fine. Okay, There we go.

Speaker C: People don't like problems. They favor solutions over problems. And CRO. It's Easy is always about, uh, discovering problems. Great product work is about uncovering problems

Speaker A: and solving them, which is why AI cannot do it fully, which is exactly why AI will never be able to fully do it. Because AI is a rearview mirror. It's looking at a history of all the problems that. It's like a compilation of best practices. That's.

Speaker B: That's the way AI it's optimizing to the lowest average.

Speaker A: Uh, from the past. From the past.

Speaker C: I always tell the story about why people buy a BMW.

Speaker A: You.

Speaker C: You guys drive a BMW?

Speaker A: No.

Speaker C: No. Good. Oh, it's.

Speaker A: By the way. By the way. The new CEO is from Serbia. They just, uh, named him yesterday. It was big news in Serbia. There's a guy who's. BMW CEO from Serbia.

Speaker B: We're gonna get all the hate from the BMW drivers in the comments here.

Speaker C: No, Maybe, maybe I, I usually, hey,

Speaker A: they should use, they should use turn signals.

Speaker C: And then, then no, I, I tell this story because it's, it's a real, it's a real project I've worked on. We did dozens, uh, of interviews with BMW clients uncovering their um, buying motivators. And this is what I also tell publicly in my talks. I said, nobody said I bought this car to impress my neighbor and my friends because I have a small ego. Right?

Speaker A: Large ego, large ego.

Speaker C: Um, and I need to compensate some trauma from my childhood or whatever. Nobody tells you that in a user research. But guess what happens if you analyze it with AI? AI is overcoming social biases and can tell you the truth about buying motivations. So if you're familiar with the jobs to be done framework. And I was fascinated with this like uh, days of research work. Ah, replaced by a prompt in AI, not the same depth, but the clarity about social rewards and motivators that people won't tell you in an interview.

Speaker A: What if LLMs just know those things about BMW drivers and they have that in the knowledge base and that's why it figured out this is the problem. Because the Internet has been saying this about BMW drivers for decades.

Speaker C: A human being won't tell you in that clarity. Never. Social conformity bias and use the research. So I see some. This is just another way of how you can use AI for the better. But it still needs some human being, um, an expert, not just an expert make sense out of it.

Speaker B: So yeah, I'm curious if you guys hold this opinion. I think is this AI solving that problem or is this more on that UX person not getting to the heart of the right results and asking better questions or like doing their own research or doing their own hypothesis generation to get to the heart of that question. And um, I don't think there's a right or wrong because I get like to me the perfect world is AI will help uncover certain parts of research that humans won't. And I think there's a coexistence of like both will help get you to the right answer because they do things in different ways.

Speaker C: But then you're not a user researcher anymore than you're a therapist. You need maybe 20 sessions until people open up and say, hell yeah, I have a problem with my ego and I want to impress people and show them how successful I am. And then they break into tears. And how should you do that with the user research? You can't yeah, uh, probably.

Speaker A: Probably not. You're probably right.

Speaker B: I guess the counterpoint to that is, like, how would you actually suss out when AI is hallucinating? Like, you know what I mean? Like, if what happens in that scenario where maybe AI is hallucinating the user's actual intent.

Speaker A: You need to be an expert. You need to be an expert. That's the whole thing. Uh, I was reading a thread, like Threads, the app.

Speaker B: Are you, like, one of ten people on Threads?

Speaker A: I was shocked, but I was reading. There was.

Speaker B: Exactly.

Speaker A: It's, uh, Instagram. Yeah. Anyway. Anyway, I. I checked it for. I don't know. It's been a while. Let's just say it's been a while. There was a, uh, scientist writing there about how up until five, six years ago, he thought that Elon Musk was this creative genius because everything he says sounds smart. And then he talked about my area, like, the SpaceX, whatever, and. And then I realized this guy has no fucking idea what he's talking about. And that's AI. AI is Elon Musk, and that's donning karaoke.

Speaker B: Yeah, it fosters Dunning Kruger to, like, an absurd degree where people just have this false confidence that they're experts in the field because. And, uh, by the way, this is like. This is why you see this in interviews too, where, like, when you're, uh. And I've been conducting. Helping, like, conduct some interviews for hiring some folks on the team that I like, work at Motive. And there's like. We've. I've noticed this where there was one candidate I was talking to who was very obviously using AI to answer questions. And it's like, you could just tell, like, I understand the concepts. And when he was answering the questions, first of all, his eyes were just reading. And I was like, all right, that's weird. But, like, a person who understands it can suss out the bad shit. And it was. It was. I'm not going to throw this person under the bus. I'm not going to call it out. I'm just saying it was very obvious to me.

Speaker C: But I hear you guys. You seem to be so anti AI. Like, don't use AI. Should I use that as a title of our episode? Guys, don't use A.I.

Speaker A: no, no, responsible. But I never said that. Don't use it blindly and don't expect miracles unless you're willing to check what it does and do some work.

Speaker B: Yeah, it's responsible use. I think we glazed you up pretty well on your use for AI and how you're optimizing for systems, you're optimizing the process and you're like, you have a brain, you've mapped it all out and you're like, these are. And I talked to Chris Mercer about this, he was super sharp and he gave me some really great advice and he's like, don't end to end AI. Don't just like, hey, build the tools that help you build apps from scratch with three prompts. That kind of stuff is prone for a lot of risk. But instead you as an expert, map out your whole process end to end and then figure out like, okay, this thing, this is a really great use for AI. When I have a, ah, hypothesis, a problem statement, data, I need you to build a test doc and here's the template. Beautiful use for AI because that's super manual TD stuff. I don't want to do it. And then other things like uh, when you have the inputs for, here's the problems that these users face. And then intelligently use AI or LLMs to help you figure out and group the problem statements across clients and then solution there. That's really cool.

Speaker A: Use case, I would go, don't use AI unless you know why you're using AI. And that's good enough advice. Like if you're just using AI because I'm supposed to be using AI, you probably don't know why and how you're using. But if you're using AI because I need it here and here and here to do this and this and this, that's a different story. And that's kind of what Shiva said,

Speaker C: but without asking the question. I mean I wanted to do like a wrap up of 25 with you guys here. We ended up with talking about AI for 5, 4, 5 seconds. Was that the most important topic in 25 AI?

Speaker A: I think it was. I mean undeniably in tech, AI is the only thing that people seem to care about these days. Uh, for better or worse, I would say for worse. I hope that changes. I hope we go back on track and just treat this as a distraction at some point. Because you're not an AICRO person. You should not be an AIC person. You should be a CRO person that uses AI to 100% to help you in some parts of your process and workflow. And we see a lot of, we see a lot of these AI magic and then it's AI SEO, AI CRO, and it's just horrible.

Speaker B: Andre, I'll put it back on you. If AI is not the most important thing to talk about in 2025, what was the most important thing for you? And you can't use the words AI at all.

Speaker C: No, no, um, to me, it wasn't, it was a lot about people and how they work together and how they, how they change things. What enables them to change things to maybe make work more meaningful. I kind of doubled down on the idea of who's our audience? It's people who are sitting somewhere in the basement of a huge company and saying, well, we could do better. This is what Ton called the optimizers disease. People who can't stand the status quo and want to make things better, and they have so many barriers to overcome. So, yeah, in many episodes, we talked about, um, this kind of stuff. We did not talk so much about AI. Maybe I would have more listeners if I pivot the podcast towards AI.

Speaker A: Oh, no, that's specifically talking about the podcast. I was talking about work and tech in general. That's where AI has to be the number one thing.

Speaker B: We all are sick. We all have diseases.

Speaker C: AI disease.

Speaker B: No, you said optimizers disease.

Speaker C: I mean, that was, the term was coined by Tom.

Speaker B: I know.

Speaker C: And with disease, he means the negative, uh, parts of that. If you can't stand the status quo, maybe you're over optimizing small things. You don't see the bigger thing. So this is what he talked about in this episode.

Speaker B: Oh, interesting.

Speaker A: I think that's ocd. Like, literally what OCD is.

Speaker C: What is ocd?

Speaker A: Obsessive compulsive disorder, where you have to clean it.

Speaker C: Ah. Ah, Okay.

Speaker B: I think there's, there's a Venn diagram of OC and optimizers, and it's one circle.

Speaker C: Yeah, no, I mean, in my world, uh, outside the podcast. Right. It's a lot about how, how companies reinvent themselves and, and, and change. There are still. So I, I, I also wrote about digital tailorism on, on, on LinkedIn, because people still think they can, like, uh, split the work and little things and create processes and then work on this. Still great, uh, outcomes. And this isn't the case anymore. So we want empower teams and how to empower them, how to convince stakeholders and bosses and why you should not call them hippos. And I sometimes think my life, um, has shifted away from optimizing towards, like, organizational development or something like that. I talk about stakeholders and C level and whatever and business cases much more than ever.

Speaker A: But that's a good thing. That's a very good thing. That, that's a natural progression.

Speaker C: Yeah. I mean, depends on with whom you are talking to. There are many people who are, as I said, stuck in an organization and not able to change things. And I always try to, as you said, Shiva, the definition of hacking, making smaller, uh, chunks and getting things in smaller projects and approaching problems bit by bit, helping them. That's what fulfills me. And. Yeah, that was good. That was good. You have to stand the, uh. Some people are really sad about why they are stuck in an organization and can't move forward.

Speaker B: Well, can I bring up one thing kind of parallel, uh, to what you're talking about with people and optimization? So I've seen this where there's people who are just like, dude, fuck AB testing. Fuck optimization. I'm, um, out. Like, I think that kind of is to your point, Andre, but when they

Speaker A: say out, where did it go?

Speaker B: Like, career change. So it's like they get out of experimentation. Um, I know someone who was like a very prominent product manager, and they're like, building experimentation and they're like, I just want to go build a daycare. Like, fuck this, I'm, um, out. I think that's not unique to AB testing. I think that's just like a general tech burnout where people are like, I just like, this is such a shitty rat race. I hate all of this. I want to go do something that feels meaningful and impactful. And I'm sure part of it is related to the idea of, like, there's CRO is getting. I don't know if the trend is that it's getting more resistance or less resistance. I couldn't tell you. Uh, I'd like to think it's getting less resistance, but I'm, um, not. I honestly can't. Couldn't tell you.

Speaker C: If you compare it to a time like 10 years ago, then you see maybe the bigger pattern and the evolution.

Speaker B: Do you think there's more resistance to a B testing? No less. Less.

Speaker C: Less.

Speaker B: I think so too.

Speaker C: It's really slow. It's really slow. You have to think in long terms to see the maturity growing.

Speaker B: Yeah, yeah, yeah, I agree with that. But I think there's just like a general burnout with people doing testing, doing digital in general, but specifically a B testing where they're just like, uh. It's like three companies in a row where I join and I'm sold this position of AB testing, optimization, accountability to do shit. And then when they're in it, they're like, I can't do fucking anything. This is. This sucks. And I think part of. And then. And then on top of that. They see all this AI bullshit that's like. And. And leadership looks at the AI bullshit, and they're like, oh, I don't even need you. I'll just have AI do everything right. And I think there's just, like, a, uh. Dude. All right, fuck this. I'm gonna go build a daycare because at least I could help some kids. You know what I mean?

Speaker C: Yeah, yeah, sure. I mean, that's what fulfilling that what makes you. You happy is helping other people. If you don't have the impression that what you do is something that people want and that helps them. Sure.

Speaker B: I was talking to Talia about this, and she was talking about, like, a client that she was working with, kind of a smaller client. She's like, it's tough to, like, help them out with stuff because they have such small traffic volume. And I'm like, they're the people who need the most amount of help because they don't understand digital like any of us do. And they trust you that even if it's not an A B test, there's a lot of research. There's a lot of stuff you could do, even just, like, baseline. I don't want to say best practices, but let's just say website foundations that you could probably help them out with. They need the help and they don't understand it, and they might get sucked into the bullshit AI, but they need someone who actually understands that stuff, and they do need the help, and you can help them out.

Speaker C: Maybe we should do an episode about, uh, burnout.

Speaker B: That'd be a good one. That'd be like a really good episode. Oh, wait, did we do that or did we do the conference episode?

Speaker C: Both.

Speaker B: Oh, yeah, that's right. We did both. I thought about that.

Speaker C: It's a long time ago. Yeah, long time ago. So. But I wanted to talk about what. What really moved us the last year. So what was it, apart from AI on your side? Shiva.

Speaker B: Like, positive.

Speaker C: Yeah. Yeah. Maybe also negative.

Speaker B: Go, go, Sonny. First, let me. I need to, like, I've been in a. I've been in a doomer mind. Let me, like, get into my happy place and I'll figure out what made me happy.

Speaker A: And I may sound like a doomer, but honestly, I'm not. I'm just realizing that this was a distraction, the whole AI and I think humanity will recover and we'll be fine. Uh, what moved me and what the most positive thing about this year was, I mean, I have to go back to the people I met thanks to the podcast and through the podcast and through the community. I mean, when I was doing my keynote two and a half weeks ago, I had 30 people in the first three rows there. 30 of my previous guests from my podcast were.

Speaker B: I thought you said 29.

Speaker A: 29. Okay, but, like, cool. Uh, but that.

Speaker C: You're exaggerating.

Speaker A: I'm glad you were not one of them. But, uh, having connections with people like that, that I still chat with daily, I talk to daily. I'm, um. Today I spoke to two of you, and I spoke to two or three more people who I met through the podcast. And that. That's. And I'll speak to two more after this call in eight minutes. That is insane to me that. None of that, uh, the thing I told you when I was on your podcast first time, Andre, I still don't believe this is happening. Like, I still need to pinch myself to realize that this is connecting people. This is real. This is actually happening.

Speaker C: I just received an email from Brazil from a listener of our show, and I thought, wow, that's amazing.

Speaker A: Nice.

Speaker C: Yeah, that's amazing.

Speaker B: How.

Speaker C: How's the weather in Rio?

Speaker A: Yeah.

Speaker C: Uh, connecting people.

Speaker B: Yeah.

Speaker A: Uh, incredible. Yes, yes, yes. I still remember my first message I got from a listener, Simon, uh, Clark. I'll name drop him. And he told me, dude, I love. This was 2021. I love listening to a podcast while I'm running. And that was. For me, that moment was, wait, someone's listening to my podcast? Are you serious about. And those moments are real, and the rush you feel from them is. Is difficult to describe to someone who has an experience.

Speaker C: Absolutely agree.

Speaker B: I think there's a similar. Just like, there is a community of experimentation, people who are incredibly thoughtful, who are incredibly down to earth. And it's like Sanni and Andre, both of you guys have just been genuine human connections that I hope sincerely we have a chance to meet in the near future, grab a beer, and not talk about any of this bullshit. And we could talk about Pluribus, or we could talk about literally anything else, because I know we could. I know we could talk about anything else. Like, if we met at a conference, I don't think we talk. We might bitch about CRO stuff for, like, five minutes, but we talk about so much other stuff.

Speaker C: Five seconds.

Speaker B: Five seconds. So, like, three hours, really. Um, but I do think there is, like, this community of authentic human people and where I am, like, I might be perceived as anti AI, and I think there's a lot of things I am anti AI about. I think it's because I'M more pro human connection and humans doing human things and authenticity. Like, there's no, I'm not, there's no fakeness about what I am on the Internet. Like, I don't lie about anything. You might not like me and that's totally your M.O. i'm kind of an unlikable person sometimes. Fine. But you know where I stand with me and you understand who I am. So I think like, authenticity is something that this community has a lot of and there's a lot of people who will just like. To Sonny's point, I've gotten similar super nice messages. I see like people wearing my shirts in the wild who purchased from me. I'm like, that's just like such a cool thing that you built something cool enough that people are like, I want to support your pod. And that's cool as shit. I love that.

Speaker C: Yeah, that's amazing.

Speaker B: There's a lot of like authentic humans who like, genuinely care and, like, you know, go to conferences. It's such a, like, fun thing. Because I don't know about you guys, I don't think I've talked that much about CRO at these conferences, at least in the later times.

Speaker C: No, not a lot. It's about connecting people. That's why I, uh, say it's a main value proposition of gms connecting like minded people.

Speaker B: Nice plug.

Speaker C: Yeah, that's it. Period.

Speaker A: It's his podcast. It's GMS podcast.

Speaker C: Will you come, Shiva?

Speaker B: So you know what you should do, Andre? You should have, uh, promo codes. So promo code Shiva gets you 10% off. Promo code Sani gets you 10% off. Never figure out which people like who, who more.

Speaker A: That doesn't work.

Speaker C: Yeah, make it a competition. Yeah. So last question. What's. What's coming in 26? What are you guys expecting? What are you, your plans?

Speaker B: I think part of me is like, I. Personally or professionally? Because if you say personally, I think part of it's just like disconnecting a little bit from like the CRO stuff and metaphorically touching grass more. Because I think that's just like there's a personal thing around, like all the ads, all the stuff that's just on the Internet, on LinkedIn. It's just like, it's exhausting. And it takes me away from any type of motivation to like, want to actually do high quality work. So I'm trying to like disconnect myself from that and do more fun stuff like skits and like, have fun and poke fun at it. Because I think we all need a break from Just like the monotonous bullshit of work. So I want to try and like, focus more on that positivity and that happy stuff. Because I think to Sonny's point, I'm similar. You asked like, what are you happy about? And I'm like, I'm not like super jazzed about a lot of stuff in the work stuff, but there's a lot of personal stuff that I'm excited about. So it's like, let me just focus on that. And I need to do work to put food on the bill, food on the bills, put food on the table. And like, I like my coworkers, I like a good chunk of my work. Um, but it's not all like, great. There's some kind of frustrating stuff, as is normal with work.

Speaker A: Yeah.

Speaker C: So if you say AI is a distraction, how would you describe LinkedIn?

Speaker A: Poster child for distraction these days. Like, I don't even know.

Speaker B: I think it's ripe for being distractible, but I try and add some amount of guardrails and cover that. If I'm like, If I read 5 AI posts in a row, I'm like, I'm good for a couple hours. I'll come, I'll check back later. I'm gonna train the algorithm. That's bullshit. I'm gonna abandon.

Speaker A: I, I think LinkedIn, the way we talk about it now that, that LinkedIn is dead.

Speaker B: Yeah, true.

Speaker A: Honestly, never going to be honestly as a healthy platform. I mean the, the Corey Doctoral and shidification book is literally about this thing happening with LinkedIn currently, which book I

Speaker C: didn't get, I think.

Speaker A: And, and shidification and shitification. Yes. Platforms becoming worse over time and by design. Uh, So I think LinkedIn as we know it doesn't exist anymore and never will exist again. So I think mostly it's a waste of time unless you have strictly business goals that you want to. But building.

Speaker B: But even then, can you achieve the business goals on that platform?

Speaker A: I don't think you can now. I think it was possible in the past.

Speaker C: So.

Speaker A: Second thing is paying attention to everything happening around AI. 99% of it is noise. I think it's healthier to just not pay attention and get the news a week later. Who cares about it? Honestly, that's going to m. Be my approach in 2026. I don't care about ChatGPT 5.15.25. It's the same thing, it's just a different name. Uh, what I will focus on professionally is I want to understand the agentic web and machines using the human built Websites, websites built for humans. And I will focus on my entire. Strictly on that, including the podcast. So this is going to be the only topic I care about professionally. How can we make our website, websites, uh, easier to understand for those autonomous systems? And that's it.

Speaker C: And, and why do you think. I know you have a hard end, but how do you think, uh, humanity will overcome AI as a distraction?

Speaker A: We'll stop caring about it because they're lying to us. Everything that, that has been promised since November 22, 2020, when ChatGPT was launched. Is it really that much better now than it was three years ago? Maybe it can write better sentences also to get a lot of things done, um, with stuff like you're doing in your workflows. We don't need it to get better than it is. This is fine. Stop. Do something healthier. Do something better. Just leave it alone. And we can use the open source models and just continue implementation. The implementation. I, I don't think we need more AI. The world doesn't need more AI. That's my take.

Speaker B: What are you looking at, Andre?

Speaker C: Yeah, Oh, I will, I will double down on, on customer centricity because I think it's. We are helping people to, to solve their problems and, and look more on the human side. AI is helping me to accelerate it, but basically, yeah, I think experimentation is a means to an end. And, um, what people really want is greater, better experiences and make an impact. So, uh, I will double down on that one. So kind of similar, uh, I think

Speaker A: every human being, every sane human being is in that same mindset.

Speaker B: There's a lot of insane human beings. Um, then.

Speaker C: So let's finish it with that words of wisdom. Thank you so much, everyone.

Speaker B: Insane. All right, zoo, you guys.

Speaker C: Next year, Everyone's Insane. That's the title of the episode, Everyone's Insane. Thank you guys. That was amazing. I loved it.

Speaker B: Thanks, Andre.

Speaker A: Great to talk to you.

Speaker C: Likewise. Have a great time.

Speaker B: Take care.

Speaker C: Bye bye bye.

Speaker B: The superheroes. The superheroes.

Speaker C: Point me in the direction you'll make the right connection.

Speaker B: The superheroes. The superheroes.