World's #1 Banking CMO: How AI Is Changing Billion Dollar Banks Today | with Tim Rutten, CMO @ Backbase

0:00Sam Jacobs: Tim Rutten is the CMO at Backbase, a 2.5 billion euro fintech powering over 120 banks worldwide. And he's one of the rare go-to-market executives getting massive measurable ROI from AI tooling. Workloads are truly either automated, some of them are agentic, you don't even touch the keyboard anymore. So with 25 % less budget, I'm actually still doubling the pipeline. In today's episode, we get Tim to go deep on the AI tooling driving those results. But we don't only focus on what's working. Tim also shares where most organizations are falling short with AI. You don't need Fable. It's way too expensive.

0:33You don't need that level of intelligence. You could actually probably do better with it.

0:37Sam Jacobs: Plus, we discuss the importance of protecting your team's judgment and taste, how to get top-tier results from cheaper models, and our takes on precisely how the AI bubble will pop. Welcome to Topline. Hey, everybody. It's Sam Jacobs. Welcome to Topline. I'm joined by my amazing co-host, Asit Zaman, the CEO of Sales Talent Agency. And AJ Bruno, the CEO of an incredible company in the incentive comp space called Quotapath. And we're joined by our good friend, Pavilion Gold member, and perhaps the most AI-pilled CMO in all of tech. His name is Tim Rutten. He's the CMO of Backbase, the 2.5 billion euro fintech behind the AI-native banking OS, used by over 120 leading banks worldwide.

1:23Sam Jacobs: With 12 years at Backbase across solutions engineering, strategy, and executive leadership, he brings a rare operator's perspective to marketing. He's also one of the few AI go-to-market leaders that's building and running AI-native marketing at scale inside a category-defining technology company. He's an alumni of the Kauffman Global Scholarship, which is something to do with Harvard, Stanford, and MIT. And he holds a BSc and MSc in International Business and Entrepreneurship from Maastricht University. Tim, welcome to TopLine. Thank you. Thank you so much. It's also a pleasure to be here, guys.

2:00Thanks for having me. We're excited to have you. All of those things that Sam just said sound really important. Yeah, you're an important person. It sounds so important.

2:08Sam Jacobs: It's almost unbelievable. Maastricht, I will say this before we dive into the show, Maastricht holds the distinction for being the place where you received your degrees in international business and entrepreneurship, but it's also the place where in 1999, I crossed the border from Belgium into the Netherlands. I immediately went to a coffee shop with my friend Derek. I smoked a tremendous amount of marijuana and then wandered the streets in the bike lane and almost got hit by a number of different aggressive cyclists. I look forward to sending this clip to Rose Down the Line. So that's what people love about Maastricht in my world.

2:50Sam Jacobs: Beautiful. Yeah. You could also refer to the Maastricht Treaty or something of the other nature. Not only is that awesome to tell that, but for those that are on YouTube, and I don't know how we're going to edit this or what we're going to do, but Sam is currently holding his daughter in his hands as he's telling the story. So she's like, what's going on right now? Oh, yeah. Let's jump in. On that note, Tim, we're excited to have you. And the thing that we were thinking about a lot in preparation for this conversation was the space that you're in. Obviously, everybody is thinking about AI. They're thinking about the impact it's having on the world around us.

3:28And there's some clear use cases that are very powerful. Chat, code, support to some extent as well. But for AI to have this huge impact that would justify all the cap head spend and all the other sorts of investments of energy time and focus that's happening, it has to start having a much broader impact than just those three use cases. And that broader impact, a lot of it is eating labor budget. And it's eating labor budget by doing things that labor does. and a lot of those things, a lot of the big chunks of labor budgets sit in companies that are in highly regulated industries. And the nuance about highly regulated industries is that you don't have that much of a tolerance for mistakes.

4:12You know, a 5 % error rate is not very palatable. And so really curious to understand, you guys are an AI forward platform in the banking space. Tell us a little bit about how AI and this world is intersecting right now. What's working, what's not working? How are you dealing with the error rates that seem to be a blotter for the sort of penetration that one would look for? Give us your thoughts. Yeah. Well, maybe first and foremost, just to describe Backbase in a nutshell, what we do. So Backbase has been around for more than two decades already. And it's basically a white label banking technology partner.

4:55And if you think about the past two decades, what happened in banking was that typically every bank needed to go through a digital transformation because they had amassed, you know, tons of different systems that were doing specific things. For instance, originally, you would have your core banking, which is the ledger where you have your accounts or transactions. Then all of a sudden you want to open branches. Great. Previously, those were physical ledgers, like literally those physical ledgers you would put on the shelf and would maintain. They needed a system to handle those. Then there was an ATM, new system.

5:29Then there was mobile, new system. And it just goes on and on and on. That also pertains to introducing new products. Again, new system. Acquiring a new bank, systems collapse almost because it all needs to come together. So that is how Backbase has been, let's say, founded on the premise of what happens if you bring all of those systems together in one single platform and then power all the different channels and all the different actors, which pre-AI was omnichannel banking. It was the best mobile banking application in retail or in small, medium business, all the way up to commercial and private wealth.

6:05Similarly, the web experiences, both for the customers and clients, as well as the internal employees that are, you know, picking up the phone to help and service ASAT in this case. So that's really the, let's say, founding nature of Bankbase, to basically remove all that friction and bring a pretty advanced architecture that would practically turn the bank into a platform type of company. So we've been doing that now for 23 years. We've got 120 banks globally that are on that technology, that are also running on our cloud, completely on our stack. And we practically run those banks from the digital side of the house.

6:38So in a way, we might be, you know, in the top 10 of largest fintechs globally, although we don't talk about it like that because we are not the bank here we are the technology partner but when ai entered the building asad back to your question we were also looking around us thinking okay so probably the form factor is going to change are you really still going to be tapping your iphone screen to do your banking activities or are you eventually going to chat conversational with the bank aren't you going to just have voice type of interactions that you know are going to find the intent of the actual user to then also resolve that intent to actually do the work eventually and deliver whatever type of outcome.

7:21So basically two years ago, we made the full decision. We're going AI native. Sounds great. Nobody knows what that means. I think still many people don't know what that means. I'll try to explain what I think it is perhaps later, but basically it meant we need to evolve our architecture to be able to bring AI to life in the bank, which is highly regulated and is highly scrutinized. Going back to your question, what is the primary objective of the bank and what are they very, very good at? Managing risk. That's what the bank does, right? They're not the most innovative. They're not going to be breaking through with new technologies, new implementations, none of that.

7:59So the first year after ChentGPT launched, silence. No bank was interested. Of course, they were doing a little bit of a pilot here or there, but in reality, there was no real activity. When Agentic entered, let's say one year, one and a half years ago, I think the windows opened up like, hey, this is actually something where we can start doing certain workloads, which we now are automating away with RPA and just classic automation. There's probably going to be a window where LLMs can interpret those intents from customers or intents from our employees and actually resolve the work to an outcome.

8:34And probably you can have multiple agents that do different things in the organization and basically come to an outcome. Would it be fair to say we're far from that right now? Because I think of it as just like, what is the best, unless and until you take a whole data center worth of compute and put it against a particular task, and a reasonable amount of compute has an error rate of three to 5%. So it's like, I would imagine there's probably no real process in a bank that is a meaningful process, mortgage claims and insurance claims and that credit card applications that can justify a three to 5 % error rate today.

9:10And so it's a lot about exciting vision for the future. But today, the true implementation I would imagine is relatively sparse. And then maybe there's, you know, in the engineering team, there's, you know, we're using clock code and there's efficiencies then maybe marketing is doing the ideation. But the true product innovation for the end customer, I would imagine, is relatively early days, really good ideas, but far from like an impact. How wrong or right is that statement? You'd be surprised on how far you can already take the technology in the context of banking. But the starting point, you're absolutely right, is mostly internal, is operations.

9:53It's under the COO because they own the branch ecosystem, the customer care center ecosystem, meaning they have thousands of people on payroll to basically pick up the phone and resolve issues. And they have a typical KPI of resolution time, like how fast from intent to resolution can you close it? And they have a KPI on what's the cost per resolution. the current numbers that are out there already 90 % efficiency gain in the customer care call center meaning you can actually take out 90 % of that cost which you wouldn't apply per se as a technology partner to small credit unions and community banks no no you work with the bigger players who have high volume where that 90 % is hundreds of millions of savings and that is a very clear super direct use case which also is obvious because I think horizontally in other industries, you see that too.

10:47And in banking, that is a prime one. Now, coming back to that error rate, you're correct. And it also tells you one thing, and that is that you don't apply AI all the way. You don't go completely with an LLM to do a full set of workload steps and reason through it and think and whatever. The reality is the actual implementations are almost 99 % and deterministic. And there's this one little percent that actually gets them to gain to reason at the right moment in time, super heavily guard-reeled, like it's almost like kicking the value out of the LLM out of the door to just get one interpretation and hence a good, let's say, acceptable error rate.

11:29Sam, don't you think it's interesting how different this business is to other businesses in the following way, which is that in most cases, you find that technology companies are faster to be able to implement AI in their product for the market. And it's harder to find the AI efficiencies internally for many. Whereas this is almost the opposite, right? It's like your internal efficiencies you're getting first, but you have to be more measured with how you introduce it on the way. It's just an interesting, like how different this world is when you really click into it. Not everybody's having the same experience.

12:04Sam Jacobs: Well, I think it's, You know what's really interesting about it, as I think about it, which is that it's not the banks, as I understand what Tim's saying, they're not building their own. They're not vibe coding their own internal customer support software. They're not vibe coding their own internal operational software. What they have is an application layer that really understands them as customers, which is back base. And that feels like the blueprint for the successful implementation of AI in production. And it's why there's so much value in the application layer versus just in the frontier foundation model layer.

12:44Sam Jacobs: Because you have to know the customer. You know, you have to understand and, you know, you have to know your customer, as they would say in the banking world. And you have to know exactly how to design the guardrails. To Tim's point, you have to know exactly where reasoning, guessing, probabilistic assumptions and conclusions are acceptable. And then where it's like, listen, we don't need an LLM to tell us what the settlement amounts should be on the specific transaction. That is a math problem and math will solve it. But I think it's possibly a template for other industries. It's really an argument for vertical AI or vertical SaaS, but the benefit again of like, the other thing that we've talked about a lot of us is like, well, Backbase is 23 years old.

13:27So it's also, you know, there was this whole discourse

13:31Sam Jacobs: when LLMs and when maybe it was like early last year, when it was all of the great companies are building around the capabilities of the models as opposed to building around the needs of the customers. And that never felt correct to me. It always felt like knowing your customer is, you know, we talk about it all the time here, and we sound like old fuddy-duddies because we're like, listen, it's still about building something that's super valuable and specific to your customer, knowing your customer, caring about your customer, and using AI as one part of the mechanism through which you deliver the solution.

14:07Sam Jacobs: That's what I'm hearing from Tim. And I continue to think that is really the future. AI is a capability of intelligence that needs to be applied to a customer-specific context. I do have a question here, though, because I think I really fully agree with Sam and Tim, though, for you. I was thinking about this as you were talking, and you see, you mentioned 150 banks. And I remember distinctly when I went to Europe, I guess it was 2012, 13, and tapping my credit card for the first time and doing a tap. And that was really innovative and progressive. And we didn't have that in the U.S. I tend to think of European financial services as being more innovative.

14:49And I don't know why that's the case, but I do. And yet with AI, it feels like because of privacy and everything that comes with LLMs, and frankly, maybe just because it's American companies that are doing the innovation on the LLMs, that there's a pushback. So one, And is it fair to say that like European banking systems are actually in this like more customer value service first and they're pushing it? And two, is that, are they leading the charge here as it relates to AI, if that's the case? I would say as much. I do know for a fact that we're in the, let's say, European context, at least 10 years ahead in time.

15:32when you, let's say, average it out compared to American institutions. As in, we've never had checks, let's say, in the past 10, 15 years. I still write checks, by the way, to my cleaners. Every Friday. To who? There you go. In fact, Angela, my wife, was just like, hey, I forgot to put a check on the counter for the cleaners. Can you put a check on it? You guys pay the poor cleaner with a check? Yeah. So that cleaner has to go and deposit the check? And you make it out to cash. Oh my God, that's ridiculous. over here. Why aren't you doing cash transfers? Like, why not? Yeah, this sounds actually cruel.

16:07Yeah, that's a meme. Do you add extra money to it to like make up for the time? It's antiquated and it's beautiful. We were in checks, man. They worked. I would really hate that. Get the checkbook out. That's so crazy. To just continue that point, and there's many more of these examples where that's just the embedded way of working. It's the embedded culture as well that still lives and somehow prevails. But at the same time, there's a few very progressive institutions in the US as well, right? That are definitely innovating like crazy. JP Morgan Chase is, by the way, a great example. Deep Pockets, big top-down CEO mandate, we're going AI native, we're going to really do this.

16:44And they are investing at such a high level and high pace that they are finding the benefit already this year, by the way. It's serious with Jamie Dimon. But then the actual long tail of regional institutions, local institutions, credit unions, community banks, which make up the majority of the amount of banks that are in the region, they are way behind. They are. But then look at the European context. What you find there is a major entrepreneurial wave that stepped into these new charters for new banks. Think Revolut wide in Europe and also entering other domains, and they do it very well and successfully.

17:22In the Netherlands, we have Bank. In the US, I know you have Chime. They actually now stood up, got critical mass are wildly profitable and ai is giving them even more tailwind to continue growing so i think the the time is in their favor right now and people are not that loyal as the typical traditional bank things so as it's related to the ai piece of it though and just privacy and the the way that the european union has been thinking about it and um from our from my point of view just been a little bit slower to move, I guess. So that's not the case as it relates to the banking and financial services industry.

18:02So to actually bring it live to production, customer facing, that is definitely on a different path than what you're seeing right now in the US, because regulation is just a little bit more tight, if not delaying. But at the same time, the FCA in the UK, they just released, I believe it was two weeks ago, the actual guidance on how to apply AI in the context of banking in the UK. That is faster than what the US has been able to deliver in terms of regulation. So at the same time, there's this frontier lab race, there's this model race, there's also the regulation race right now. Who is the first one to call what the framework is within which financial services can operate?

18:38Does the lack of that framework in the US mean that there is more or less friction for them to do it? Because is it that there's more of an open field, be wild, be crazy, forget out, But right now the controls aren't there, whereas Europe has been faster to create the boundary walls within which to play. Or is it that in the US, these companies have been governed by older regulation that is now dictating how they behave in this new realm, and that's causing more friction for them, holding them back, whereas this regulation that's coming out of Europe is actually more progressive? Like, which one of these is it?

19:16so the uk is most progressive unfortunately they're not part of the eu anymore so they have their own framework but they're the ones to be called out here being at that pace i was actually on the phone with colin payne recently he is the fca head and he's like we are innovating with our banks we're doing it together i similarly had the regulator in the us in my podcast recently they're on the same trajectory they're co-creating it um so they're trying to remove the limits while still being true to the actual consumer, right? Regulations, therefore, trust and safety and making sure that those banks are actually, you know, keeping that trust in the total economy.

19:53Because of that collapse, we have a very different issue to look after. But basically they're being constrained by the current regulation, as in that's how they're looking at how progressive they're going to be. When you say collapse, are you, I mean, we've heard all of this talk recently with Mythos and concerns around just AI taking vulnerability, risks that have been inherent in systems. And there was the head of the FCC or the head of Federal Reserve having a call with all the heads of banks. A lot of concern there. Are you talking about that type of collapse or are you referring to something else?

20:27It's part of it because basically the regulatory framework will also dictate how do you keep a secure operation and what do you do if things do go south? How fast can you act and how can you ensure that you remain a trustworthy banking institution? So that's on the Mythos side and basically how secure can we be and how vulnerable are you today? Mythos was the leading subject, but the reality is also other entities, non-American or non-European, if they get the same capability, can we stand our ground, yes or no? It's a very more sovereignty-related type of conversation. When it comes to the question of the sad post, like, okay, so are they now holding back on the innovation?

21:07this is the verbal answer you will get. Yes, because we don't know what the regulation will really do. So I'm not going to invest because if I need to pull back, I'm losing time, I'm losing funding. So I better put it where I can actually deliver. So think banking executive here, how many hundreds of millions are you going to pour in without knowing what the next week will look like? Well, at the same time, the majority can actually not even execute. They're still living in this completely disparate ecosystem of disconnected systems that they've stitched together themselves. I think you all know one thing you need to have is holistic data, holistic intelligence, and then you can drive whatever in a very effective way.

21:44So they're actually very far away from doing it to begin with. And yes, having regulation that is clear would certainly help speed it up. Sam, when we look at this AI build out, obviously we all know that there's bubbly behavior, which is reasonable you've written about this. knowing that you're in a bubble is not rocket science when you're in it like it's fairly clear that doesn't mean you know when it's gonna burst doesn't mean it's a bad thing it's fairly logical but for this thing to not blow apart the investments we're making into ai you have to see this return on the revenue side as well and so as the investments are increasing this return has to increase as well and when you hear about some of the friction that ai is going to have to wrestle through to be able to gain proper substantial revenue out of banking, out of pharma, out of healthcare, like out of these industries where there's so much revenue potential, but there's a lot of complexity to be able to unlock that revenue potential.

22:49Part of it is the AI has to get better with less error rates. Part of it is that we have to have the right frameworks in place for these companies to operate in those areas. Is that the big concern that by the time we figure out how to get to these industries where the revenue sits, we might have overshot on the spend. And this is going to take a lot longer than people might think it is going to take.

23:13Sam Jacobs: That's a great question. Yeah, I think the fundamental concern is what you said, Asid, which is that the build out on the CapEx spend and the revenue are too disconnected temporally, meaning you don't make the money soon enough. You know that you make it one day, but you spent billions and billions of dollars to build the data centers, to buy the GPUs, you know, to invest in the capability. But I think the near term, to the point of all the conversation that's happening in social media right now, I think the biggest risk for the bubble deflating is the inversion of the business models. OpenAI and Anthropic, the labs, Their businesses are built on selling tokens.

24:03Sam Jacobs: The business model of Anthropic is that fable, nay mythos, is expensive, right? That the tokens associated with that compute and the new capabilities are expensive and desirable. And the issue with open-weight models and open-source models like Kimi and anything coming out of China presumably are that the business model of OpenAI and Anthropic are now being called into question. If open-weight models are extremely cheap and therefore the tokens associated with them are extremely cheap, then any equity asset bubble that you have related to the ability for Anthropic and OpenAI to generate massive amounts of ROI on the invested capital of those data centers is now called into question in the short term.

24:53Sam Jacobs: I think that's what's at risk and that's what people are worried about, which is a related but not exactly the same thing as what you were asking. Like if everybody presumes the way that all these companies make money is selling tokens and we can trace the dollars through the token flow and then all of a sudden it turns out selling tokens is a really terrible business, well, that's a thing that would compress, right? The equity markets right now are predicated on selling tokens being a good business. Yeah, that's interesting. I'm sure you see this, Asat or AJ, but if I think about the token economics right now, it's so early still.

25:31It's so early. So we're almost thinking of calling out, well, tokens are a terrible business, which I think you're right, Sam, but we haven't even seen the actual AI transformation across the globe. We've seen a few nice examples and there's a few, let's say, big brands overnight. Think Cursor, one example, overnight, massive success. look at sierra going crazy lately there's just a handful of these examples because they're the ones that grew up with it and they basically built a business that's on that rails that's on that new infrastructure and in a way the way i was thinking about it earlier today is it seems like we're where the frontier labs are competing around electricity eventually right because all these models are getting incredibly intelligent and i can tell you already for banking there's sufficient intelligence.

26:18We don't need more intelligent models. We need smaller models, more efficient models, and we need better infrastructure around these models to have the lowest error rates, to your point, Asad. And we need a clean, almost standardized architecture so that every bank can feel properly governed and properly utilizing the technology. But eventually, to me, it's actually just electricity. And it can also not be held within a continent eventually because of open source because of open weight models. And it reminded me of a paper from a guy called Nicholas Carr, professor. He wrote a paper way back when and stated, IT doesn't matter.

26:55And he made the point that the initial transformational wave of IT and internet was giving massive GDP growth to certain economies and to certain types of businesses, but eventually it would roll off because it would just be like electricity. Everyone has certain capability of IT in their operation. Most likely, and that's maybe the thing to have a discussion on, I believe with AI, it's going to be the same thing. It's actually, it is a commodity the way it's now being made available to the market. Tim, so question for you, how many of your customers have an on-prem part of their business? Do they like, I mean, I'm sure all of them have some level of cloud, but there's probably a lot of them have an on-prem piece, correct?

27:38Practically all. practically all will have a subset of tech internally on bare metal yes so as we think about these open source models and this on-prem do you see a world specifically to financial services and maybe this is true today where all these models are just kept local and are basically on-prem today even is that is that where we're headed yes and no yes because that is logically what the bank will do because they want to contain risk they want to control the unity economic costs. So eventually that will happen. There's a few of these steps already taking place, but even more so, a bank shouldn't choose one model.

28:17That would be silly. They need to be model neutral. So for some workloads, it makes sense to tap the frontier lab with the latest and greatest, highest intelligence. But for the majority of daily workloads, which are, mind you, highly repetitive, they're pretty deterministic because a bank is not reasoning whether they will give you a loan. They know there's just checkpoints and policy. So actually applying AI completely in the capability it has to a bank is actually not smart to do. It needs to be very metered, very well architected, very well guard-reeled, and then it becomes wildly efficient, effective.

28:51But for that, you don't need Fable. Trust me, you really don't at all. It's way too expensive. You don't need that level of intelligence. You could actually probably do better with a small language model, which is specifically trained for financial conversations or financial type of transactions, which is already also happening. I think where I lined up on all of this is that I hear the argument that a lot of the more intelligence that we have is good enough. I think Harvey's CEO was saying that I don't need a smarter model. I can build a many billion dollar business with what we've got today.

29:28And I find that, like, I don't know if I can agree with that. Like these models are good, but they're not that great right now. And the amount of revenue they've unlocked is kind of proof of that. They have like idiosyncrasies. They have problems within them. The highest level of intelligence you can get is good, but it's too expensive. Even the highest level of intelligence you can get with a reasonable cost basis in terms of the compute you give it has an error rate that makes most meaningful work not palatable to put through it. And so really what we have is support, you know, the Sierra type of use case where you're replacing a call center worker.

30:07Like, you know, you pay$10 an hour to someone or you replace them with AI. Like, I don't think the quality bar is that high in that use case. Code is great. Like, it's very, very meaningful. It's probably the most meaningful use case. But beyond code and support, I really don't see, other than personally using chat to, like, give yourself a pep talk. There's not much more it's doing in the industry right now. And so I think it would be great for these models to get smarter and smarter. And I think the smartest model, the hesitation to use it for even a menial task is cost. But if you could have a PhD do a really, you know, low complexity task at the cost of a non-PhD person doing the same task, you'll choose the PhD person because they will be able to do it better, sharper, smarter, with less errors, etc.

30:56in theory, right? It's just that to hire a PhD to work in a call center doesn't make financial sense. Hence, we will find the person that's willing to do that job today. So I think we need better models. They need to become far more intelligent. They need to make less mistakes. We need to learn how to implement them at a cost structure that made sense. I think all of those things need to happen. And the thing that gives me a little bit of the anxiety is that we need it to happen relatively quickly for all these investments. To hit the numbers. sense to, yeah, these investments are crazy. They are crazy.

31:30Alphabet's going to announce their earnings later today. And you just think about the cap et spend that these companies are doing. And revenue doesn't have to match it, but it has to be on a particular curve. And we can't, like that curve, it can't just rely on these three use cases. They've gotten us here. Something else has to get unlocked.

31:51Sam Jacobs: So there's like two counter, counter argument one, you know, point, counterpoint. Point, we don't know how big these markets are. These markets are going to be massively bigger than anything we understand. And we just haven't seen the capabilities yet. So we have to be long AGI. Counterpoint, we might have to be long AGI, but that doesn't mean that Alphabet, Microsoft, and these other businesses are the ones that are going to win. Same way if you bought Yahoo or AltaVista or Lycos in 1999 or 2000 and you didn't buy Google. like you could have been right about the internet and wrong on all of the early leaders where do you fall on that like you sound actually bearish are you are you short to market are you are you have you moved more into cash no no i'm riding micron to the sky right now just holding on it's up 12 today it was down 6 yesterday it's a wild ride um That's how you like to live, if I understand you correctly.

32:53Living on the edge, Sam, living on the edge. I stopped smoking weed, so I need something else to give me like a little rush. That's very sad. Yeah, there you go. I think these things have to, I believe in the long term, long term, long term, right? Just like anyone who looked at the internet would believe in the long term, long term, long term of it all. But I also don't want to sit in 2000, 2001, 2002. I would rather not be in that sort of like desert land for a three-year period. And it's undeniable that there's a risk factor here, that we could land up in that situation, not because this thing isn't real, but because we don't match revenue with investment and something breaks here.

33:37And the thing that I'm thinking about more and more is we can't ride the coattails of code generation support and chat for the next chunk of revenue that's needed for this investment to keep making sense. More use cases have to get unlocked. I can't see those use cases today. I really can't. Like, I can't see meaningful stuff. And that's why I'm like, hmm. Like, something has to give for meaningful, real use cases to keep presenting themselves. I haven't seen anything so far beyond those three things. And I believe the markets will be larger and all of that stuff. I just don't see what the next thing is yet.

34:15And so I'm really hoping that that breakthrough happens or the next use case presents itself. And my worry is that those use cases sit in places where our tolerance for probabilistic outcomes is very low. And that's really the nature of AI today. So it's like... Like banking. Yeah, banking, healthcare. Like banking, yeah. But mind you, if I look at the, as an example, if I translate it to my world, so banks are making investments. Over the past decade, they would have a digital transformation fund and a big theme that they go for for multiple years. And they invest hundreds of millions in digital transformation.

34:50That budget is marginalized by now. And it's actually a replacement budget. And the majority is reallocated to our AI native rewrite. So how as a bank do we rewrite our processes, our org, and everything we basically believe and do on how to operate a financial institution? meaning funding shift. It is happening. I see those dialogues. Now to actually digest it and do it, it's going to take years. So I'm similarly like you, a little bit bearish on timing, massively bullish on the future. It will happen nevertheless, but it's going to be a very rocky ride. Well, Tim, just to actually clarify that point you just made.

35:28So when you say funding and digital transformation, does that mean that there was a corp dev arm to some of these banks that was investing in a more of a VC way to startups and now they're directing that funds to maybe just a fewer players that are these larger players uh open ai or anthropic or whatever that's where the dollars are now it was like the accenture budgets that they had right like everybody was working with extension mckinsey and tauda consulting and they had all this money going to these like digital transformation stuff that was like happening and i'll give you an example you have uh danske bank biggest bank in the Nordics, they have a program from the board.

36:04It's called Forward 2028. That was a four-year program with a certain number of investments, like hundreds of millions, to completely modernize their architecture, infrastructure, and their digital channels. Big strategic push. Up until four years ago, the major headline there was digital transformation, digital banking, omni-channel banking. The headline now is agentic banking, agentic servicing, gen AI, customer care center optimization. It's basically trying to apply AI in the context of banking and the budgets are flowing there. Like that was mostly my point and certainly they will have an arm that makes some investments here or there.

36:46That's not the biggest thing per se. Have those budgets gotten even larger then? Like the digital transformation budgets compared to where the agentic AI budgets are much larger now. Interestingly enough, the bullishness on it is... This is the scary part because it's all this money ages. So like I talk to people at the labs and they're like, yo, we keep getting these calls from the banks and everybody else. And they're like, we have all this money. How should we spend it? Basically, right? Because the spreadsheet checks out if it works. Yes, the AI lab is like, let me tell you how to spend your money.

37:19No problem. So we will come in and we have all these ideas for you. Great. But what if those ideas don't lead to anything that gets into real production, real like move the needle type of stuff. I don't think that YOLO, baby, YOLO. This is the scary part. But is there a financial crisis that's coming from this? Like, is there a reckoning? It has to be, right? Because at some point, if the revenue doesn't keep increasing, then at certain points, you'd be like, oh my god, we've spent all this money and the revenue is this much. Because there has to be the balancing of these two things. And the balancing of these two things has to mean that AI has to go deeper and deeper into the economy and create actual value, not just get money that gets spent, but money that gets spent in a way that creates recurring revenue, which means value gets unlocked.

38:09You know, if you run an experiment at JP Morgan, that basically is like, that's cool, but it doesn't make us more money or save us any money. So it's not really like it's not something we're going to double down on. So what? That's the issue. But that's maybe also the interesting thing in time right now. Those examples are there. Actually, in particular, JPMorgan Chase released their financial statements and it's very serious what they're getting out of the programs. Lloyd's Banking Group, already 50 million last year, tracking to 100, if not 200 by end of this year in terms of actual ROI in millions.

38:44That's not a lot. But if you look at the actual detail it is maybe 0.1 percent of the use case of the actual total operation and to tim's point it's so early we're so early austin yeah yeah yeah they tell the capets when how early we are right like that should maybe we should talk about spacex uh as well then that's not how it works you have to balance these two things out and right now it's a bit crazy this episode of

39:17Sam Jacobs: The top line is brought to you by Compound Growth Marketing. The way your customers buy has shifted before they ever reach a rep. They're asking an LLM who to trust, who to shortlist, and why. Decision-making is impacted by investment at the top of the funnel. And most B2B marketing teams under-invest because it's difficult to measure it. That's the gap, and it's expensive, folks. Y 'all, do you know what impression I was doing? The guy, the Morris, what is his name? Anyway, he's the guy that he selected Graham Plattner as the candidate for Maine, you know, one of those political consultants. Anyway, that has nothing to do with compound growth marketing.

39:58Sam Jacobs: That's amazing. Which is this gap for many of the fastest growing companies. And cybersecurity. They work in cybersecurity, compound growth marketing. Their predictable demand system unites paid media answer engine optimization, and AEO for those not in the know, and GTM engineering into one discipline engineered to produce pipeline. You can actually forecast backed by over$100 million in media manage. No vanity metrics, no guesswork. If you want a free audit on a funnel media, including AEO, reach out at compoundgrowthmarketing.com and mention Pavilion. Reach out at compoundgrowthmarketing.com and mention Pavilion if you'd like a free audit on your funnel.

40:39Sam Jacobs: If you don't do that, Topline will unfortunately go out of business. So please reach out. Sam, when you, I was thinking about this compound growth marketing, you remember the term growth loops and how much that was used as a term? I think so. And it was this idea of like, I will. Yes. So like to me, it's each action you take today makes tomorrow easier, faster and cheaper. So why don't we do that with compound growth marketing? Thank you, everybody. Asit, any closing thoughts? I think AJ should do a diagnostic. He loves SEO. This is SEO 2.0. AJ, please do a diagnostic. It's part of my compound growth marketing strategy as SEO, for sure.

41:19Sam Jacobs: EGM, folks. They've got a predictable demand system and it's great. So thank you, compound growth marketing. Go to compoundgrowthmarketing.com and mention Pavilion to get$5 off your next Happy Meal. Thank you. This episode of Topline is brought to you by Pavilion and a new program called Pavilion Scale. Listen, if you're a go-to-market leader, you already know the job gets lonelier the higher you go. Fewer people you can actually talk to about the stuff that keeps you up at night. Scale is our answer to that. It's built on the same model we use for Pavilion Gold. Small councils of about 10 operators at your level with a coach meeting regularly.

42:03Sam Jacobs: Not another Slack channel you'll mute. Real peers, real problems, real accountability. It's invite only, and the first cohorts are matching in August. Topline listener Ryan Burke is going to be one of the founding members. He's also been a guest multiple times. So listen, if you want to be in a room with Ryan Burke, if you want to be in a room with 8 to 10 people at your level, building companies and go-to-market, if that's the room you've been looking for, go to joinpavilion.com and apply. Also, you can actually message me on Pavilion Slack if you're already a Pavilion member. But tell them wherever you go that Topline sent you.

42:44Sam Jacobs: Jackie Leahy, Jackie Robinson, Charlie Moss, we want to see you in the Pavilion Scale program. The other thing that I think is massive risk is these delays that you keep hearing about in terms of the data center build-outs. And that's also pretty crazy and scary. And so there's like a bunch of stuff that like we're all on like the shaky ground here. Right now, it matters Only, but shaky ground for what? You're on shaking ground if you're trying to retire in the next three years, for sure Right? If you're counting on Shaking ground means that we might be doing business in a very blah market for a couple of years Because any time the bubble bursts, you've got a two, three year period of light Back to 2023 Asit's in Toronto, he's been spending too much time outside The air quality index in Toronto is like 270 You know that day?

43:35I thought it was great. I was like, what nice cool weather outside. I went out. I did like a bunch of deep breaths. I came inside. Sherry's like, do you know what's happening? I was like, I don't know, but I'm very lightheaded right now. You just spent nine cigarettes while you were outside for five minutes, Austin.

43:49Sam Jacobs: I would just say, I think what you're describing will happen for sure. And I think it's totally okay. Yeah. It will need to happen. C 'est la vie, you know. To your point, it's disconnected. But let's look at the positives, Sat. and I'll just stay in my lane when it comes to software development and what is happening in the realm of tech, which by the way, that is where the bubble sits. The rest of the world has no idea what's happening. Like just have a conversation at any birthday. I feel like at this point, it's kind of the cat's out of the bag. People have to know that this is a massive bubble.

44:22They don't. At least if I'm at a family gathering or a birthday or I talk to my friends from high school, they have no idea. They're in the regular economy. can we we're going to come back to these inputs this is actually a really good set way for us because we're going to come back we're going to talk about what you're seeing that's actually the good and the positive and all of this you're going to make me feel good about the world before the end of this i hope so but before we do that we have a surprise for you because i didn't tell you we're going to do this uh but if you've been listening you know we're going to do this which is our favorite segment quiz quiz let's go okay so quiz it's a trivia quiz today again i i mean i'm enjoying this trivia stuff and so i'm gonna tell you guys a little bit about um i'm gonna give you some facts and then i'm gonna give you some options as to who those facts are about and you have to Pick which famous entrepreneur we're talking about.

45:17Cool. This is like out of left field. Oh, this is awesome. I'm excited right there. I'm really excited about this. As a teenager, he and a friend built an illegal device for making free long-distance scrolls and used it to print roll the Vatican. Fact number one. Fact number two. At an early job, he was eating a special diet which was which he thought he didn't have to bathe that often. and his co-workers were quite upset about that because of which he was asked to take the night shift. And number three, other co-workers said that his preferred method for stress relief was soaking his bare feet in the office toilet.

46:01This is a rumor. We're not sure about this one. option number one Vinod Khosla option number two Steve Jobs option number three Steve Wozniak I was going to say Jobs

46:19Sam Jacobs: it's Jobs I mean I think Woz is like the obvious one because you imagine him to be this slovenly like person you know like this big fat slob putting his feet in a toilet

46:37Sam Jacobs: And then the other part of the problem is like, I don't remember, did Steve Jobs work anywhere besides Atari? Oh, did he? He worked very infamously at Atari. I just, it's gotta be Steve Jobs, that bathing story is pretty well-known. It's definitely not Vinod Khosla. I read a lot about jobs, but not, not the specifics, but I don't expect Walsh to do this. So it has to be Jobs. I love how Jobs is expected to be put in his feet in a toilet Why not? He was a wild guy It is Steve Jobs There we go You are correct Let's hear a little bit harder That was an easy one Not too easy The bathing one was easy Okay He was born with an entirely different last name The surname he's famous under came from his adoptive stepfather Number one He's bankrolling the construction of a clock inside a mountain designed to take once a year and run for 10 ,000 years.

47:34And he's personally financed a deep-sea expedition that hauled the Apollo 11 rocket engines up from the bottom of the air. I think this is Bezos. It is Bezos. How do you know this? The clock is definitely Bezos. The clock is Bezos. That's well known. What a stepfather thing. I didn't know the stepfather. If you look at you guys, after months of losing the actual go-to-market stats ones, straight whims here what was his original last name i actually don't know that i don't know but

48:05Sam Jacobs: i know that he's adopt you know his his stepfather he considers his father and uh i guess that guy's from is he from mexico he is a cuban man um yes this expedition was really wild like to go and find the i thought that would make person think that maybe this was a mosque but you guys are well read. Question number three, what a disaster these two ones have been. In middle school, he wired up a homemade messaging network so his family's house were paying his dad's dental office. Alright, Zuck. It's Zuck. How do you know this? I didn't even know this one. What the hell, guys? You guys are so tuned in, Sam and AJ.

48:45Sam Jacobs: So well read. Well, we know that his father was a dentist. Like, that's how you know. I did not know this. I did not know his father was a dentist. What was your other two clues? What were the other ones? He stole an idea from twins. In Prep's Rule, he was the captain of the fencing team. I didn't know that. And at 17, he turned down a million dollars from Microsoft and AOL for music. I knew that. I knew that. Yeah, you knew that? Well done, guys. I didn't know all of those talks. I'm learning as I'm putting these together. I'm learning to... After this disastrous performance on my side, wait till next week so you guys wait um we are back to topic number two tim you were about to tell us of all the good things you're noticing with ai you wouldn't get positive and optimistic about this tell us a little bit of what you had in mind yeah perhaps more on the long side on where i see let's say the day-to-day operation that i look after are changing so as a cmo of backbase i run the marketing org of course highly collaborative with my cro and we basically decided together, let's just take the assumption that this is all true, what we're reading around what it means and entails when you become an AI native operation.

50:02So let's do it for our go-to-market function, like together to bring the sales marketing operation on the same rails or the same infrastructure when we can collaborate together as such. And that is now, I think after a year of work, evolved into a thing we call GTMOS internally, which is basically a Vercel application using the full Vercel stack tapped into our Salesforce, tapped into our finance tools, tapped into all the different, let's say, corporate tools that we have, which hold the data. And the unlock has just been unbelievable in A, how our teams are able to work. They are literally 10x, our AEs, our BDMs, my EBM managers, my product marketeers, all roles are equally impacted.

50:47Workloads are truly either automated, some of them are agentic, You don't even touch the keyboard anymore. So the decision I made was, if we're going AI native, I better be it myself, right? So that was a trigger. And actually deeply stepping into it and deeply doing the work, I can see what the future looks like. And we're making that transition. Is any of that making you more money or saving you existing spend? Or is this net new? Or is this net new benefits, advantages, et cetera? Like we're new capabilities that we have. They're new things we're doing that we could never do before. An example is we had Gaurav from COO and he's like, we can now put together 500 versions of some campaign that we just couldn't humanly do before.

51:35That's like this net new thing you can do. But it's not necessarily saving you labor budget today or necessarily making you all that much money either today. it's just this new capability which is super interesting right like we're developing new capabilities you can couldn't do the split now you can do the split new capability that's cool where does it fall in those three buckets so in the way we have approached it and what we've also executed on double the pipeline 25 % head cut on my budget still so with 25 % less budget I'm actually still doubling the pipeline ooh I think that is interesting.

52:19And in one year I can tell you. That's not nothing. No, I mean, Tim's been talking about the ROI. I'm curious as to how he's done that. Like how, what have you actually seen? Is that just an uptick in keywords that you're seeing as from a demand and SEO? Like what is actually driving that? Yeah, I love your example. It's the actual, well, it's not SEO or geo, although that helps and we do that work. Double pipeline, are you doing more keywords?

52:48i'm selling to boomers they're still googling like uh whatever there's plenty of like plenty of the demand side on inbound is still coming from google like that's still a really big channel don't hate on pain and actually don't get me wrong uh my type of buyer is a little bit more senior in their career and they do still start on google but more of the traffic is flowing to llms but nevertheless, I think the real, let's say the number one thing I would share, Asad, is the actual operation runs on GTMOS. So the way I would like to work and the framework we have to run the business is running on GTMOS.

53:23So there's no other way to work. So the way AEs, BDMs, and my ABM managers work together, the system dictates it. It's not that they have a different taste of the day or culturally, they do different things. They don't do outbound. They don't do events. they don't do whatever. No, no, no. They all run as one. They run the same. Meaning, my interaction and execution against the market is just 10x better than ever. And that changes the output. It's still the same vision we're executing, but we're executing it in like an engine that just doesn't take no for an answer. about this. Because you've got definitely an AI-filled team then, right?

53:56Like, you're doing this stuff. You're at the edge of like bringing it into the function. I'm having this experience where I'm interacting with people that are using AI. And I'm seeing what they come up with when they use AI. And I'm really judging them a lot. Like it's changing my perspective on that person in a very like meaningful manner, right? Because I feel like, I don't know, I have so many more tells about people now with their use of AI. Give us some examples. This is very interesting. Tell us. Let me give you an example. So if you use Claude to create any sort of artifact, right, like a PDF or a, it has these like things it loves to do.

54:39It will always like put in very like light gray font. If you're doing a deck underneath confidential made by this date, this, it's just these like things that always does. and if you actually pay attention and care you would be like this is a very like ai thing and so let me just get rid of that stuff let me make it feel less like aie and if you don't do that i'm like did you not catch it like where's your judgment where's your taste like where's all of that and so i'm being very judgy like i'm just being honest like i'm being the judgiest i've ever been, but I feel like I have more things to judge.

55:19And I have a really tough time now when I see it to not see it anymore. I'm like, how are you dealing with that? Because in an AI-pilled team, I imagine you're seeing so much of this stuff.

55:31Sam Jacobs: Especially because I think GTMOS creates enablement materials and creates briefing reports. So to Asit's point, if you've got a group of a large group of people that aren't really, you know, nobody's really a reader anymore. Most people aren't writers they might just be accepting slop as like the natural output and thinking that you know thinking that it smells like roses when it smells like poo poo yeah because it's just better than the shit that they made before and it's impressive right and it's fast and you get mind blown and whatever but it all looks the same feels the same it's actually not great so that that was actually the full thesis behind having a gtm os which you literally log into and the system itself dictates how it's delivering output, be it work, be it PDFs, be it plans, be it reports, what have you.

56:18Funnily enough, GTM OS on its own, 90 % is deterministic. We're actually not reasoning across all this stuff all the time. We only use it for certain moments in time. If we do account research, you leverage EXA, you combine it with Firecrawl, then you get an output. There's a specific flow where the output gets massaged the back-based way, so we find the five strategic projects described in their words and basically every workflow has its own little boundary in which it operates and which it cannot step out of so that the quality is high then when it comes to content generation same thing you have tone of voice you have guard rails that we don't have ai slob there's like 22 rules that are applied if they don't check out the agent needs to rewrite and continuously learns to rewrite so that the we only do this for long tail content, not for primary high value content, but long tail, you can actually generate it.

57:10And trust me, it is really, really strong. It's actually the same as in banking that you don't want to have these error rates. You have to contain it. You need a whole, let's say, architecture around it to get the quality out. And yeah, funnily enough, if you actually use GTM OS every day, it's actually just a very headless way to interact with your data, with your Salesforce, with your opportunities. I think that's really cool. Like I really do like the interface of it all. Like, I like that you have more signal. I just worry, like, you know, I know of companies that have made some real mistakes when it comes to planning with over-relying on AI from analysis, committing things to boards, you know, that actually shouldn't have been committed, but it's the way it's.

57:55My biggest thing is actually, I worry that right now, a lot of people that didn't have a still are using AI for that thing. And if you didn't have a still in an area or judgment or taste in that particular area, you have no mechanism to stress test the quality of what is coming out. But you're just somehow really excited, this psychology of like, if you couldn't make decks easily, but now you can just create decks, no problem. If you couldn't write a memo easily, but now you can write five memos in five hours, I think there's this like excitement or mean rush of being able to do that thing, but you don't have the mechanism to stress test your work.

58:36And so you're putting out things into the world that is very bad. It's very bad work. LinkedIn is a perfect example of this, by the way. Like you just have to stroll that newsfeed and be like, oh my God, what's going on here? tough but asid question on you like who's giving feedback to everyone of these people are you calling when you read this whether it's your team or externally do you call it out uh immediately hey i don't it's not worth the time like internally of course you know you think of guard drills and all of that stuff right like because it's your company it's generally like do i care about so many people that much like who am i like i'm not jesus walking around trying to made everybody great at everything.

59:19Like if I'm if it's you who does it, I'll tell you. But like, I'm not like some police force that's growing around LinkedIn being like bad, just like judging it from afar. I'm like, what an idiot. And I think there's going to be a lot of that. Like, I think this is on individuals, like, I can tell my friends, like, I'll come to you if I see you do something. I when Sam was using Clotbot, like we were like, Sam, that's bad. Don't do that. He was like, I adjusted. Sam will tell us when we're doing it, but how much of this responsibility for the rest of the world are we all limitating? Well, to your point around the skills aspect of it, we're not teaching, we're no longer teaching people how to work at a level that's acceptable or adequate.

1:00:05We're taking what they give us AI slop and we're feeding that to Claude and saying, how do I respond to this? And feeding it right back to them. I mean, that's a very huge problem with what we have going on. Yeah. Maybe the other way to look at it is this is also just a moment in time. And right now your LinkedIn channels are flooded with this crap and people are annoyed by it, right? They are really actively anti-AI. It is happening. It's going to take another six months. It's less popular than like some of the worst criminals in the world and like all of that stuff. Like, yeah, this is true.

1:00:40I didn't read that magazine, but that sounds fair. I think it was like, yeah, it's a thing. Like the popularity levels are like really, really bad. Yeah, so anti-AI is now a thing. And on some channels, it's really the thing. I think especially also in GTM and marketing, this is now the topic. But also this moment in time, in six months, algorithms has changed. And actually the game has renewed its vows on how to either distribute content. And Substack right now is massively booming because people are moving off of LinkedIn. Like things are just moving around to new spaces and eventually there's going to be something new.

1:01:16I don't know what that new thing looks like, but I'm sure that Sam is going to be there and it's going to be wildly successful. Thank you, Tim.

1:01:22Sam Jacobs: Just Sam. The parameters are just going to change. Because he does good content. I think it's like a very multi-layered, I'm trying to think of what I think about what we're saying and I don't even know. because if everybody becomes a more like a proficient but uninspired writer, is that, like Asad, I agree with you that everything is slop, but maybe you and me are the only people that care. You know, I'm trying to think of what are the right metaphors, you know, and I can't, one of them is Ikea. It's like, well, is the existence of Ikea a bad thing or a good thing? I'm like, I think it's a good thing because now lots of people can have relatively nice furniture, even if it's all completely, you know, mass produced.

1:02:08Sam Jacobs: It's like, does that mean that when I bring in my bookshelf that's handmade and if there's a carpenter in there, I'm like, oh, that's really beautiful and that's a piece of shit. But I don't know that that means that having lots and lots of people that can express ideas a little bit better is a bad thing. Maybe it's more like fashion or like trends. Like, so if we look at jeans, right, They go through these like every few years, you go from like skinny jeans to white lead to boot cut to like skinny again. Like, you know, the style keeps evolving. And I think it starts with at the edge, the cool people start doing this.

1:02:44Then everybody starts wearing really baggy jeans. And so the cool people are like, screw that. And they get like really skinny stuff again. And maybe it's a little bit more like that where it just changes where the advantage is and what you need to do to have an edge. I mean this is why I think the point the one point not that there's only one but I do think that this idea I think it accentuates talent not necessarily levels talent I think I listen I use obviously

1:03:13Sam Jacobs: I use AI to write my newsletter I'm just really good at it you know and I know how to reframe certain now there's still probably sentences where if you put it through an AI grader like well AI wrote that sentence but I understand the game and I can't do it. dance with the AI so that instead of an M dash, I might use a colon AI, at least my, you know, chat GPT thing that I used to write with likes, uh, fours, like it likes examples of four things. So I make it examples of two things, you know, like I have the sense to know how to play with it, but that's because I'm a decent writer beforehand. And I think that to your point, that's like how do you h &m and zara emerge from this world to your point where it's like fashion is ubiquitous but now there's still a talent to be had and how you use fast fashion and low-cost materials to make something stylish so that people actually can't tell what you know all of these women like what's on it's it's an it's a hermes handbag it's a birkin with a bunch of stuff that It costs literally$20.

1:04:20Sam Jacobs: That comes from Uniqlo, H &M, and Zorro. Yeah. Look, as the guy that still writes checks and cares about SEO. Yeah.

1:04:32I like, I'm not long for this world apparently, but I like, Sam, I totally agree with you. If you have that foundational, if you have the skillset to begin with, AI is just enhancing that skillset. If you don't have it, which lots and lots of people don't, then it's really easily noticeable and it's annoying and frustrating to view. I wonder what the still set is. Part of me thinks it's my insecurity. Like I probably care too much about what the world thinks. And so I think more about how will the world receive this. And it probably comes from a deeply insecure place, right? Like, I don't know if it's a net positive thing.

1:05:08Like it's probably better if I can just throw it out and not care what anyone things. But yeah, I don't know what the still is even. Tim, I think the question I have for you is, as you sit in your seat today, you've done what many a few years ago said was impossible, which was to make a 20 plus year old organization feel AI native internally and externally in terms of product as well as in terms of internal use cases of it, specifically in good market. That's really just cool. And it's cool to see it in Europe, right? Where everybody is like, Europe's so far behind. It's like, actually, you guys are really, really at the edge of what you guys are doing.

1:05:49I think it's very commendable. When you are sitting in the seat you're in and you look out towards the end of next year, what are you excited for being possible that you're like, I think I'll be able to do this in the next year that I can't do right now. But if I just give this thing like a marginal rate of improvement. I'm going to be able to do all these other things and how exciting. Like what comes to mind when you hear that? In the context of what we're discussing, like how far can AI really help you to, you know, drive operational excellence or to drive end-to-end delivery? I'm going to get to Christmas with real-time marketing, meaning I'm currently piloting a way with the team to have a pipeline approach to how we bring strategy to execution.

1:06:36sounds very vague. So let me give you a specific example. If you take all the ingredients of how you build a brand that goes to market with a value proposition and a tone of voice and everything, you basically have strategy, you have context of the company, you have value proposition, you have buyer personas, you have their problems, use cases, wedges, what have you. All of that context we have down to the T because that's what informs GTMO is when it runs. It consumes that context. If you then take that context through a pipeline of stages and gates, I'm certainly going to land towards the end of the year that I can literally spin up new pages and value propositions within an hour that land on my backbase.com landing zones, look and feel incredible, have interactive demos right off the bat, and are enterprise-grade.

1:07:25So they're fully secure they can convert there's utm text there's everything next step i can also easily generate abm campaigns for specific accounts for specific persona so that they land with those accounts um what i'm trying to say is that is impossible today like humanly that's just you need to scale up to way too many headcounts which i cannot really budget for so that's what the tech is going to bring it's going to bring pipeline you can see that you're like it doesn't require like a continual learning breakthrough to get there. This is just like, as this thing gets marginally better, I should be able to do that.

1:08:00For me, this is the architecture of the infrastructure. And it's just rethinking how do we actually deliver work and where do we tune narrative? Where do we tune structure? Do we have, do you have an engineer on your team? Yeah, we do. Okay. We do. Interesting. But those ideas, these ideas, for instance, this inspiration comes from Vercel because Vercel and V0 as an example is nothing more. It's a lot, don't get me wrong, but it's a very specific harness. It's a specific architecture with specific SDKs that Claude just understands and can build within. And hence, you can build applications within five minutes that are pretty much production grade.

1:08:40Just apply that to enterprise marketing websites. Have a look at Ploy. Ploy can consume your website and basically generate pages like nobody else today. Look, Ece, your world of SEO is getting more advanced. There's things happening there. So that is one. So real-time marketing through a pipeline approach while still having the craft at heart. And number two, there's going to be autonomous workloads. It is going to happen. So there will be roles that don't have a human interface per se anymore, but they're just running all the time and they're getting better all the time. and think about research, think about intelligence, think about going through all of our conversations with prospects, what is resonating, what isn't, what is clear, what isn't.

1:09:25Like those types of learning loops will actually come to life and inform company strategy. This has been great. Fantastic. On that note, we're going to have you back. We're going to have you share at the end of the year what's happening, what's not happening. But we have our last segment now. Bulls versus bears. Bulls versus bears. We're going to ask you guys questions and you're going to tell me if you are bullish or bearish about that thing. We're going to start with Sam so that Tim, you can get a bit of a flavor of what we're doing over here. Sam, are you bullish or bearish about LinkedIn as a social media content platform?

1:10:03Have we talked about this one? I don't know, but I've been thinking about this a lot. We're chatting about this. I could not be more bearish. Tell us more. Because you're like one of the biggest profiles on LinkedIn. Like this is something you've invested so much time into. You've gained so much momentum.

1:10:18Sam Jacobs: It must be because we were texting about this that I feel like we've already talked about it. I feel one mentor once said, your org chart is your destiny. You know, other people might say your balance sheet is your destiny. This LinkedIn is clearly a company that doesn't know what the hell it's doing. And there are many, many departments and many, many senior people. some of them from the media industry. I've been in meetings with some of them. They all have different projects and different strategic priorities. There's not a clear consensus on valuing the user or the content creator, which to my understanding is what makes LinkedIn valuable for the most part beyond the resume database.

1:10:57Sam Jacobs: So they just, unless they do a massive rework, but I don't see them being able to do that, I can't - Who's the CEO of LinkedIn right now? I don't know. Would it take Reid Hoffman coming back? or Jeff Wiener. Yeah. One of them. We know was driven. He was driven. Jeff was very good. I just, I think that they don't know what they're doing and the consequences, the algorithm is always changing. You know, if you want to use it for, listen, for me, it's the best and it's been an incredibly valuable lead generation and business development tool, but now I don't have reach. I can't get to the people that are supposed to follow me.

1:11:31Sam Jacobs: I'm not sure exactly how. Am I writing as consistently as I was a couple of years ago? No, but I write every day. hey, I'm trying to put something out there every day, and I'm reaching a tiny, tiny fraction of the people that are ostensibly my followers. And I compare that with what we're seeing on X. And what you see on X, and why is it an org chart thing? Because there is more clarity around who gets to make decisions, I feel, in an Elon-led organization where there's just so much, there's less politics, it's just numbers and data-driven. Nikita's in charge, and he's making decisions. And so you can see that people are gravitating, at least I am.

1:12:11Sam Jacobs: You know, I was the person that was like, I don't know about X. It's so noisy. It's, you know, it's racist and mean. Now it's where I go because it's where the smartest people in the world are debating the most interesting ideas. It's not really a BD tool for me yet. But if I want to be part, which I do, you know, if we, if all four of us, which I think we all do, we want to be part of the global conversation about what's happening in tech and in AI, it's not going to happen on LinkedIn. You know, Dean Ball had this thing because now he works for OpenAI and he wrote this thing about how anybody that wants open weight models is decelerationist.

1:12:43Sam Jacobs: And it's in the context of this new Kimi model that I guess, you know, the Chinese company released. And there's this whole massive debate around the value of open source versus closed source and what it means for the frontier labs. It's really interesting. And then, you know, the other active conversation is the value, you know, how good was the Odyssey? I tend to think it was amazing, but lots of people didn't like it. but it was interesting conversation. None of that is on LinkedIn. Zero of that is on LinkedIn. So I am a huge bear on LinkedIn and I'm spending a lot more time on X and a lot more time on Substack.

1:13:18Sam Jacobs: You know, as marketers, like I had a call with the guy that does Adam Robinson's cold emails because I'm like, I need different channels. Like this is not working for DemandGen anymore. So as a marketer, I'm looking for different channels off LinkedIn as well. like basically in every which way it has just gone to shit like it's not driving pipeline you don't want to be there for content you're not learning anything that I always find it like horrible that the b2b social network is where the least interesting business conversations are happening that's like it's it's just so jarring I agree with you so that's my I am a bear tim right end with a joke tim are you bullish or bearish about the form factor of a phone and a banking app on that phone being how i will be banking two years from now two years i'm gonna be bullish still for now um i don't think we are ready as a society but also not from the banking domain per se to completely drop that form factor.

1:14:26I think there's going to be an early adopter group on a new thing. I don't know what that thing is yet. I mean, we were hoping on the OpenAI, John Ivy relationship to deliver something already. I think that's that little device that you can click buttons onto. Right now there's a lot of court cases. Yeah. So I think for the banking apps themselves and the interactions that we are currently used to with our iPhones and our Android phones, it will certainly be challenged I think it will be disrupted. The form factor is maybe not even an actual physical device. It might be even more wild, more sci-fi is my current take.

1:15:02So I don't see the, let's say, most relevant population to the bankers, which are the more senior audience, like AJ here in the room, of course. Still writing checks. I don't see them changing their devices yet until there's an incredibly next best alternative which it feels very it's not there I don't see any signal just yet so for now I'm still bullish on it most likely it first changes on the typical form factor we have today becomes more conversational more voice driven maybe the frontier labs find their distribution there initially which OpenAI is clearly going for and Anthropical probably fast follow that you basically interact with your GPT which is connected to your finances through Plaid, well, that is just an easy warm-up to potentially the new form factor.

1:15:53But two years, it's not fully changed yet. That's a very interesting take. And I would tend to agree with that. AJ, we're going to end with you. Are you bullish or bearish about the idea of spiffs in sales teams moving forward? it has been a thing that's almost like a standard we always use spiffs to like drive short-term behavior etc i've always thought like not sure how useful they are um how good they are are you bullish or bearish will we just always have them so much is changing we so here's what i'll say i'm i am bullish on it's related to this uh i am bullish on my own i'm changing the question sort of.

1:16:40I'll choose my own question. I reject your question, Austin, and offer my own. I am bullish on companies. So typically in like a financial planning period, you start in August, which is coming up this year, and then you get board approval. Sam, when did you get board approval for your financial plan this year? I was like March. Yeah, that's the typical thing. You laugh, but that's pretty typical for companies January to March timeframe to get actual, mine was January. I am very bullish on most companies using AI to be able to speed up that process that whole financial planning. I completely disagree.

1:17:26Okay, well that wasn't, it was my question to myself. Hey, I'm going to debate you on that. Okay, save more. Here's how it's related to it. spiffs can be a really really really great test experiment on how you're incentivizing the behaviors of your team where we have failed is truly understanding the data and the analysis on what those behaviors are so that we can actually implement spiffs to a level that actually makes sense within the financial planning of the organization given that i believe that financial planning will be sped up this year, I believe that smiffs and incentives will actually go into a much more calculated way to change comp plans and incentivize the behaviors that actually matter.

1:18:16And so I'm bullish on it because I'm bullish on companies being more on top of their shit when it comes to financial planning. I think the chance that this ecosystem gets better planning this year is 0%. You should have asked me that question. I don't think there's any company that was late last year that will be like meaningfully faster this year, which is I think those delays are down. Yeah, when are you going to deliver your financial plan this year? When will you actually get signed off for next year?

1:18:43Sam Jacobs: I don't know that there will be a financial plan for next year to be completely honest.

1:18:50Sam Jacobs: It'll be like, it's like so, my business is changing so much it's like a series of like milestone-based scenarios. You know, it's like, if this goes well, this is what it'll look like. If this goes well, this, we're not in steady state. You're making those scenario models so much faster than you ever have before. That's for sure. I mean, I would like, but I just think it's an always-on kind of thing. I think the idea of like, there's an annual planning process for a company of my size, you know, we're just doing 150 mils, so not that big. Just kidding. Just a drop in the ocean. Super small. No, but seriously, I think we should move to like a six-month planning process.

1:19:32Sam Jacobs: And it's really like, this is the zone of probability. But again, I have like a less involved board at this point. We're quarterly. And so to that point in that world of how comp plans interact with it, on a quarterly or even six-month basis, you're going to be able to make changes to those plans. For sure. And therefore, SPFs will be valuable. Yeah, but what I'm challenging you on is that I don't think that the yearly process is downstream of the speed of analysis. It is just humans coming to agreement and differing opinions and board friction and all those things are all the same, specifically in venture funded businesses.

1:20:10I think maybe in the rest of the world, it'll be different. But that's my take. I think we've done this enough, Sam. I think we've had a long episode. Neil has been going to have a fun time. It's been a great timed episode we've ever had, actually. But we're going to catch it. We're going to make it efficient. It's going to be beautiful. We're definitely going to, unfortunately,

1:20:25Sam Jacobs: if you're out there listening and you want to hop over to YouTube, you will not see a picture of my daughter because I will tell Neil to edit it out. You will not be visible. We get like a baby emoji face. Actually, I would be willing to do that, but you can't see, you know, you got to visit me in person if you want to see her beautiful face. Folks, would you smash? would you give it a smash the subscribe button would you give us five stars wherever you get your podcasts would you join our Slack community the Topline community and be an active participant Tim will you be an active participant in our Slack community for Topline yes sir do you promise I think he promised are you a man of your words I am a man of my words absolutely I'll be there thanks for being our guest it's awesome to see what you built and keep going and I'm glad that the band's back together folks we'll talk to you next week thanks Tim Bye, everyone.

1:21:15Sam Jacobs: Thanks for having me.

1:21:22If you are looking for more Topline, check out the Topline newsletter at toplinemedia.substack.com. That's toplinemedia.substack.com.

1:21:36Sam Jacobs: Or if you're a video person, check us out on YouTube. Topline-media is what you want there. Have a delightful day. everybody