The Best AI Tools for Marketing in 2026

Does your experience with AI sound a little something like this? You've been prompting for 20 minutes. The output is polished, confident, and completely useless. That's what happens when AI doesn't know anything about your business.

HubSpot's AI works with your actual customer data. Every interaction and every signal your team has generated over time. So instead of prompting and hoping, you ask once and ship it. Check out HubSpot.com, the agentic customer platform for growing businesses.

What would you say would be the five latest AI tools that you have tested for marketing and give us feedback on all five. Start with your least favorite at number five and your favorite at number one.

I can't think of my least favorite because if I'm using it... Order it in any order you want. If I'm using it, I'm assuming it's useful. Okay, here, I'll tell you what. So, here.

So, I've been talking a lot about Hermes and Open Claw, right? What I've learned from both of these is this. They're beautiful, very useful if you use it in the right way. But if you overload it, it becomes useless and it degrades very quickly, right?

And so, I use... I would still say I love Hermes, I love Open Claw, and I use them a lot, right? Because... When you combine Hermes OpenClaw with your codex and your cloud code and they build with each other, you're able to make babies with it, right?

Like product babies with it. So let's give some real examples. Let's talk about a successful example and an overloaded example when you think you're overloading it and you're putting in too much or asking it to do too much and then you're not getting what you're looking for. So here's one example.

So my team actually made a creative agent called Picasso that will spin up creatives at scale. Add creatives? Add creatives, yeah. And a lot of people can do that, but ours is focused on, hey, what's actually performing?

What are your brand guidelines? And also what are your competitors doing? And we have a designer in the loop. And I was like, huh. I want your GitHub repo.

I told the team, I was like, give me the GitHub repo. I'm going to fork it. So forking it means I can take it for myself, right? So I took it for myself.

I dumped it into Codex, which is OpenAIs. And I said, hey, I want you to basically make this into... where I can send Neil cold emails or Noah cold emails and say, hey, Salesforce, we made some ads for you over here that are on brand. Here's three, would you like more, right?

And so that I built with Codex and then it told me to finish off the build using Hermes. And so that was a situation where that's something that we already have like a cold outbound infrastructure. It added it into the infrastructure, which is instantly and is able to send, okay? Like that's one example of it.

Where it fails, where all this stuff fails, open claw and Hermes fail because if you add too much to its disc or too much to its memory, it starts to, I can't find it anymore. Neil. I don't know.

I don't have that capability. Oh, but you do. It's like, Oh, I appreciate the confidence. I don't have it anymore. So you start to go back and forth with it.

It starts to become really unreliable. So I will say it's a love hate relationship with Hermes and open claw right now. Um, And Codex and Cloud Code have been amazing in that respect.

I will say what's overhyped, I think, is Higgsfield, right? The reason why I say Higgsfield is overhyped, not because the product sucks, it's because you run out of credits too quickly. And so it basically makes it useless if you can't really use it that much. Yeah.

And one quick thing when you're talking about, what is it called? Forking. What Eric's doing here is he's creating his own version and modifying it and making any changes that he wants. So that's why you would go get the code.

So that way you can just do whatever you want and go create your own version or expand upon it. Yep. And I'll kind of close it off with this. So obviously, okay, best AI tools, love, hate with OpenClaw, with Hermes.

I do love Codex and CloudCode. I think you need to pay for the $20, sorry, $200 a month version. Because here's the reason, Neil, I bet you don't know these numbers. So for $200 a month using the OpenAI 20X Pro version, how much usage do you think you're actually getting from it in value?

Don't cheat. Yeah, I already know the answer because I saw it on your sheet. Okay, so it's $14,000, right? Assuming you use it. Yes. Okay, and if you're using Claude, how much value do you think you're getting?

The Claude max $200 a month. How much do you think you're getting? Maybe $30,000, $40,000? No, $8,000. And here's the thing. Remember there was that time period where I spent $12,000 in that month, right?

So I've always felt open and I was very generous. I'm like, well, I can just keep going on this, right? And it did feel like I was spending $12,000 to $14,000. But Claude, you can tell it eats it up pretty quickly.

So what I would just say is maybe buy both if you have the means to do it. What's the pricing for both? $200,000. So you have $20 version, you have free, you have $20 version, and then $200.

So they both have $200. I thought it was $100. They do have $100 version, but there's also a $200 version for Cloud. There's a $100 version, but if you want 20x more usage or something like that, you get 5x or 20x.

But OpenAI, is it $200 or $100? It's $200. I thought I was paying $100. no you're paying you're definitely paying either 20 or 200 i'm pretty i'm not paying 20 yeah yeah it's you're paying 200 so i i think those stick out oh i do like uh i do i do love the hrefs mcp that's been great i do love the riverside mcp as well the beehive has an mcp riverside has an mcp what do you use a riverside mcp for

So Riverside actually has the ability to look at your, like, let's say we're recording this. I can say, hey, find the best moments to clip. So you can use their magic clipper in there. You can pull the files that you need.

You can pull transcripts from it. So I can just work in there. And then I like the Beehive one because you can just have it draft emails or you can have it. It's going to be able to, I said, I messaged a Beehive growth guy the other day.

I was like, hey man, your MCP needs to give me the ability. So whenever we're doing a webinar that we can make a special lead magnet and spin up a page immediately. And he's like, that's coming. I was like, you know what I mean?

Because you're doing webinars, you're doing different topics, right? But if you have a special resource for each one, that's great. And then you can tie it into your socials too. So those are just a couple of examples.

I would say using kind of the main models, the main autonomous agents, and then using MCPs from tools like this because I love Ahrefs too because their brand radar is really good as well. Yeah. Yeah.

And when it comes to marketing, one interesting thing that I've noticed is every time Claude, specifically Claude, or Anthropic, not Claude, specifically Anthropic, announces a new model, it calls a whole SaaS industry to tank. I'm talking about all the publicly traded stocks.

And what I've noticed, this is just my two cents, from everyone who I've talked to that really spends money with them, what kind of company do you think they are?

A SaaS company? SaaS or a tech company that has tons of engineers. So it's just like they could put out a notice saying that, hey, hey, guys, we don't actually think we're going to end up disrupting all these companies like you, the analysts, actually think.

The NVIDIA CEO has stated that a few times. And every time he states it, the stocks go up and then eventually they go back down. But Anthropic, from a marketing standpoint, I really haven't seen them really push out any narrative because the customers paying them are the ones, a lot of them are getting hurt.

And if those companies went away, Anthropic would hurt because they spend so much money on coding, right? Software companies, they need to have tons of engineers. That's the primary use case, literally the primary use case for paying a lot of these corporations like Anthropic.

But if they said that, I believe that would also hurt their valuation. So they don't say it. It would. I would also say one thing too. If you're Google right now, I would just try to squeeze them on costs, squeeze them on pricing because it squeezes the entire, it squeezes OpenEye and it squeezes Anthropic.

You know what I mean? No, they are. Yeah. But more. Yeah. Yes, you saw the announcement. It was OpenAI and maybe one other player they were talking today about potentially reducing pricing even more.

That's going to squeeze them all. Did you also see the winners from Anthropix IPO? No. from Anthropic IPO. There's a really cool chart. Are you in on any of these with these funds that you're in?

No comment. So, Anthropic, no. So if you look at the Anthropic winners, where is the chart? Anthropic, which companies will make the most? Let's see. Okay, Amazon holds a 21% stake.

Yeah, and Amazon owns a 21% stake. Google owns a 15% stake that can deliver hundreds of billions in value for both Amazon and Google. Microsoft backed OpenAI. Salesforce turned a $50 million in 2023 into a $5 billion anthropic stake ahead of the IPO.

But a lot of these old school companies are the real winners for, From this, how the heck do you fund all this stuff with all these billions and billions in expenditure? You need the big players to fund it all. So even if they have competition, they're making money from the competition because they own the largest chunk of the competition.

Like, look at how much the Anthropic founders actually own of the company. Have you seen their percentage ownership? No. Anthropic founders ownership percentage.

It's tiny. Not in dollar amount, but percentage. Each of the seven founders of Anthropic individually owns less than 1% of the company's total equity.

Dude. Yeah, Google owns 15%, and Amazon owns over 20%. How much did Sam Altman have had? I think he had a good chunk, right? I bet you he's still going to get paid.

Sorry, not Sam Altman. Sam Beekman-Fried. A large chunk. I believe he owned like 10% or something like that back then. Man, they say he was the... one of the best investors because he did that and he did a cursor and there was a few he did a lot of good ones you know robin hood yeah he owned nine percent i think when they were worth like 10 billion dollars now robin hood blew up they went down yeah uh from their peak but they're still an 83 billion dollar company but they're worth more than that dude call it you know less than a year ago the stock was at 140 now it's at 92 i showed you marvel it jumped it fell down to 245 yesterday it's up to 280 again

Do these things jump? Have you been trading them? No. I'm not a trader. The only thing I'm going to trade, Neil, is I'm going to bet on the Spurs to win it all.

Did you see the game where they're down by 20-something? They're up by 29. So I think they're still going to win. So we'll see what happens. You think the Spurs are going to win?

I think they're still going to win. So I'm going to put some money on that just for fun, right? Because if you put $1 in, you get $3.50 back in profit if they win the next three games. So what are you going to bet on that?

I'm not going to bet a lot. I'm going to bet like, I don't know, four to five grand. Oh, okay. Yeah, so we'll see. That's too risky of a bet for me. I think the Knicks are going to win.

Put a grand. I don't want to lose a grand. What do you mean? You could get $3,500 in profit. But I believe the Spurs are going to lose. No, they're going to win.

Dude, the two games that they lost, and we can move on from this, but the two games that they lost, they lost by one, and they're both boneheaded. Boneheaded mistakes that they made. So I still believe they're going to lose. They are very likely going to lose, but I still think they can win.

Okay, if the Knicks win... You buy us lunch the next time we all eat out. And if the Spurs win, I'll buy you guys lunch. No, no, no. If the Spurs win the series, I'll buy you guys, Erwan.

I still have a bunch of those jars. So, you know, I was talking to our mutual friend yesterday, Yaniv, and we're like, oh, they're going to win. Because at that time, he's like, oh, the game's over. He went to go play tennis, right?

And then he came back, they lost. And he's like, now's the time to vet. so i'm like yep you're right i'll do it but yeah i i probably won't even put in that much i'll put in like a grand okay put in a hundred dollars a grand a grand makes it more exciting you don't even get me started on you okay all right i don't bet you did before all right so we're not really gambling though you used that that was gambling are you talking about stocks you're talking about the casino no sports

No, I made a killing. But that's still gambling. Someone, Jay Fenn, his buddy had this whole math or formula on sports betting. Years of track record.

I didn't even know half the teams. I just followed the bets and I was making money on it. It was great returns. Yeah, but that's still gambling. It's still gambling, but I call that a calculated decision.

Isn't that just business? Yeah. Quick break. If you want to run personalized LinkedIn ads and have personalized landing pages to convert your customers at a much higher rate on LinkedIn, check out Carrot.

That's K-A-R-R-O-T dot A-I. Carrot allows you to do things that LinkedIn ads does not allow you to do right now. Check it out. There are publicly traded companies such as SEMrush and Sitecore that are using this right now.

And you can use this to get ahead because if you don't use this, it will take you weeks, if not even a little longer than that to make these ads. Because again, you cannot do this on LinkedIn. So again, go to www.karrot.ai and we'll see you on the other side. Okay, so I want to go into these costs real quick too.

So we kind of briefly touch upon this, but Neil, here, let's just do this. The real cost of a ChatGPT Pro and Cloud Max subscription is this, okay? So if you're paying $200 a month for ChatGPT, I just want to show the chart over here. So it's $14,000, okay?

$14,000 in what? In value. So I'll go over that in a second. And then if you pay for Cloud Max, which is $200 a month, you get $8,000 in value. So here's what I mean by that.

When I spent that 12 grand on Cloud API costs, right? Like over here, let's say... I felt like I was getting this much in value because when I'm using chat GPT, it feels like I can just keep going. I feel like I'm cheating.

I'm like, oh, this is too good. I'm like, why can I just keep going on it, right? And it does feel like I can kind of go to distance on this 14,000. And I do actually run out every week, okay?

Now, when I'm on CloudMax and I'm just using the subscription, I run out so fast. And this is about like a little more than half right here. What do you run out doing?

building stuff. Like I have my agent doing a lot of, my agent's like, I'm having to do research. So for example, today I was doing a podcast interview, right? This one doesn't take a lot of tokens, but I'm like, what does this guy talk about?

Do the research over here. Oh, find me candidates over here too. So it's sourcing candidates too. Yeah, so it's sourcing candidates and it comes up with good angles to reach out to them personally.

But everyone's like, oh my God, thank you so much, Eric, for doing research on me. I would love to have the conversation, right? And it's like, how did you find me, Eric? I was like, I never found you.

And you don't want to just use There are versions like the Cloud Coworks and all that kind of stuff that just help you with everything. That is using Cloud Cowork. But Cloud Cowork takes from my Cloud Max the $200 a month.

Oh, so it's pulling from there. Yeah, that one's pulling from there, that Cloud Cowork one. I thought Cloud Cowork, like my team uses Cloud Cowork a lot. Yeah.

I thought it just takes from the API. No, no. So it takes it from their $200 account. Yeah. And just so you know, like the way I have it set up for my team is sure they have their subscription, but then there's overages and you can start to see who's using on overages too.

But my point of saying this is that, okay, semi analysis, which they do a lot of research here. Like if you're paying for Cloud Pro for $20 a month, it's worth $400 a month. ChatGPT plus $20, you get 700. So ChatGPT is actually far more generous with their subsidies and this is not going to last forever.

So I think if you're trying to go hard on it right now, you might as well pick too much. No, I think they're going to go more hard. more subsidies yes the reason i say that is once they go public and their shares are liquid and people start selling they want you to they want you to come over not just that let's say if i'm google and i make 100 billion dollars from one of these companies and i spent 10 i'm like oh we got 90 in profit let's just use the 90 billion dollars and just make our stuff even cheaper and just and cut them down now and cut them down even more yeah Yep.

Yep. So I just think it's interesting. So semi analysis, like they do really deep, really deep research here. So that brings me to the next point over here.

So if you look at what the founder of ramp said, so he said this, right? So, you know, check this out. Basically, basic AI tasks are 700 times cheaper, whereas frontier AI tasks are 300 times more expensive.

So this is Eric Lyman from the CEO of ramp. So when you look at this right here, right, you look at the routing opportunity. So talking about token routing over here. So when you see,

So Frontier Intelligence, bigger models, more reasoning, like Fable 5 right now, it's 300 times more expensive per task, right? So if you're using chat cheapity for level intelligence, again, if I'm cleaning your toilet or whatever, I don't need the best intelligence, right? So you get lower intelligence, same task, smarter routing. It's 700 times more cheaper per task.

So you need to have good token routing. And I'm in some of these CEO chats right now where some of these people work at large enterprise companies. And they're like, yeah, this is a major point of control. tension right now.

And you're seeing this, like people come to us saying, hey, how can you help us with token optimization now? I'm sure you're seeing some of that too. In one of our software divisions, we are on track to missing profitability numbers. And you want to guess why?

Too much on tokens. Yeah, too much on tokens. How much on tokens? A couple million? That's overage? Yeah. No, the overage is, if I recall, it's close to 500 grand.

But that's after deducting efficiencies and hiring less. So if we remove some of the expenses we had, we didn't spend as much on marketing. We didn't spend as much on hiring.

Some people got terminated. Some people churned. If you include that, plus us spending more on tokens, right? And we spent substantially more on tokens, the delta and where we're off, even hitting revenue numbers, is getting close to around $500,000.

So that means we spent over 500 grand more than what we budgeted on tokens because I'm including the savings from the employees, marketing, and all of that. So the tokens ate that up, that expenditure, as well as more. Significantly more. Significantly more, yeah.

So anyway... Which is becoming a big problem in a lot of companies. And it's just like, no, dude, give me something that's 90% there that's like one-tenth the cost. I'm okay if it takes a little bit longer.

And so someone on your team, I'm sure, is already figuring out buying local infrastructure and then the open routing and all that stuff. Yes. I started seeing bills for computers with heavy RAM and all this kind of stuff. And I was like, all right.

Are you buying the Dell ones? I like everything comes back. I'm like, man, I haven't bought an NVIDIA thing for a long time or a Dell thing, but those companies are back. Like the Dell ones are huge now.

They have a ton of memory. You see their stock, their stock bill. Boom. Yeah. Because of the server division. Yeah, we're dummies for not buying. I saw it coming.

Like all these things, that list of stocks I sent you, I was looking at them for months. Yes. And Dell from that perspective is more, I would call, corporate friendly from a computer and server standpoint than Apple. Because everyone's talking about Apple and, oh, they're all sold out of high RAM computers.

It's easier to just go to Dell to get those. Dude, Apple gypped me. I bought a new monitor and then the camera's all broken now. And I'm past the return period.

So I don't even know what to do. Yeah. You didn't notice right when you got it? I didn't turn that camera on for weeks. Ah. Yeah. So anyway, I want to talk about something, Neil.

So I want to talk about the difference between selling AI services right now versus marketing services because I'll jump in on the marketing services calls and I'll also jump in on the AI services one because we're pushing the single brain stuff for it and all these pilots. So what I've noticed, Neil...

on these AI services calls is it's collaboration. It's like when you talk about building this brain, you talk about building these AI things for the clients. They're like, it's always like, what else can you do? Can we do this?

Can we do this? Can we do this? Now, when I look at the agency side, it's like, oh, it's much more of like you're, we're talking to, we know how your business model works. There's a bunch of different agencies we're talking to.

You're like a vendor, right? Yeah. And when you move into kind of the client relationship, it's like, what have you done for me recently? Right. And then with agency, I hate to say this, you can disagree with me here, but the problem is people have to kind of get on all these calls with agencies all the time because they feel like they might have to push the agency.

And I'm saying for most agencies, maybe not for yours. Right. And so you're a little bit of like, you know, they have to, not that they want to get on calls with you, but they feel like that might need to happen sometimes. Right.

And I'm not saying that that is for most cases, but that does happen. Right. No. Again, when we do these pilots on the AI services side, once we show them the light, it's like, oh, well, can you do like five more things over here?

Every single week, it's can you do more? Can you do more? Can you do more? And I'm like, well, if that keeps happening, then... Might as well just focus on hiring more engineers and just more AI-pilled people.

Because on one side, things are moving so quickly and the market's pulling. And on the other side, the market's still definitely pulling, but it's not as engaged as the other. And then one more thing I'll say about marketing services. We do have legacy clients where they don't want you to use AI, right?

Because they don't understand it too, right? They don't want you to use AI. They want to get on calls with you all the time. And that's maybe because they're more old school.

The new clients that are buying AI services, they don't want the calls all the time. They only want to talk about collaboration and they only want to talk about new ideas and they're down to test a lot more. So it's just interesting to kind of see both right now because I'm like jumping in, you know, between the two, so. Well, I think with more of your single brain division or company, because it's something new, not just for you, but a new concept for companies, they're much more collaborative because they've never done it before.

and they're curious on your thoughts, and then they're giving you their feedback. But on the agency end, most of the companies you deal with have hired agencies and worked with them in the past for many, many, many years, some even well over 10 plus years. They already know what they want, what they're looking for, and they're just very direct. You know, it's funny.

I've gotten on some of these calls with some of these. I was on a call with like a founder this week and he had founded two companies that work with single grain right now. So I just happened to get on that call because I thought he had a really interesting product because, you know, I have an allergy to mosquitoes. Right.

So he founded a new mosquito product that's super like, so I want to get on the call to learn more about the product. Right. And he's like, you know what, Eric?

I really hate agencies, right? I really. And then he had two people standing behind us like, we hate agencies too. I'm like, oh, interesting. Why? It's because of the reasons we talked about, right?

Like because they've had so many jaded experiences with agencies. So let's go back to the mosquito thing. What was the solution? Was it another heat pen?

No, no. The heating pen is actually not good for you. You also used to use, right? Yeah. So I used to use, so I have this thing called Skeeter syndrome after the pandemic.

So after I got the, you know, the little shot. So I'd get bitten and then this like giant pus filled dragon ball would happen on wherever I had the bite. And so our mutual friend Yaniv was like, oh, you should use the heat pen, right? Not saying he's wrong.

I was like, oh yeah, okay, makes sense. But, what you want to do when you get the bite is actually you want to put ice on it and then you want to take like a like a benadryl or something and then you put some cortisone on it and it doesn't happen anymore so that's how you get rid of skeeter syndrome his product is called shoe bug which is like it's shoe bug right and it's like organic i think and uh it doesn't smell it's like all natural compounds it's like this is good um so you can check it out if you want you don't have a mosquito problem they tend to like me more uh we'll leave it at that thank you all for listening and we will see you all tomorrow