Marketing School - Digital Marketing and Online Marketing Tips

Anthropic's New MOAT Is Insane

24 min
Apr 22, 2026about 1 month ago
Listen to Episode
Summary

Eric and Neil discuss Anthropic's competitive moat in enterprise AI, which centers on permissioning and governance rather than raw intelligence. They explore how institutional AI requires deep organizational knowledge bases, share strategies for enterprise deal-making through relationship-building and introductions, and discuss the tension between business growth and family time.

Insights
  • Enterprise AI's scarcity is shifting from model intelligence to permission/governance capabilities—companies need robust systems to control what AI agents can access and execute
  • Institutional AI success depends on deep, organization-specific knowledge documentation (skill.md files) rather than general capabilities; going deep in one area beats being everything to everyone
  • Enterprise sales at scale require multi-year relationship building and peer introductions from trusted networks, not just product quality or marketing tactics
  • AI agents are generating measurable ROI through cost optimization and operational efficiency, but implementation requires careful governance to prevent costly mistakes
  • Geographic and market-segment arbitrage exists for AI solutions—technologies adopted in the US take longer to reach other countries and SMB segments, creating expansion opportunities
Trends
Enterprise AI adoption bottleneck shifting from model capability to governance and permissioning infrastructureRise of AI agents for internal business operations (cost analysis, ad generation, cold email infrastructure) with measurable ROIInstitutional AI requiring deep organizational knowledge bases and SOPs as competitive moat rather than model intelligenceSlower enterprise adoption cycles creating geographic and segment-based arbitrage opportunities for AI solutionsCompliance and governance (SOC 2, NDAs, non-competes) becoming critical differentiators in enterprise AI deploymentAI-driven cost optimization and operational efficiency replacing headcount reduction as primary value driverPeer-to-peer introductions and relationship-based sales becoming more critical as enterprise deals scaleMulti-language and localized AI solutions for non-US markets and SMB segments still in early adoption phase
Companies
Anthropic
Central focus of discussion regarding their competitive moat in enterprise AI, specifically around permissioning and ...
Single Brain
Eric's AI marketing services company building agents for enterprise clients with governance infrastructure and cost o...
Zapier
Recently launched governance features for AI agents, cited as example of companies addressing permissioning challenges
OpenAI
Referenced for ChatGPT's limitations in travel planning and decision-making, illustrating AI capability gaps
Foundation Capital
Jaya Gupta from Foundation Capital tweeted the insight about permission being the scarce asset in enterprise AI
Disney
Used as example of expensive service (cruise pricing) and ChatGPT's poor decision-making in vacation planning
Apple
Referenced as example of large enterprise target for multi-figure deal negotiations
Microsoft
Referenced as example of large enterprise target for multi-figure deal negotiations
YPO
Young Presidents' Organization used as networking platform for generating enterprise pilot deals and relationships
ClickFlow
Sponsor tool for AI SEO content creation that sounds natural rather than AI-generated
People
Eric
Co-host discussing enterprise AI implementation, governance challenges, and business scaling with focus on operations
Neil
Co-host focusing on enterprise sales strategy, relationship-based deal-making, and whale hunting for multi-figure con...
Jaya Gupta
Tweeted insight about permission being the scarce asset in enterprise AI, sparking main discussion topic
Quotes
"The scarce asset in enterprise AI is shifting from intelligence to permission."
Jaya Gupta (Foundation Capital)Early in episode
"Models are now good enough. The harder question is whether companies will let them act inside real systems."
EricMid-episode
"Institutional AI is, for example, Neil and I are both marketers... what ends up becoming more important is the skill.md files that you have throughout your whole organization."
EricMid-episode
"I spent 7,500 on tokens last month. But the savings on that, if you do 500 grand divided by the 7,500, that's a 66.66x return."
EricLate episode
"My kids keep saying i don't need the room i don't need the toys i don't need any of this i want my dad."
NeilEnd of episode
Full Transcript
Anthropic sees a moat, Neil. So what do you think this moat is, if you already guessed on this moat? This is too up to, so I don't think you're going to get it. So try it. No, I don't think I'm going to get it. I wouldn't say it's enterprise, you know, or adoption or anything because they're already at, all the enterprises are adopting whatever platforms they want. I don't know. So this is interesting. This guy, Jaya Gupta from Foundation Capital. So he tweeted this. So Anthropic sees this moat. The scarce asset in enterprise AI is shifting from intelligence to permission. And we're seeing this right now, Neil. So same thing, you're getting demand for agents. We're getting a lot of demand for agents and we're building out these revenue agents right now. And what's happening, Neil, is our agents, because my team wants to have agents like mine where they can ask it to do whatever and it has a lot of content. It's connected to all the data that we have internally. And there's all our SOPs, all these things, right? And what I'm realizing is that permissioning is a very big deal. So obviously you shouldn't let everyone in your company see all of your QuickBooks information or all compensation information, right? You shouldn't have that. And so you need to, so models are now good enough. The harder question is whether companies will let them act inside real systems. So will you let them write code? Will you let every one of your team members merge changes with your code, right? Or touch infrastructure or trigger actions? That is a big deal. And you need to have this governance in place. And some companies are launching this right now, like Zapier just launched something in terms of governance and so i think you're gonna have to piece me a lot of these things together we're doing the same thing as well but that's why we're we slowly started to test on ourselves first with our single brain unified intelligence and then we're slowly rolling it out to clients too but they're all you know i'll say all of the clients right now they're all on pilots um because we want to make sure that we're building with them every single week because every client has different needs um and that's why like i'll tell you like having to get sock getting ready for sock to compliance like all these things are a huge pain in the butt as you know um but we want to make sure that we're building these things right, as I'm sure you guys are too. So governance matters, permission matters quite a bit. Otherwise, it's going to become a poo-poo show at your company, and you don't want that to happen. Yes. And also, when you are leveraging some of these marketing tactics, let's say people are going to, marketing agencies are creating agents for others, or you're going to ride on this way. Keep in mind that if something has been out there for six months or a year and is already popular and like, man, I missed the boat. If you switch the marketing angle to focus more on enterprise, you'll be shocked at how much slower companies adapt and you can penetrate. And then if you're like, well, in the US, they're already starting to adapt. That could just be in the US. And a lot of other countries are much slower to adapt in the US because a lot of this technology gets rolled out from San Francisco to the whole US and then to other countries. So what you can do is say, hey, I'm going to end up creating this for SMBs or enterprises in Portuguese or in Hindi or in whatever language and then adapt for those countries that you know will eventually follow through, but they're not there yet. Yeah. Remember we talked about this, Neil? Institutional AI versus individual AI. So individual AI could be like you using ChatGPT to help it figure out recipes for you or plan your travel or whatever. Okay. Institutional AI is, for example, Neil and I are both marketers. We both know quite a bit on market. I won't say we're the best in the world, but we have a lot of SOPs. We have staff that help us. And so what ends up becoming more important is the skill.md files that you have throughout your whole organization because that's how these agents read this stuff, right? It's all these skill.md files. The more that you have and the deeper they are, the more of an advantage you're going to have. And you want to go institutional in terms of deep in one area and not just try to be everything to everyone. Yes. Did I give you the example of how AI really messed up a family vacation for me? Yeah. You told me this like three times where it spent like 7,500 over or something like that? No, no. How it put me on a cruise ship. Did I tell you that? No, no, no. Go. How Michelle wanted me to spend more time with the kids. She accepts, as my wife, she already knows I work a lot. It's more so the kids are starting to cry. Like even my daughter said, you haven't come to any of my school functions. I'm like, I volunteer during lunch and stuff. She's like, you haven't seen my open, you won't come from open house and see the homework I did. You haven't seen my open cloth. and then and i'm like honey i'm so sorry i had a conference book three months ago or some meetings and i can't in a different country i can't just fly back from italy and then fly back over and uh so my wife planned a disney cruise okay using chat gpt shitty internet coordinated with my assistant so she knows like when my calendar was free and all that kind of stuff um so the Disney cruise or from Vancouver to Alaska, JetGPT ends up breaking down like, oh, here's a perfect date in June. Keep in mind, I had a few dates open for many months or a few weeks open. And it tells my wife how June is the best time. All right. So then we tell some friends who have kids similar to our kids' age that are in that area, like, hey, do you want to join us and whatnot? Hands down, everyone we told, some of them may end up joining us, but hands down, everyone we told, said, why are you going in June? July is way better. June is not a good month to go on the cruise. So my wife goes back to Chad, hey, you told me June is the best month to go. Everyone's telling me July. And they're like, oh, you make a good point here. But the reason I pick June is when you look outside you can still see a little bit of snow at the top of the mountain which looks beautiful So that why we say it the best month But then you can do crap outside because of the weather and it not conducive I was like you picked a crap month and made us waste money on a non-refundable trip and not just a little bit of money. Flights, I think, were like $4,000. The cruise, $20-something thousand dollars. I do not know how the heck people can afford this Disney cruise. I was pissed off at the price. People don't have options. this isn't a fancy room this is like a motel six looking room i'm like how the heck does it cost this much money they just charge that much and get away with it but chat gpt this is how chat gpt screwed you out of 30 grad probably more but yes to start off with and it picked the wrong date probably 50 and then i'll have forget the cost on that i'll have crappy internet for the whole week so i'll be very unproductive that will cost you even oh yeah it's gonna cost you way more but I get that one because I've been working so much and Eric sent me amazing X quote or quote from X from this uh guy talking about the net worth versus spending time with kids and how you can't take it with you and I agree with him and I've been trying to make it better but I actually found I've probably traveled more this year than last year and it's probably worse I was just gonna say before you said that I was like I need to remind Neil to reread the IRR meme so the whole idea with that one is nobody your daughters are and your your son's not going to care at age 21 how much your irr was at 31 versus 25 so you can spend more time with them right i was going to say you've actually regressed this year so you need to reread that yeah it's gone a little crazy you've done really well you've cut back traveling a lot if i'm not mistaken like i've noticed way less trout and you spent a lot more time on operations on your business like actually doing the work instead of going out there and trying to build a brand, collect leads and all that kind of stuff. What has been, out of curiosity, if you had to put some tangible numbers or percentages, what would you say, rough range, because I know it's too hard to exact, but you probably have a rough idea because I know you would have already adapted if it was hurting you. Yeah, so here's what I can share. So I will say, I was, Neil, like we can do a quick comparison. At the high point, I was probably traveling half of the year. So 140 to 150 days, sometimes even like 170. or something like that. I think you're probably on track for probably a little more than that, it sounds like. I think I'm on track for like 65% to 70. Okay, so Neil's worse than I am. But I used to be kind of around that. What I found though, is that me focusing on my business, I'm working harder than I ever have. And so usually Neil and I would be focused more on, or at least for myself, I'd be focused more on strategy and driving distribution. And this to me is part of driving distribution. But I'm doing more than I ever have done. So I think now, Neil, you can't just be a strategist. You have to be a doer as well. And you and I have always, we're willing to kind of pick up a shovel or like clean the bathrooms if we have to do it. And so I would just say that I've been more productive than ever because I'm not distracted by travel. And I'm able to quickly meet with people more. I'm able to quickly just, you know, huddle them or something like that, right? So communication cycles are a lot faster now. And I would just say that without me focusing on this, I wouldn't have gotten the single brain thing out, which I think single brain is the AI marketing services company of the future. And that to me, building that future is very important. Yeah. No, I think that's a great use of your time. But let's go back to your organization, Single Brain. I know for a fact, AI has helped you save. In addition to whatever you spent, if you look at total savings, it saved you probably over a half a million to a million dollars and told you where to save money without reducing headcount, like without having to fire people or anything like that. It's actually saved money in a positive way. Yeah. So what I can report on is, so I actually made a video on this yesterday, Neil. So my agent just on cost savings and loss, I spent 7,500 on tokens last month. But the savings on that, if you do 500 grand divided by the 7,500, that's a 66.66x return. Just on that alone. And I made those cuts within three days. So we cut a ton of software costs. We did cut contractor costs, but we didn't cut. What I can proudly say for right now is from that, we haven't tried to cut full-time costs. And that's a good thing. I think it's like a lot of people are saying, oh, we're good. We got to cut like 20%, 50% here. I talk about it a lot, but that's not something that we're actively doing, right? Now, also, on top of that, Neil, the way my team uses it now. Like, sure, I spent the $7,500, but that was the toll booth to get everyone else using it like this. So everyone else is getting a lift. They're like, oh, I built this over here. I built this. I saw someone on the team build this. We have something now that can crank out like 400, 600 ads, and we're hooking it into a single brain. We have like an API for that. And then end-to-end, having Alfred, our agent, build out cold email infrastructure, which we've talked about a little bit. So it's building out all these end-to-end things. Had I not spent the time and the money on that, we wouldn't have been able to show the team that we can do it and unlock it for everyone else. And that to me, like, I think we can see the gains on that. It come back to me in like six months or so, but I think we'll have harder numbers there. But also, Neil, the single brain website generated in two hours, the enterprise leads coming through that right now. Like, that's real. That's real pipeline. Quick plug If you want to win at AI SEO then you got to check out ClickFlow at clickflow It creates content that actually sounds like you not AI slop And there a free 14 day free trial Okay, back to it. Yeah. Yeah, no, I'm with you. And I've been in the opposite. I haven't been in the trenches as much as Eric has this year in business. My role this year, if I had to say, I've been spending, and even a lot of last year, but it's gotten worse this year. I spend a lot of my time networking and doing meals and doing presentations. When I say presentations, I'm not talking about going to a conference, even though I do that quite a bit and I just finished one up earlier today. I'm talking about like going into companies and meeting their executive team, doing presentations internally about business strategy, marketing strategy, how they can adapt, what happens when low cost competitors enter the market, how should they adapt their marketing messages and pivot, still try to maintain and grow market share. And we'll do lunches and dinners with these people and get to know them. And the reason this is important is I'm playing like a game of whale hunting, right? I don't actually hunt whales, but you guys get the point in which... You call this a spearing, Neil. Spearing, yeah. What I'm trying to do is land multi-seven or eight-figure deals. and it's a lot of those, it doesn't matter how good you are with AI or marketing or any of that stuff, a lot of that comes down to relationships and whining and dining and meeting and hanging out and just getting to know people because people don't like signing big contracts with someone they don't actually know. I'm not talking about, oh, I see Neil and Eric on the internet type of note. I'm talking about sit down and have a drink, even though I don't drink alcohol, but like, you know, have a meal or break bread or whatever you want to call it, like really getting to know the person, their family, and what they do for fun and all that good stuff. You know, what's unexpected, Neil, is the pilots that are coming through right now for SingleBrain, a lot of them are, to your point, relationships. And a lot of them are through what? YPO, right? You know, I talked about it, but that YPO is networking. What Neil's talking about is a form of networking. I think that's very, everything comes down to relationships at the end of the day. And I think it is high leverage for Neil to be doing this. And Neil, I got a question for you. What happened to that nine-figure deal? The nine-figure deal did not close. Although I got a verbal on an eight-figure deal yesterday. So that's not bad. And when I say verbal, it wasn't verbal. We were going over the last red lines on a phone call. It's basically done. Yeah. So what Neil does, you want to remind him what you do with the outreach? Because I think it's something everybody can do. It's free. You might as well do it. Yeah. So what I really think about is who am I connected with? that knows other really influential people at other companies. And I will try to convince the other person because I built relationships with all these people that I'm connected with to go and talk to the person that they know at a larger corporation like Apple or Microsoft or whoever it is. And not only introduce me to the person I'm trying to meet, but explain to them like, hey, I want to introduce you to X, Y, and Z. you know they're doing some really cool stuff in your space and i'll provide the examples and then have them say like this person is really legit you should be working with them they're amazing and i found that i get really massive contract from that strategy now you may be able you may be thinking well neil this is really hard for you to me to replicate and it doesn't matter if you use air or not it is going to be hard for you to replicate the things that work out really well is i built these relationships over a long time so you got to put in the work there I built up a personal brand, which makes it easier to convert and close people. So you got to put in the effort and put in the time and years there. And my corporation has an amazing track record, a lot of awards. We're skilled to size and have tons of case studies that we can end up showing. So all those combined with the correct introduction from the other person who they don't look at as just like an employee or someone beneath them, but someone at their same level, like a peer that they respect, that works really well. It's like if someone has a hot flying startup that is now worth $4 billion and they look up to someone and that someone does an intro about me to them, there's a really good chance I can close a big deal. Yeah, you know what I'll say? So two things here. One, impact, confidence, ease. So those of you saying, okay, I can't do that right now because to Neil's point, from an ease standpoint, Neil can easily reach out to people because he already has this network and that makes it easier because he's built this reputation overnight. So you're not going to be able to do that overnight. But from a confidence standpoint, I'm fairly confident on scale of one to 10 that this is like, this works. Literally, because Neil's talking about it right now. Impact can be huge. So let's just rate this real quick. One through 10 on all of these, right? So impact, to me, this is like a nine out of 10 at the least. Confidence, probably like a 10 out of 10 because you're proving that it works. But ease is where it's easier for Neil right now because he's putting the work over God, what, 24 years of doing business, right? Yes. And so you might as well start now. It's okay to say, even if you start now, you compound over the next five, 10 years or so, and you start to execute this over and over, you're gonna do a lot better than your competitors. Yeah, and to clarify on one thing, actually two things ease I would probably rate this as like a one out of 10 or two out of 10 max I think it just takes a lot of years before it produces ROI And the second thing that I want to clarify here is when I ask for the introduction, I don't just hit up people asking for the introduction. So let's say if it's a friend, let's say it's Eric. And let's say he knows the CMO at Apple and making this up, for example. I would say, hey, Eric, I see that you're connected with so-and-so that the CEO of Apple, do you know him well? What's your relationship? And then I start diving in with Eric and I'll just ask some questions. I'm like, hey, do they respect you? Do you know him well? Like, what do they think about you, your marketing strategies, all the stuff you put out? This will give me a really good understanding that if Eric does the introduction, that it will help or it'll hurt or do nothing. Because if it's going to do nothing, I don't care for Eric to make the introduction. If it's going to hurt, of course, I don't want it. Because some people are connected with others on LinkedIn and they don't know him like, oh, they used to like me. They don't like me now. And you'll be shocked. That does happen. But the key is like, if Eric says like, oh, I know him really well. I'm like, oh, how well do you talk to him often? And Eric may say, yeah, I talk to him once every month and I've helped him actually solve some marketing problems and I've given him four or five hires and they really respect me in these areas. So yeah, I have an amazing relationship with him. I was like, cool. Can you do this intro? This is what I did. Here's a public case study of appris competitors they spent you know four million dollars uh with us and look what i wouldn't probably disclose the price but i'm like they spent a lot of money with us here were the results i want to do this for apple and you know i have an opening because we stopped working with them i can now work with apple let me know if they're interested or i'm just going to go back to the other company and see if they're interested and i'm not trying to play the companies off of each other i'm just letting them know like i either can do one or other and maybe neither uh any of the companies want to work with me in the future and that's fine as well so i'm not going in being like if you don't sign up with me i'm going to go to your competitor i'm not trying to do that it's just more so if our contract's large enough sometimes people make you sign a nda or no uh not nda uh what is it uh non-compete not non-compete but you can't work with competitors I forgot what the exact term is but in the context it's like you can't work with X, Y, and Z competitors but yeah, it's just like that's an example of a strategy that I found that to work well I think it's good so a couple of things here I have two things so one, those of you listening to this right now take this segment take the transcript from this if you're watching this on YouTube go to youtubetotranscript.com and take the strategies that we're talking about that sound interesting to you and figure out how you can identify these what Neil's talking about right now if you have some of these relationships what's my play? look at all my LinkedIn connections. Here's the export of my LinkedIn. How can you do this? Oftentimes, Neil, because my open claw has all my LinkedIn connections in it. I'm just like, hey, who do I know here? Hey, I'm going to Mexico over here. I'm going to Mexico City. Who do you know over here that's an entrepreneur? Who should I, because I'm pretty good at setting up dinners and setting up these small events. And so I'll do that, right? So you got to play to your strengths. Here's a question, Neil. Theoretically, this can never stop because you're always going to want to grow more. So when are you going to stop doing this? I need to slow down because I think it's affecting my family in a negative way. Like my kids are starting to cry being like, I'm not involved. It's not even spending time because when I'm with them, I'm really like, I bought a crazy Formula One big Technics set. My kid likes big cars when my son does. He's not old enough to do it. So I'm sitting there doing, there's nine or 10 packs in there. I'm doing a pack a day and I do as much as I can with him. And then I finished the rest of the pack on my own. Technics is like a Lego or something? It's sold by Lego. So it was in a Lego store. It was like a big Formula One car. But he gets burned out after a while because he's four and it's meant for 18-year-olds. So I'll help him push pieces or he helps him push pieces. And then when he's done, I'll finish the rest of the pack. And then every day I'll do one. But it's not the quality time that they're looking for. I even do dinner dates with my kids, just like me and the kids by myself. Give wife a break. She doesn't ask for the break, but I want one-on-one time as well. But the kids are more upset that I'm not there for their school moments. That even though my parents weren't always there because they had to work and I understood it, I have the optionality to be there for my kids financially. My parents didn't. My parents was more so if they weren't there, it would be less money for food on the table. So they had no choice. I have the option. And that's where my kids are getting hurt because when I tell them, oh, but I'm working hard so you can have a pretty room and all this kind of stuff, my kids keep saying i don't need the room i don't need the toys i don't need any of this i want my dad so they're not wrong correct but here's also like one i'll say they're not wrong but that that's on you at the end of the day to decide what you want to do but also i think about my past as you're saying that i'm like did i really care how often my parents were there not really me neither not really i maybe the only time i cared was when they like my drumline competition my senior year because we're actually good that year. That's the only time I cared. But other than that, I don't really care because I enjoyed the work and the work at that time was playing video games. So that was the work for me. And you were already working. But we're all wired differently. So anyway, that was a real question. You can play this game infinitely if you want to. So anyway, that's a good place to end it. We will see you all tomorrow. Don't forget to rate, review, subscribe.