Stop Competing on Features: The Real Way to Win in AI Agents | Rob Kaminski, Co-Founder @ Fletch

They come in, they actually kind of know their stuff. They're oozing with insights. They help us shape and frame all these different strategies, but then they don't choose. And that's just bad leadership. The reason why founders and executives get paid the big bucks isn't necessarily for their, like, in-the-weeds smarts. It's because they have to make these tough choices that require trade-offs.

There are dozens of, or more than dozens of AI agent companies coming up right now. Unless they're domain-specific, they all sound and read Very similar, very similar words on the landing page, very similar claims and same category jargons.

If you go to someone who knows what shoes are and you're trying to like pitch them, hey, I've got this thing, it's going to make walking better, it'll make your feet feel comfortable. They'll be like, wait, it's a shoe, right? You're better off just saying, hey, I've got this badass shoe. You're going to love it. Let me tell you why. If you have landing pages and social posts about everything under the sun, all these different segments calling yourselves all different things, The AI won't know what to do with you. So like, I think back to a founder I worked with, phrased it beautifully. He said, you know, Rob, all 6 of these suck and it's not your fault. They all just suck because they all have trade-offs. And it's sort of this whole exercise is like, choose your heart. We always joke with our founders, if you wanted certainty, you should have been an accountant, not a founder.

Today I have Rob Kaminsky. Rob has positioned more than 600 B2B startups. including Notion, GitLab, and Superhuman. So today he will teach us why a great AI product isn't enough to win and what founders should actually do about it. Welcome, Rob.

Thank you. Appreciate you having me.

Can you share a version of you that AI search cannot find or I cannot find on LinkedIn?

Absolutely. I live down in Austin, Texas with my wife Courtney and our cat Mia. We're really into the outdoors and fitness and we love to travel. So we're trying to pick out fun little places to go every year. Grew up in Cleveland, Ohio. I actually started as a biomedical engineer of all things and took a very meandering path into product and sales and eventually product marketing, where I now work specifically on strategic positioning for B2B software companies as a part of Fletch that I co-founded with my business partner, I'm so excited with this mini workshop today.

I'm so lucky to have Rob here to help me think about, you know, how to position one of the companies that I'm working with, actually. So Rob, I'm seeing, you know, this seems to be trending. What I'm seeing is there are dozens of, or more than dozens of AI agents companies coming up right now. To be honest, I think unless they're domain-specific, They all sound and read very similar, you know, very similar words on the landing page, probably very similar claims and same category jargons. Everybody's claiming leveraging some sort of AI agents. So today I brought you one I think is actually doing something, you know, innovative, and maybe you can help us to figure out on camera how you would help them stand out.

Absolutely.

I will try my best to explain what it does. The idea is called Agents Flow. It's by my friend Nicole. So the idea, it's a coordination layer that sits above agent frameworks like LangChain, LlamaIndex. You know, it's using a learned coordination to make multiple multi-agent systems more reliable, auditable, And much cheaper. The founder's analogy is this. So she's saying that, you know, this is the central nervous system for multi-agent system. And she calls it like the octopus. So 9 brains and 1 central and 8 semi-autonomous. I want to ask you how to start about thinking about this.

This is a great transition. I'm going to share my screen to kind of take you through this. So what I've done is I've basically taken what you've sent over, which was a kind of a version of the elevator pitch you just gave me. Of what this thing is, what it might be used for. And I'm, I put it through the methodology and the framework that we've developed with 600 companies have gone through a version of this positioning audit process. And so I'm gonna take you through that and then this'll open up into like, how can we position this thing? Because what we see, especially with new and novel products and technologies, is there's a lot of different ways you could approach doing this. And all of this comes with different trade-offs. And so the goal is to get to clarity of what those options are. So that you and the founder you're working with can actually commit to a decision to start building out marketing. And often it has an influence on the product development itself. And so for those of you not familiar with Fletch, this is all that we do. We're a very focused product marketing consultancy that just does positioning. And our whole aim is to identify the positioning strategy, pick it, and then translate that strategy into the homepage. As Angelina talked about, we've worked with lots of different companies and lots of different verticals. And what's cool about this process is like, we're not market experts in any given space. Granted, we have worked with probably hundreds of AI companies in different areas. And so we're sort of dangerous, but all of this is going to be predicated on your insights and the insights of your founder when we actually go and make some choices. And so I'd like to start kind of back to basics. Most people have a decent understanding of this. When we think about positioning at Fletch, we think of it as the strategic decisions of where we're gonna be memorable. And so most people think about, oh, positioning, that's where you live in the customer's mind. That's the output of good positioning. And so the act is us making choices about where we can live in their mind. And when we think through this and the way we developed our framework is really working back from the customer's point of view, back to that definition. And what we discover is that people think and really organize all information into categories. It helps. It's kind of just how our brains work. We put things into buckets so that they're easily recognizable. And so you think about some of the basic product categories out there: chair, car, shoe, phone. We do this so that we can remember the information. And then the next level of this is then identifying, well, hey, who are the leaders in any of these buckets and categories that I could actually hold in my brain? And so with that in mind, here's the downside for anybody building cutting-edge technology software products is that often your product or your technology that you're building can fit into multiple buckets. And so if we look at, to an example that everybody knows in the iPhone, they could fit into the phone category. They could also fit into the computer category. And so this was a choice that they had early on in their marketing of like, hey, what are we gonna call this thing? How are we gonna explain it? And it has huge implications on a lot of different things. Number one, it'll likely dictate where you're gonna go to market, the channels where these different products show up and where you're gonna end up getting compared to your competition. And so it dictates the channel, it dictates the competition. If you call yourself a phone, you're competing with phones. If you call yourself a computer, you're competing with computers. This will then influence the differentiation. Just like, why you? Why are you better? And so on the smartphone side, hey, we're smarter, they're dumb. On the computer side, we're small, right? It fits in your pocket, they're huge, you can't carry them around. Even your laptop only fits in your backpack. Other things to consider through this category lens is that categories are different sizes. And so we might look at things like market size. Is it growing or shrinking? Which, by the way, categories are all doing this at different rates. This stuff's pretty widely available now, especially with like AI research to go figure out, is this category big or small and is it growing? Other considerations that some categories are, we'll call them real, where if you were to ask somebody or even ChatGPT here, who are the top 5 CRMs? They'll give you a list of the top 5 CRMs. Some categories are not real. Who are the top 5 revenue platforms? Even AI is like, wait, what the hell is a revenue platform? That can mean a lot of different things. It's a signal that there's a lack of maturity over that category, right? So this, at least not yet, speaks to the maturity arc. Another kind of layering in on this is some categories stick around, right? They haven't changed in years other than the underlying technology, but people still use them as heuristics to explain certain products that are in these categories. But some do not. PBX, on-premise telephone systems, no longer a thing. PDAs, nobody says that phrase anymore. Enterprise fax servers, 3D televisions, right? These were things that were real categories that were emerging and have since gone nonexistent. And so we have to consider these when we're thinking about, oh my God, what category am I going to put myself in? Other elements of categories that we need to consider, and this really brings us into our model a little bit, is this idea of, we call it maturity mixes. And this comes back to the awareness level of, do people, have they heard of the name of the category? Do they understand it? Do they know who the competitors are? And so when we look through a category lens, what we've discovered is you can conceptually break up a market category into 2 groups. And we're now thinking through the customer's lens of, are they aware of the category or are they not aware of it? And if they're not, the only reason they would even conceptually be in the total market of the category is that they're in a situation or have a job where the product or solution may actually fit their situation, but they're just not aware of it. And so when we look through this lens, looking at a given category, many of us have seen these technology adoption curves, and so we'll borrow from that. This is what we would say a mature category looks like, where the majority of people in the TAM for the category have heard of it and they're aware of it. And these are things that we buy every day, right? Mobile phones, gym shoes, email, where when you say that, people can conceptualize what it is. They could probably tell you who the leaders are and who they might be choosing between them, and they could start to get into things like attributes. They're aware of the communities. That's all signals of a mature category. An emerging category would look something like this, where the majority of folks who fit into that TAM are not aware of it, so you can't just say the name of it. These are things like personal knowledge graph builders, at-home neural stimulation devices, and synthetic user simulation platforms. These are all real categories that are starting to form, but to the uninitiated, they'd be like, wait, what's a synthetic user simulation platform? Which arguably anybody that's doing customer research is in the TAM of that type of tool, but they're just an immature audience of that category. And so strategy 1, and these are the lenses that we're going to walk through for Agents Flow as well, is to position in a category for the buyers who are aware of the category. We think of this a lot as stealing mindshare and market share using the anchor of the category as the main heuristic. And so this audience is very familiar with the primary job to be done of the category, i.e., like what would I use it for? They usually can name at least one vendor in the category. They have a sense of how to evaluate those vendors. They have an understanding of the lingo and they've probably even taken part or seen some of the communities and conferences that tend to support these very mature categories. And so from a positioning standpoint, what you do in this case is you call out the category name and you basically explain why you're better, right? This is the most straightforward path I would say that most people understand from a positioning front. And so through an example lens, we basically say, hey, here's the web browser category. Nearly everyone has heard of web browsers. Granted, there's some people that still don't really use the internet. So theoretically, there's some immature audience out there. And from a positioning standpoint, DuckDuckGo calls out the category, calls out the others in the category that they're competing with, and says why they're different, why they're better. Unlike Chrome and other browsers, we don't track you. And we love this example because they're saying something that Chrome will never say. We call this binary differentiation, where Chrome will always track you, baked into their business model. And so this is like pretty powerful from a standing out standpoint. Another example of this strategy is Figma. Dylan Fields was really kind of loud about this when they launched Figma, where he just called out his competitor. Where the audience they were going to, that audience was already very familiar with interface design tools. And so rather than try and say, hey, I'm creating this whole new thing that you'll use it to, you know, build UIs, UXs, all he had to do was say interface design tool. And they're like, oh yeah, like Adobe that I'm using right now. Where they just say, hey, we think Adobe has some limitations to it, primarily that it's single player, and so they came in with this collaborative angle to differentiate to this very aware audience. Benefits of this approach is effectively that there's budget line item for this. You don't have to educate people on why they need these tools. The market is already mature enough to know, oh yeah, I need to buy these types of things. But the downside is kind of purely competition, and so it's really difficult to stand out. The other thing around category selection is that sometimes deciding which category you're in will come with the baggage, whether it be pricing, Sometimes there's negative connotation that comes along. Like I think of AI SDRs, which is starting to become more mature. They've taken a lot of flack lately because some of their advertising is, hey, get rid of humans. And so now that category is like, are people going to keep calling them AI SDRs? We will see as far as products that might even fit into that bucket. The other thing to consider when you take this strategy, you're talking to an aware, a very aware audience of the category. So you can't pretend like the category doesn't exist. So using kind of a silly example, if you go to someone who knows what shoes are and you're trying to like pitch them, hey, I've got this thing, it's gonna make walking better, it'll make your feet feel comfortable, they'll be like, wait, it's a shoe, right? Like you're better off just saying, hey, I've got this badass shoe, you're gonna love it, let me tell you why. And so that's strategy 1. Strategy 2 is similar but has a nuance. So we are gonna select a category to position ourselves in, but we're gonna focus on buyers who are unaware of the category. This is the one that we see most often with the startups we work with, especially in the AI space where there's these categories that are starting to form and are emerging, and you have to take a different approach when you're differentiating to this audience. And so working backwards from the audience, this audience doesn't know the category exists, is not familiar with the vendors that are in that category, as they don't even know it exists in the first place. They're not even sure what the main job to be done of the category is, and they have no understanding of the lingo and communities that may be starting to form out there. And so when you're positioning for a category unaware audience, the way you do this is not using the category anchor. You anchor on a job to be done. And for those not familiar with job to be done, the way we define it is basically where is this audience putting energy? What are their tasks, workflows, initiatives? Like what are they doing in essence where this might fit into their world? And what we basically do is say, hey, you're doing this thing, there's a better way to do that. And you're introducing this new category to them. is a way to think about it. And so if we flash back to DocuSign, maybe 10, 15 years ago, e-signature platforms really weren't a well-known thing. There were some other vendors in the space, but the market kind of looked like this image on the right where the majority of folks have never heard of an e-signature platform. And so when you speak to this audience and you're positioning for them, you can't say we're the best e-signature platform because we'd be like, what the— I've never heard of that. What is that? Instead, you have to speak to the job. Hey, you're trying to get documents signed, but you're doing it now manually. You gotta get a courier. That's really bad. Don't do that. You should get our new solution. It does it automatically online. We call it an e-signature platform. It's more of an educational approach when you're using the job to be done to this unaware, this category unaware audience. Same thing with Calendly. This is a big one for us cuz it's really easy for people to kind of tie themselves to. And Calendly's been very consistent in this positioning leading up to like 2025. Only recently did they change actually into strategy one, whereas you could see here while there were some other tools starting to build around automations and scheduling, Doodle and Acuity probably never heard of those. Most people didn't have a tool like this, and so what they did instead is like, "Hey, you're scheduling meetings, but you're doing it in this really manual way. There's lots of back and forth emails." Hey Angelina, you're free tomorrow at one. What about Friday at two? And they're like, "Don't do that. That's stupid. You need to get this dedicated tool." All education-based. And as I said, what's interesting about Calendly, as the market matured, they actually recently flipped over to the mature, almost the mature category route, talking to mostly an audience who's aware, where now they come out and they're like, we're the number one scheduling automation tool, 'cause most people are aware of that as a concept. And so unlike the other one, where in this case we actually, we wanna pretend that the other competitors don't exist 'cause this audience is unaware, they're not shopping for solutions. And so we see this a lot with startups, um, that are going after greenfield where they think their competitor is like, oh, this other company in my YC batch. It's like, yeah, but that audience has never heard of them. They don't know the category exists. They don't know the vendors. And so that'd be like kind of going back to our silly example, you know, we have the best neural stimulation device. It'd be like, well, wait, I don't even know what that is, right? You need to speak to the job when you're introducing this. And so that's strategy number 2, which brings us to strategy number 3. And then we'll get into the context of Agent Flow to kind of see what this starts to look like and some choices we have to make. But strategy 3 is the position to create a new category. And this is what all founders love because they think new category, they think I'm going to be the leader, I'm going to make so much money. This is the hardest one to pull off and actually has the most risk in almost all cases. In this view, what we're effectively saying is that our product doesn't fit into any existing or emerging categories and we want to basically go rogue where no one's going to ever hear of the thing because it's a totally new category because we're just making up the name of it. And so this whole audience doesn't know you exist, doesn't know these types of solutions exist, and they don't know what the main job of this new category you just made up even is. And in a lot of cases, they don't even have the job where we have to first convince them, hey, you should be doing this thing. And then once you start doing that thing, then you should buy my tool for doing that thing. Right. And so that's a lot of education risk kind of just right up top.

Yeah, I see.

So, oh yeah. And so our approach to this when we're working like through audits and we're working with companies to figure this out, We first kind of look product first. It's like, what did you build your product to do? And this is where we go back to the Jobs to Be Done framework. Again, we're looking at tasks, workflows, initiatives where the capabilities of the product could actually tap into these things. And so a couple examples, when we're doing this, we tend to see a hierarchy of jobs where you, as you start to list things out, you can kind of organize them from specific to broad. And so a company like WorkOS where they're like, well, what does your product do? It's like, well, it lets companies offer SSO. They can do multi-factor auth, they can offer role-based controls. And if you put a high-level cap on it, it's like, it's really about making your software compliant with enterprise requirements. It's like, okay, that's a real job somebody has, but there's no dedicated category around this. And they're basically betting they can create it. And so all of their language is around the job, and they're basically introducing this concept of a new dedicated, like, enterprise requirements style category tool that they're trying to evangelize. Vanta did something very similar when SOC 2 was not a job that people had, right? This was a new thing that companies suddenly started, uh, had to start doing in order to sell enterprise deals. And so there was no SOC 2 automation tool category out there. There was just this new job that people were starting to have to take on. And so it was purely educational for Vanta. Hey, you have to do this new thing right now. The only ways to do it are manual. You should get this new tool. And you'll notice some similarities here between this and even the Calendly example. The big difference is that there really wasn't anybody doing it, right? And there wasn't this kind of wave as people, that people talk about of like, oh, awareness of like, oh yeah, there's tons of these, like ride the, ride the VC wave. This was purely category creation. Similar to Gong, right? They basically had to convince people like, hey, you're already kind of training your reps, but there was no real tool specifically for recording calls, analyzing the calls, and giving feedback. Instead, they were basically competing with the status quo of like, hey, we give our salespeople a pitch. Occasionally our director will sit in on a call and we hope that they do a good job. They basically built a whole tool specifically for this new job to help deliver the perfect pitch, right? Evangelizing conversational intelligence for sales. Now, the pros of this is it's the wild, wild west. You pretty much get to make up the rules. You get to decide how you frame the job, and you kind of get a head start on the distribution. Although I look at this head start as more so a negative because you actually have to go do all that education. You have to spotlight the job. You have to talk about why it's important to address and the problems of not addressing it. And then you have to convince new budget line item to be allocated to your tool to solve this. Takes a long time. We've worked with bunch of folks who've done category creation. The one that stands out is one of the co-founders of Gainsight, Sridhar. We worked with on one of his follow-on startups and his words exactly were, despite Gainsight having a ton of success, I will never do category creation again. It was this 10-plus-year journey, so much pain, so much education, so much talking to people that are like, oh, this is cool, but then not getting any money coming for it until much, much later. The other thing, and I brought this up, is when you create a category, you don't get to tap into the hype cycles, right? And so if we look at some of the third parties that tend to evangelize some of the category creation, excuse me, some of the emerging categories, we see like YC calling for startups. They basically are spotlighting future potential categories, and this can help jumpstart especially early startups' activity to sell into spaces like this. The other risk here is if you say, oh no, it's a new category. But it's not that different from a category that already exists, you can actually create like a negative impact from the market perspective. And so DocuSign, kind of flashing forward, tried to then create a new category of intelligent agreement management, and they've actually done a bunch of funny terminologies of categories. But everybody looked at the main job of this and everyone's like, wait, isn't that just e-signature? And so it's not different enough where it's like, this isn't helping anybody, this is actually hurting. You're now making your core product that clearly fits in an existing category Harder to understand. And so those are the 3 strategies. I'll pause there for a moment if there's questions, but this might be better to dive into Agent Flow and just show you like, all right, what do their angles start to look like from a positioning standpoint?

Do you want to hear about my interpretation about Agent Flow? I have some thoughts on this. And based on your explanation, I felt like maybe initially I was thinking let's create a new category. And now I think maybe, maybe we shouldn't.

It's like. You can do it. You know, I, I get, I'm risk averse to it because I know how challenging it can be, but for some startups, it's the right choice. You just have to know what you're signing up for. That's the big thing.

I remember I watched a TED Talk and they were, it's about how to sell everything. And the main statement I remember was, if you want to sell something that's unfamiliar, you have to attach it to something that's more familiar to people.

Love that.

If you're trying to sell something that's very familiar to people, you need to add something that's unfamiliar, which is more strange component, which is an innovation piece. So you kind of have to have a good mix of the two in order to sell everything.

I love that. And it's pretty spot on to this. We are— another way to reframe that is we're looking for common ground. And it's amazing how easy it is when we're building these new things to talk right past our audience because they just don't— there's nothing to hook into. And so that's what's interesting even about the elevator pitch, some of the phrases you were using, and this is what we'll see in a second. Does anybody understand those, right? Like agent control plane. It's like, that sounds so cool. It's like, but what is that? I have no idea. It's like, well, you know, and it's— I'm not, I'm not the litmus test here. So when we're working through this, the question is, does your audience know what that means? Do they know what they'd use it for? Do they know who the players are? And if the answer is no, it is a signal that we have not found that common ground, that similarity that you were talking about.

Give me a second chance for a second try.

Of course. Yeah, no, and that's— and again, this is what's really fascinating, and I forget who truly owns the quote, when they talk about, hey, the future is already here, it's not evenly distributed. And so your pitch to those building in your space and your founder space, Nicole, I think her name is, that probably works. And what the tricky thing is, is especially in these new and emerging markets, is sometimes the first customers are other tinkerers and innovators. And so you start using all this crazy language that everyone starts to make up around a new space, only to find out when you go sell it to the mass market, everyone's like, huh, I— what? I've never heard of these types of things before. And so there's this interesting growth wall. And if you're curious about that, anybody who's listening, like, go dig into Geoffrey Moore stuff and Crossing the Chasm. He talks about it quite a bit. Because we operate in that kind of same modality.

So this is my interpretation. If I were her, how would I reframe the story so that it's easier to understand for the audience? So for instance, you already know how to maybe manage a company or manage a team, right? You hire people, you give them projects, and as a good manager, you will watch like Who's doing a really good job? And maybe next time you will put those people on the right projects. And over time, the whole company gets better, it gets sharper. So Agents Flow does that for your AI agents. So that's the one thing that I— if you think about every task an agent is running is a performance data point. So Agents Flow will watch which agent is handling what and how well it's doing it. And at what cost, right? Next time when a similar project or similar task comes in, it will assemble the right team on the fly and route the work. You can have actually a live dashboard and you actually know who's doing what and how much they're costing, what's working and what's not working.

You found an angle of familiarity, but I would challenge a bit. And the only thing here is Analogies can be dangerous because we are now speaking in abstractions. And so the moment you talk about managing people, you know, it's like anyone who's managing people is like, great. But the real underlying job, and you eventually got there, was managing and coordinating agents. And so the— and then you'll see this here shortly in our options. How many people are thinking about that activity? Are they actually— are they putting time and energy into it? Because the challenge might be is like, if they're not, than as, all right, you got me to understand it, but there's no market for this right now. And so what we're trying to find with a lot of our positioning is like, hey, this is a new thing. Is anybody doing this now? If so, where and how would we explain it to them? And so I, I quite liked your example, but again, the analogies are super dangerous in explaining this.

Definitely need your expert help on this, but then it is indeed an emerging category in a sense because Such things don't exist yet.

Yeah. Well, let's talk about this. This is a good transition. I wanna share with you the couple lenses. And again, there's probably more than I'm even gonna share, but with the materials I had, I'll kind of give you a sense of like, hey, where do these 3 strategies apply? What do they start to look like?

Sounds good.

And so I've got, I think there's 4 bets, arguably 5, that I wanna take you through here. And they're all, so first one, I'm gonna call 'em category positioning bets, where there is one approach where we position you in a We'll call it an existing category. The question is, how mature or immature is this? And I did note in one of the messages, there was a comment that said, we are not an agent orchestration framework. However, I will say a lot of the materials, they talk about, hey, this is like a framework for like orchestrating agents. And so there's this funny thing sometimes of like what we say we're more than. And we hear a lot of companies say that. Oh, and even Calendly's done this. We're more than a scheduling tool. It's like, no, know your scheduling tool that has a lot of cool things that you do.

Yep.

And so this is one bucket, right? This is a very real category. And conceptually in our model, we would, we could in this approach say, you know what, this is a real category. We're going to position to the people who are very aware of this. So focusing on the category-aware segment and come at them and you can start to see, and you said this earlier, the differentiated approach, the new and kind of unique thing about this. And from what I understand about the product, You're the only self-learning agent orchestration framework where the whole angle here is like, hey, most of the agent orchestration frameworks like LangChain that are out there, you gotta hardcode in kind of like how they're gonna think through applying which agent goes where. Not us. We're the only one that continuously optimizes the model, the— not the hardness, the harness, and the agents used for each task, right? And so this whole strategy, we think of this through a bet landscape. Is basically, hey, we think there's actually a segment of people who are aware of these agent orchestration frameworks. Maybe they have one or they've evaluated one or are thinking about getting one. And we think we can stand out here by calling out our uniqueness against the LangChains that exist in the world. The risk, and this is where there's trade-offs, LangChain's pretty well known, pretty well funded. And so like, are we willing to compete head to head, recognizing that we're a startup with limited resources? Is there enough pain and risk associated to the way this happens in these other frameworks? Or is this kind of a nice to have? Because that can create a little bit of friction when you go and try and sell this. And then can you, and this is kind of a nuanced one, can we handle the objections associated with the hardcoded approach? Because as I was going through this, and we see this with a lot of AIs, when we start to say the AI will do everything, then you start to get like, well, wait, do I have any sense of control? Is it going to be a black box? How will I know what it's doing? And so sometimes that creates that. really adoption friction. And so that's bet one. On the flip side of this, we might say, hey, this is the right category we want to position for, but we actually think the majority of our opportunity is for people who have really not adopted this yet, or maybe they're starting to build with agents, but they've never bought a LangChain before. Where— and I just did a small one here— we would message them totally differently. Hey, you can't let agents run wild. You need to coordinate their efforts. And again, this is really short, short-winded in this piece, but we would speak to some sort of job. And so back to your explanation is like, is anybody trying to already orchestrate or coordinate multiple agents? And the answer might be yes. Then we would just frame it around the job of like, well, hey, the way you're doing that now, you know, are you creating hard code internally with your own approach? Are you just creating linear tasks? And we'd be like, oh no, don't do that. You need to get an agent orchestration framework. to go do this, where we're sort of evangelizing their first buy of this category. So I'll pause there. That's, that would be like option 1 through the 2 lenses. And it's all audience-based of like, where's the opportunity? Where's the energy being put? And where do we think we can actually go execute from a marketing perspective? Because to do this strategy, much like you saw in my deck, a lot of education, we have to figure out how to get in front of these people, convince them this is important, get them to carve out budget line item to go buy a new tool that they don't have today that we're replacing. And that comes with its own kind of downside.

I have several questions around this. So this is our option 1, right? We're, we're competing.

I would call them 1A and 1B, but yes, they are collectively. Yeah.

This is, this is an existing category. This is fairly new. I can see on your graph that it's only the first part of the curve, but this is the biggest market right now that people are kind of familiar with. Agents, orchestration to some extent. And one of the differentiators you mentioned about is self-evolving. There's something also people are talking recently about token maxing, right? How are you spending the money? You should really watch out your wallet. That could be a differentiator.

So this is, this is where we just get into the mistake. I'll start through the mistake lens. Most often what happens is founder builds a product. They're like, oh, there's like 10 different things it does that others don't do that are so cool. The thinking that they usually go through is if I could just get anyone to know all the things it does, they'll see how valuable it is and they'll buy. It's pretty logical. However, that's not exactly how the market consumes that information, because the aim of a good positioning is memorability. And so the more things we add, we're introducing now risk of like, can we actually be known for that thing? Because when you start to add a bunch of them, so like for example, we just walk this out from a messaging perspective. We could say, you know, there's a variation of this where we'd say, hey, the only agent orchestration framework, you know, that reduces, you know, cloud costs or token usage, you know, something along those lines, right? That's pretty good from a differentiation standpoint. But when we clean this up, these are 2 different arguments though, right? And so when we try and merge these, kind of, and I'll walk you through what like the 3rd one would be. We then have to summarize. And so if we want to bring both these arguments in, what would happen— I'll move these down— is we would basically say, we're the best agent orchestration framework because you're not going to get it all in the whole thing. And then it's like, okay, they're saying they're the best, and then we would have these subarguments: we save costs and we are self-learning. And so we actually start to create now some friction of like, wait, will anybody remember this? And, you know, the best is obviously not gonna stand out. Everyone's gonna think they're sort of the best. And so in general, the recommendation is use the framing that's gonna stand out. You can always layer on more of those value props as you get into a conversation like this, because there's dialogue, we get to ask questions, but we're looking for the, we call it the 2D positioning and messaging. What's the, like the word of mouth that we wanna carry that like if someone was in a bar and they bumped into this product, And they, like, even if they were to ask me, hey Rob, what's Agent? What's this product all about? I'm not gonna remember all of it, but I'd be like, oh, they have this agent orchestration framework that, and their whole thing is that they can reduce all the cloud token costs. So the cloud token costs they're doing, or they have this agent orchestration framework and what really sets them apart is that they're self-learning. I'm not gonna bring it all together.

So which one should I choose then?

So, oh, I like this. So this comes back to Where, if we go to this audience that's aware of this, where's the perception on where the challenge is with these other frameworks? In other words, where's the biggest problem, right? Are people who are using, you know, LangChain and the other products here like, man, I'm really having a hard time like optimizing this thing. I can't, I can't get it to like really deliver the results. Or are they like, hey, I'm using this, it's pretty good, but I am burning through a ton of tokens. Each of those would dictate a different message. And this is where, again, we probably don't know, but we have to make really an educated gut feel bet of like, you know, this is how the market's shaping up. This is where I think I could win. And especially since we're putting ourselves against some well-funded incumbents, which one of these is the most defensible, right? Because, and that's the other thing I don't know if I had in my risk component here, it's likely that these other tools are gonna start building for some of these other jobs at some point. And so it's kind of like, where's your edge the sharpest is usually where to stay.

So what I think is the best feature that I have that's harder to copy by incumbents should be the first version I should go with?

Yeah, in general. And this, I would say, it's such a good question. This is where so many founders get stuck because what the temptation is to do a data analysis, but even that limited sampling, and it's not the wrong thing necessarily, But when it comes to positioning, this is where we're talking about, you know, people, they use the phrase like skating where the puck's going. This is kind of our version of that. We can't see around corners. This is an immature category, maybe mature in some areas. And it's kind of our bet of like, where can we carve out our space in a way that's defensible? And that requires a bit of a vision to go and do this and push this into the market. And those who wait to see what people care about will already be behind because someone will have already filled that space as the market really does dictate what this category should and should not do over time.

Is this a one-time thing?

Oh, good. No. However, the very best companies, it's, it is about a longer-term bet. So what I mean by that is over time, your positioning will certainly evolve. The ones that we've seen in lots of different categories are the ones who've been able to remain consistent with their strategy. with their strategy altogether. Meaning if like, you know, we go back to the silly examples I had, like if we go back to like DuckDuckGo, if they just did this for 6 months and then started talking about other things, it's like this positioning wouldn't live and exist where now people think of DuckDuckGo as the privacy-focused browser. And so in general, there should be a longer-term thinking, but when you're early, and so my assumption with this product is because you're early, this may change as quick as 3 to 6 months. But I would say on average with most companies that have traction, most startups that have Your positioning shouldn't drastically shift unless you're making a pivot and likely should only alter every 12 to 18 months or so as you sharpen your angle into the market.

What if a company started having new product lines, like adding more features? Because in the meantime, when they're interacting with the market, the product is also being developed.

Yeah, they're basically introducing risk to the business and it's a question of how you want to grow your business. And so if I just, let me just duck into a different asset here. I got a nice fun visual for this one. And when you look at even the biggest companies in general, they are known for one thing, right? And so, you know, we think this is actually— we'll start here. So even these companies that eventually went multi-product, they owned a space, either a category or a job to be done. This is the aim of good positioning. Now, if you're gonna layer on more products, it just comes down to the question of Are you ready to operationalize that? And because I've been in companies that they basically broaden too quickly and their company starts to look like Company A, where now we're selling multiple products into multiple areas. This is expensive to fund. And so when we're working with these teams and we say, hey, you know, tell me about your marketing team. And, you know, we'll look at these early stage companies and they're like, well, it's me, a half-time marketer. And we're trying to maybe hire one person. I start to get skeptical that they can be in all places. And so what we tell them is, hey, from a marketing perspective, you're probably better off owning a niche. And then through the sales and customer success process, you can upsell and cross-sell without having to do marketing. And so like even taking our consultancy as an example, we just do these positioning sprints, but if and when we want to expand, we get asked to do a bunch of stuff, sales decks, LinkedIn content, all those other things. We would do that through the sales process, but it wouldn't really impact our front-facing positioning. We would still wanna be known as we're the group you go to if you're in B2B and you need help with positioning and messaging. The other stuff can hide behind the scenes. Okay. I wanna walk out the rest of these 'cause these will be fun and then it'll probably trigger a bunch of fun questions on your end. And so this is just one bet when we say, hey, this is a real category. Who do we wanna position for? What are the different angles that we might go do that? There's another category, AI agent training. Which in the materials you sent was mostly immature and seemed like it was more research-based. But there is folks who are aware of this. I would argue this maybe fits even closer. And this is where we'd have to kind of split hairs of like LLM training versus agent training. Is that the same thing? Are those different? And I actually don't know the differentiation here, but I just wanted to show this would be a totally different angle where we would say, hey, we're an AI agent training tool of sorts, which what I wanted to spotlight in this example. is it changes our competition, right? To the folks who are actually buying these types of tools, we're no longer competing with LangChain anymore, right? Because that's doing— it's carrying out another thing that a different category has. And then there was even some categories around agent management, agent builders. I excluded these because I felt like at least the way that the product's set up today, it probably doesn't have the features that some of these other big incumbents do. And they also skew towards the governance side and the building side. And I just didn't see that in the product, but maybe these are other things. And this is where the only reason I actually created this slide. Categories come with the baggage. And so if you call yourself an agent management platform, which you're gonna start to bump into, like, you know, Microsoft Copilot, Salesforce's Agentforce, AWS Bedrock Agent Core, they're gonna expect, oh wait, yeah, but where do I build my agents? And it's like, do you connect all these tools? And that's where like you and your founder will be like, wait, do we have any of this functionality? Because if we start calling ourselves this, The buyer is gonna expect us to have some of those features. And so those are some of the category bets. This is where we move now into more category creation bets, which I think these are actually quite interesting. And again, the idea here is that there's not a category anchor we can really hang our hat on. So it all comes down to like, what's the job? In other words, when you show up to a prospect, they're gonna be like, what would I use your product to do? When would I use it? You could, and this is like maybe the most specific, and there's a couple flavors of this. We could show up and we call it, we think of this as an add-on positioning where you show up as a long chain agent optimization add-on where from a job to be done standpoint, you're basically saying, hey, we're a better way to optimize your agent orchestration if you're a LangChain user. Now this comes with some risks and some questions associated to this. Does anybody use this language? In other words, if I'm using LangChain, and actually, is it chain? Chan? Actually don't know. If I'm doing this, am I putting energy into optimizing the orchestration where I'm looking at outputs? I'm going back and changing the iterations of the agents and changing the hard code. If so, this positioning works where we would come with some narrative like this. Hey, you have LangChain. It's great for orchestration, but it has some gaps. It's got some things it doesn't do, especially when you're trying to optimize your agent orchestration. Where we'd be trying to position as like, hey, keep your tool, but you do need this very separate specialized tool for this job. Now, the risks that come along is like, wait, isn't LinkChain gonna build this functionality at some point?

That's my question.

Yeah. And so how long does this opportunity even exist? And so like, there's a couple examples. If you think of like Salesforce with this style of strategy, there's a lot of companies that position Effectively as a Salesforce add-on. We saw this with Chrome extensions as well. We just worked with a company. Oh my God! What is their name? Upside, Slide Up. One of those. They basically do Office 365 stuff, but specifically for financial advisors, and they position as an Office 365 add-on and do almost $100 million in revenue. And so this is kind of a ecosystem platform fast follow strategy, but it comes with the risk of like you know are you just going to be eaten up by you know this. Product as a whole. There's a broader version of this where you don't say you're a LangChain add-on, but you could say, hey, we're an agent orchestration framework add-on. Same thing applies, but when you go broader, you actually introduce now even more variety because do all of those platforms that fit into that category have the same limitations as LangChain? Probably not. And so this is where you have to probably start to run then sub-campaigns because You know, this obviously works to someone who's using LangChain, but if I'm using one of the other ones and I see this, I'll go, this isn't me.

Not for me.

Yeah, exactly right. But it's pretty specific. It's anchored in the job. And then there's, you know, just to show you what another flavor of this different job, and this is the one you talked about where the job is now token optimization, where we would just speak to the job and I'm doing it specific to Claude. It's got the same sort of challenge of like, is this all? AI token cost? Is it just for Claude? And so, so there's some segmentation specificity stuff that we would ask, but we would say something to the effect of, hey, reduce AI token cost by 70% without sacrificing quality, where we would then truly position as this AI token optimization tool to this unaware category who's never bought these kind of tools before. But obviously very different than the vision and even the way you initially explained it of like, hey, this is about control planes and orchestration. Where sure, that's sort of the secret sauce of how we deliver savings, like of where we look, but it wouldn't be the leading piece. We'd be talking to an audience that's very cost conscious overall. And the question here, is this the right job? Because there's some other things I saw that the product did. You know, do we position specific on like automated model selection? You know, we talked about orchestration or harness and agent selection. These are other theoretical jobs that could be positioned for where we can make the argument of like, you need a specific tool to go do that. And so I'll pause there. 2 quick lenses. There's probably more categories. And where we usually spend our time with founders is basically mapping all of these and then asking the hard question of like, which one can you actually execute today? And I'll lead in and you probably have a lot of questions. Most often what tends to happen is startups, when they think about positioning, they go too broad, too fast. Where what our recommendation usually is like, hey, where can we find the beachhead? Where can we get dominant recognition in some space where we layer on more segmentation, get really, really specific, similar to what I did here, to then broaden into some of these other areas as we grow.

Your option C of the variations of, you know, niching down more on like agent orchestration at, um, Or reducing cost, you know, or even almost sounding like we are an extension or like Chrome extension in a marketplace or in an ecosystem. I get it. It's very easy to understand and probably targeting at very clear pain points. But what about our grand vision of like, this is a required management system for your whole AI agent team? You know, where's the grand vision?

Yeah. Yeah. And, and this is the, the usual challenge for early stage. And so what sometimes happens, and this might be the right path for you all right now, is you may see all of these and go, I don't like any of them. They're too small of a market. They don't tie into my vision. I think we do more than all of these things. There's a world where you could position broadly, basically articulate the vision, which will probably be a version of this. I would imagine, hey, you know, speaking to orchestration of agents at a high level and optimizing and saving costs. The risk is, and I think that's kind of where you guys are at now, is like, is, can you actually be known for that thing where theoretically someone would shop for that, where you guys would stand out? And the answer is most likely no. And so in early stage, sometimes just explaining the technology and then seeing if you get pulled in any of these directions. Can be the right positioning strategy. We just need to be hyper-aware that it's an experimentation strategy. And what would happen probably is you would end up with customers like this where, like, if you did that for 3 to 6 months, you would find that, hey, some people came to us for cost savings. Some people came to us because they're using LangChain and they wanted an upgrade. Some people came to us because they didn't have anything for orchestration. Then you have a decision to make because this business won't scale. Foundationally. And so when we look at this, we try and take realistic view of where are you at, what are your resources, and what's the next milestone that you're trying to position for. And what we usually see is that most startups have a license to be specific, but they have to be willing to ditch their ego, ditch their vision temporarily to earn that initial market share and then use that funding to expand. But I know it's hard to kill, it's hard to kill the vision and you may get some early adopters there, but that's likely not where the majority of the opportunity is going to be.

So what's your suggestion for early stage to test in terms of timeline and milestone to test your positioning for, let's say, the first 2 years?

Yeah, it's going to be different for each market just based on the distribution that's in place and also just the scope of things. Overall. And so it's hard. There's no like specific experimentation. What I will say is that when you're in that experimentation phase, I would try and not stay in that space for more than 6 months. Because you're just going to get pulled in a lot of different directions no matter what you're doing. As far as the how of that, I would avoid A/B tests unless you have some high volume way of bringing people to something like a site. But even then there's too many variables. You're going to learn a lot more in interactions like this. And so if you have a positioning, the aim would be, how do I get in a room with somebody for 10 or 15 minutes, give them my, my spiel of a pitch and see where they lean in, lean out, get confused. Obviously, you then need to see if you can actually close them and bring them into the product. But that's where you're kind of just getting into product-market fit seeking. And I get the sense that that's kind of where this project is at as well.

Yep.

And so in the early days, and this is where we get really frustrated when we hear a lot of other people talk about positioning, especially for startups, because there's some folks that assume that every product has some differentiation. And I'll tell you, working with hundreds of companies, that's just not true. And so sometimes what can happen when we're working on these is we work through all of these angles, and then we realize, oh my gosh, we don't have differentiation no matter what we do. I don't think that's the case here. And so I would encourage to lean into like, Who are you building for? And like, don't build for everyone, which is the trap that a lot of people fall into because they think more opportunity, I'll build more things, I'll go broad. You're harder to buy, your story will be less believable. And remember, the main issue you're going to have is no one knows who the heck you are at this point, right? We don't have the brand recognition to go and do this.

What the heck is really positioning about? Is it really the words I use on my landing page? Is it a wordplay?

No, it is not. It's to me, like, that's just the expression of the strategy. Ultimately, positioning boils down to target market, like how you're segmenting things, and why you're better. And again, whether it's mature and immature, you can get into all of that. But it's usually a segmentation problem, and really just figuring out who you're building for. And it's crazy how many founders, they build on top of the tech, and they think in terms of look at this big TAM, What we don't like to admit as founders and marketers, markets are wildly fragmented. And we kind of know this at a human level. We think of ourselves, especially here in the States, like everyone's an individual. We're all unique. We have unique circumstances. We come from different backgrounds. Companies who buy software and products are the same way. They're unique. They're in unique circumstances. They have unique tool sets. They have their own team with its own baggage. Not everyone is the same. And granted, marketing, we try and put people into these buckets. You know, it's our own heuristic to make them easy to speak to, but the reality is everyone's in a different situation. And so what that means is like, we have to make some hard decisions of like, who are we actually gonna help and who are we not gonna help? And a lot of founders aren't gonna do that. So tied to your question, it's not what you put on your homepage necessarily, although I can usually look at a homepage and tell you how they're thinking about positioning. It's much more the really key decision of like, who really is our ideal customer right now that we're going after? Most people just don't know who that is. And you'll hear things like, our ICP, we have 6 of them. And then they go and list them out. It's like a company with 6 ICPs is either extremely successful and has the team and resources to go after different segments, or they're kind of a delusional early stage startup that doesn't really know who they're serving just yet.

Okay. Finding your target before you shoot. While you're shooting, you're defining that, that target. Speaking of landing page, do you think landing pages are dying if the interface is becoming like chat, a prompt, an agent? I still see landing page as the symptom or the representation of a positioning. So if everything changes to chat, so where does positioning even gonna live?

Uh, it's an interesting question. I don't think they're dying just yet. Especially the way that these AI search works, and I'm by no means an expert, but I've spoken to many experts. Right now, a lot of the AEO, GEO stuff is still rooted deeply in SEO, which is really rooted in, hey, what are you publishing out into the world? I love that it's now bringing into social. And so what's happening is basically like what you publish out on kind of the old school internet and social media platforms, it's gonna be the summation of that that these AI systems pick up altogether. And so I don't think they're dead. I think they're the feeder to AI to actually figure out what to call yourself. And this goes back to specialization versus broadness. If you have landing pages and social posts about everything under the sun, all these different segments calling yourselves all different things, the AI won't know what to do with you versus if you have a pretty consistent 80/20 of like 80% of the time we are, we call ourselves an agent orchestration framework. There's a pretty good chance that the AI is going to pick up when people are searching of like, oh yeah, here's a list of well-known agent orchestration frameworks because it's all that they talk about. And so I don't think they're going away overall.

Is there one thing that you really hate about founders that you keep telling them and they just don't get it or they just don't execute it?

I think it's such a tough question. There's a, there's a few things. I think the ones that there's a difference between understanding good positioning and messaging and then actually having the leadership skills to execute. It's probably the biggest issue. And so the way this shows up in our work, and we work with a lot of talented founders, is they come in, they actually kind of know their stuff, right? We run a version of this exercise, they're oozing with insights. They help us shape and frame all these different strategies. But then they don't choose. And that's a— that's just bad leadership, right? The reason why founders and executives get paid the big bucks isn't necessarily for their like in-the-weeds smarts. It's because they have to make these tough choices that require trade-offs. And so like, I think back to a founder I worked with, He phrased it beautifully when we got to this point of like, hey, it's time to position. And they had like 6 or 7 different, truly different strategies they could go down. And he said, you know, Rob, all 6 of these suck and it's not your fault. He's like, they all just suck because they all have trade-offs. And it's sort of this whole exercise is like, choose your hard. Go into a competitive space and see if you could battle it out. Go into a non-competitive space and hope that you could educate. And then the spectrum in between. They're all really hard, like, and they come with tons of uncertainty and you can't predict it 100%. And so we always joke with our founders of like, if you wanted certainty, you should have been an accountant, not a founder. Like entrepreneurship is just really hard like that. And so it's making those tough decisions and then sticking with them. And that's probably the most frustrating thing for me is to see a really good strategy only being run for 10 days where they'll, they're like, oh, I love this. And then, you know, we check back with them and they're, they're back doing the old broad thing and they're like, yeah, I, And they basically chickened out. And we try and call them out on that when they're going through our process, because this requires commitment.

This whole conversation, we keep coming back on how do, how does a founder earn a position in any market? It's going to take time and there's really no shortcut.

Yes. And the only outliers to this is if you have some unfair distribution advantage, which I hate to break it to 99.9% of the founders. But if you're not famous, you don't have a huge professional network, you don't have existing partner network, you don't have a PLG network effect in your product, you got to do it the old traditional way. And yeah, that's the time-consuming thing. And you just have to repeat yourself over and over for years until people suddenly are like, man, that's the person that always talks about X. And it's like, yes, that's when you know it's starting to work. But too many people, they quit before it starts working. You know, and this stuff takes 18 to 36 months to truly carve out a reputation. At least that's the fast route. And most founders don't have the staying power for that.

How long did it take you?

It took us about 2 years to really start to get the recognition. I would say because of how active my co-founder and I were on LinkedIn in our little bubble that we were focusing on, you could make the argument that within 9 to 12 months we started to get attention, but far from recognition. And then even for us, we had to shape what that recognition was because we were still kind of squirming with a lot of different ideas, right? Are we, are we the homepage guys? Are we the positioning guys? Like, how, how do we frame this? And so you'll even see an evolution in our positioning where we actually carved out recognition around homepage and homepage messaging. Even though we were doing positioning work, that was the thing that people knew us for. Like, oh, you guys are guys who do the homepages. And for us, we actually were like, no, we, we're trying to carve out this positioning expertise cuz it's really what we think we do well and where the value is. And so that's where maybe even the last 2 years, and you'll see it on our homepage and the timeline, everything shifted towards positioning. And so I, and even now there's a ton of people that don't know who the hell we are that are in this audience space that like, and you know, our space is interesting cuz it's very fragmented where not a lot of people hire positioning consultants specifically. And so what do we do? We just keep talking about positioning and repositioning and educating around it and And we have to keep doing that, you know, even thousands of posts in and hundreds of podcasts and tons of speaking engagements, there's still more work to be done. To us, it's repeating ourselves. But if we stop repeating ourselves, we have no chance of truly carving that out and getting even more recognition.

Do you know why? Because I think the founders are making the same mistakes too. And the newer founders coming in and builders becoming founders, we all have the same problems. We all think the same way.

For sure.

Right.

And we have the same issue. And I will tell you, like, having a co-founder goes a long way. We have this inside joke, Angelina, where like I told you earlier, people ask us to do a lot of different things that isn't positioning work. Our whole approach is like, no, we wanna be good at one thing. We wanna be known for one thing. We think we're gonna get scalability and processes and automation in that way. And so whenever one of us gets like, you know, squirrel syndrome of like, oh wait, somebody just asked me to do this, let's go do it. We, the jab to the other one of us is like, oh, that's product number 64, product number 132. And it's just a quick reminder of ourselves of like, hey, let's stay focused on what we really do well. We'll chase all the downside opportunities after we truly get this business in a place where it's like operating even without us. And, uh, but when you're on your own, it's hard because new opportunities come up. And especially when people are waving money in your, in your face, it's hard to say no to those things.

Have you thought about quitting during those 2 years?

I look back on those 2 years, we were running so fast, but like Anthony and I, and this, this kind of gets more into personality. We love, like, when things work. So like, we did these— I don't know if you're familiar with Enneagram. But we both are, I think the number's 3, which means we're sort of success-oriented, or at least perception of success. And so for us, we don't care what kind of work we're doing, as long as it's growing. And we like building things, we are happy campers. And so we were fortunate that as we in the beginning, we were having all these micro successes, even when we weren't really making that much money. And then as that started to compound, like, that was just so fun, even though it was a ton of work.

Yeah.

And so that's the long-winded way of saying, yeah, no, not at all.

You have to be, you have to do something that you're really passionate about.

Yes. But I would put a flavor on the passion is you're not going to love every aspect. Like you have, and that's why, like for everyone, you have to kind of understand what you enjoy about the work. There's a lot of the work we do that I truly do not enjoy, but I frame it within the picture of the things I do enjoy of like working with new founders and exposure and building things. But it's like, oh God, I hate some of these work aspects. It's like, yeah, that— it's just hard work. And so that's where conflating, like, oh, I have to love everything I do in my job— I personally don't think that's true.

But you cannot be hating everything. If you're hating 100%, then don't do it.

Yeah, that's what you got. You got to find something to enjoy, right? Like, I'm with you on that. It can't just be empty money chasing.

Thank you so much, Rob. I have a traditional closing. Could you recommend me or our audience a book? I can see that you have a lot of books behind you, so maybe you can recommend us a book to make us better founders.

Ooh, I actually have a few and it kind of depends on who your audience is. The book that I find myself rereading, because I just love all their stuff, I think they're great writers and I also think it's great advice regardless of where you're at in the journey, is Rework by Jason Fried and DAH, the Basecamp guys. It's just, there's so much knowledge built into that book. Every time I reread it, I just get another insight. If you're really curious about positioning and messaging work, especially in new and emerging markets, I can't recommend Geoffrey Moore's Crossing the Chasm enough. That's another one that we frequently go back to, and a lot of our models and frameworks are really rooted in that thinking. I just think it's really, really strong. And then the last one, and this is maybe more for service-oriented business or those who are curious about productized service models. Which is how we're building Fletch. There's a book called Built to Sell by John Worrelo, and it's basically a, it's like a narrative structure and it talks about a full-service marketing agency that's really struggling through growth and systemization and their journey to productization. And what you'll find in reading that is a lot of it comes back to the stuff I talked about today of getting really specific, focusing on one thing, being consistent, building processes. And that was really influential to us as we were building our business.

Thank you so much, Rob. I really enjoyed the conversation today. And I feel like one thing that I still remember, the hardest thing among everything you said is to focus on one thing. I think this is the hardest thing in this whole conversation.

For sure.

Thank you for a reminder.

You're very welcome. And it's hard to focus in the age of AI and there's a lot of shifting things under our feet. But if you, it's sort of like not actually taking a full swing. And so try and stay focused the best you can, give yourself an actual chance to make it.

Yeah, especially for the technical founders I talk to, every— everyone's racing to build more features, more products, ship faster, spend more tokens. So it's really hard to trim down. Thank you, Rob.