Ep 613: AI Is a Stack of Two-by-Fours. What Are You Building With It? (Plus Meet Gary and Blanche)

22 May 2026 · 25 min · 12 chapters

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

The episode argues that AI is “raw lumber” (e.g., a stack of two-by-fours): by itself it’s not a finished product, but it becomes valuable when humans build systems and outputs with it. Guest Brayden and co-founder Jeff discuss integrating Claude into D2C workflows for funnel congruency and faster experimentation. Key examples: using Claude Projects with Motion to analyze top ads and landing-page mismatches, then generate a new PDP in HTML; running page-speed tests via Claude-in-Chrome, then optimizing pixels/file sizes to improve mobile speed (client PDP ~7% vs improved ~88% mobile) and achieving strong SEO/accessibility. They also connect Claude to Microsoft Clarity heatmaps/recordings and A/B testing via APIs to generate V2s and split-test.

Guests

Brayden (resident AI enthusiast) and Jeff (co-founder; builds AI/D2C agency systems). They also mention internal AI team members Gary (media buyer) and Blanche (creative strategist) running Meta account analysis via Slack.

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Chapters

Tap a time to open that second in VO

Understanding AI as Raw Material

0:30 to 4:19

Explore how AI is a foundational tool that needs creativity to be valuable.

“The minute you start going deeper and learning how to start building and crafting things and making stuff, that's where the value starts to come from.”

Integrating AI with Project Management

4:19 to 7:11

Discover how integrating AI tools can optimize landing pages and ads.

“So what's been the most useful integration of AI into an all killer, no filler product?”

Brand Sensitivity and AI-Generated Pages

7:11 to 9:09

Discuss the challenges of using AI for design-sensitive clients.

“It's like 100 % SEO, 100 % accessibility, 100%.”

Leveraging Data Analysis with AI

9:09 to 11:03

Learn how AI can analyze data efficiently for marketing strategies.

“And it's like kind of up to the human in the loop to have the taste to be like, okay, like we should put this video in the hero and we should have this video down here or this image here.”

Introducing AI Team Members: Gary and Blanche

11:03 to 14:03

Meet the AI tools being used for media buying and creative strategy.

“Go check out this for the split test data.”

Creative System Updates

14:03 to 14:35

Explore the updates on their creative system and its integration into communications.

“You know, right now it only answers to Jeff.”

Exploring Claude Design

14:38 to 16:01

Learn about the capabilities and early experiences with Claude Design tool.

“Jeff, I'm sure Jeff will be doing some lunch and learns about it, but it's, yeah, it's interesting research.”

Introduction to Higgsfield

16:03 to 17:50

Discuss the features and capabilities of Higgsfield for creative projects.

“where you can just go into Figma and download a Figma file and then upload it to Cloud Design.”

Connecting Claude and Higgsfield

17:51 to 19:16

Understand how Claude can generate assets through Higgsfield integration.

“So now you can have Claude go and generate assets for you within Higgsfield.”

Leveraging AI for Creativity

19:16 to 19:52

Discover the potential of AI in enhancing creative workflows and tasks.

“But one caveat is just that you can blast through a lot of tokens pretty quickly.”
Show all 12 chapters

Teaching AI like an Employee

19:52 to 22:26

Learn the importance of training AI systems as you would a human employee.

“Great kudos to the branding team at Higgsfield because the Higgs field is the field that CERN discovered, which is the reason why any of us have any mass at all.”

AI Limitations and Punctuation

22:26 to 23:30

Discuss the limitations of AI in writing and the evolution of language usage.

“And you really have to make sure you take out those M dashes.”
Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:00Before we jump into today's all killer, no filler episode, a quick word about who makes this show possible. The D2C podcast is brought to you by Pilothouse, the performance agency behind some of the fastest growing D2C brands in the world. Creative, media and customer journey all under one roof. Performance and brand without the tradeoff. Every Friday, we hand the mic to a Pilothouse operator to break down what's actually working in their space right now. Want a team that treats your growth like their own? That's Pilothouse. Head to Pilothouse.co and now on with the show. The minute you start going deeper and learning how to start building and crafting things and making stuff, that's where the value starts to come from.

0:37I have Claude in my Chrome extension and I was like, look at this page speed test results, go through everything and come back and tell me what needs to be optimized. Higgsfield is really interesting. The cool thing that they've recently come up with is another connector to Claude. So now you can have Claude go and generate assets for you within Higgsfield. So then all of a sudden, all these tools are starting to connect into the central database. This is like a real secret. This is some valuable stuff.

1:30Triple Whale centralizes all your data, equips you with measurement tools you can actually trust, and layers in best-in-class AI to turn that data into clear recommendations, showing you what's driving growth, what's hurting profit, and exactly what to do next. It even takes action for you. With Triple Whale, you can quickly spot trends, catch issues before they impact revenue, and uncover opportunities you would have otherwise missed. It's about moving faster, making smarter decisions, and operating with real confidence. That's why brands like True Classic, Whey, Press Juicery, and thousands of the fastest-growing D2C companies rely on Triple Whale every day.

2:06They're saving time, increasing efficiency, and unlocking insights that simply weren't possible before. If you're ready to stop guessing and start growing with clarity, now's the time to check it out. Visit triplewhale.com, that's T-R-I-P-L-E-W-H-A-L-E dot com, to learn more. It's All Killer No Filler, and I am here with resident AI fanatic, Brayden. Welcome back to All Killer No Filler. I wanted to start today's podcast, I just was asking my co-founder, Jeff, because he is doing some really interesting things with AI, with D2C and agency, the company we're building. And I was asking him for his thoughts before I jumped in this podcast.

2:44And he gave me this beautiful, almost poem that I want to start this off. Jeff said, someone handed me an analogy. So someone else's analogy. This is a third and fourth hand analogy, but I think it's really good. AI right now is like a giant stack of two by fours that everyone's been handed. Free wood stacked in your yard. Take what you need. But the thing about a stack of two by fours is by itself, you can't really use it for much. Sure, you can lean on it for a minute, but it's not a chair. It's not a table. The value of two by four shows up when someone builds something with them. One person could frame a house with it.

3:13Somebody else with the right foam and fabric builds a beautiful sofa. Someone else a tree fort for their kid. Same wood, completely different outcomes. Nobody frames a house exactly the same way. Nobody builds a sofa the same way either. The lumber is identical. But how you use the lumber is what actually makes it useful. That's where we're at with AI right now. AI is a raw input, not a finished output. I think that's worth repeating. But what you build with it is what makes it different from the next person who got the exact same shipment, which is really interesting. I think it's so tempting to use it in its raw format at times because it's so polished and so good and so persuasive.

3:47but I think it's a really great reminder that you got to keep that human in the loop. Definitely. And that like, you can't really just be stacking or rearranging the two by fours by not really going deeper. But like the minute you start going deeper and learning how to start building and crafting things and making stuff like Jeff said, that's where the value starts to come from. Build a whole house, build a whole ecosystem. And so that's what we're talking about today a little bit are all the exciting ways that you are integrating, I think Claude specifically, into some of the other tools that make this industry tick.

4:19So what's been the most useful integration of AI into an all killer, no filler product? There's a great story. So it all kind of started happening once I figured out projects. So within Claude, you can create these project files. And I started creating project files for each one of my clients. And some of my clients have Motion, which is a great platform. And one day I was poking around Motion and I saw something within Motion that said, like, connect your Claude. And I was like, oh, yes, I want to do that. So I quickly asked the client if that's okay. They said, absolutely. And there's like no better feeling than just like clicking on Opus 4.7 and being like, go check out Motion and come back and tell me what you think.

4:56And so that was a real opener for me because all of a sudden it came back with a lot of valuable information. One of which was it built a report on, hey, here are your top ads and these are the landing pages that they should be going to. And it was able to like analyze and see like there's a mismatch between the messages that your ads are saying and the landing pages that they're going to. A lot of these ads were going to landing pages that were like PDPs or collections pages and stuff like that. So there was like an immediate like light bulb that went off and the conversation kind of, you know, with me and Claude just kind of evolved to this place of like, oh, well, like what would the best ideal landing page be?

5:33Or what were, I think it started with, what are the top three landing pages that we could wireframe and make? And then it evolved to like, okay, well, can you just build it with HTML? And then it's just like all of a sudden it's like, I've got this page that is built. Obviously, I had to understand how the folders work within project files and so that you can drag in assets and stuff. But I was dragging in full-size images, and it was breaking them down into smaller pieces to optimize for page speed, load times, and stuff like that. Just amazing. So all of a sudden, it's like I've got this PDP.

6:04It's analyzed the current PDP. It's made this new PDP. I'm pulling the listing images from Amazon. Have you ever noticed how Amazon listing images, they do a great job of like selling you the product when you like swipe through the thumbnail pictures or the carousel so it's like i i so incorporated that have some great like videos from tiktok added them in like unreal so i created this page and i threw it into my chrome browser i got it uh launched and i threw it into my chrome browser and i was like oh let's run a page speed test so i run a page speed test right and usually they come back like when we're like manually building pages like with with other platforms they come back and they're not great this one was okay but I like have Claude in my Chrome extension.

6:46And I was like, look at this page speed test results, go through everything and like come back and tell me what needs to be optimized. And then I took all that information and gave it back to Claude. And then Claude goes in and just starts like, you know, moving the pixels around so that they load at a certain time so that it's like not slowing down your page, like your initial paint. It's like shrinking files down. It's doing all this stuff. So anyways, I put the page back in and it's like phenomenal page speed. It's like 100 % SEO, 100 % accessibility, 100%. I think the best I could get was like 88 % on mobile for page speed, but it was like far better.

7:22And actually testing the client's PDP, it was like 7 % for page speed, like really, really bad. So what does Instapage think about this? It's tricky. Like, I think the thing that I'm finding is like, it doesn't do a great job. If you've got a client or if you have a brand that's very picky on design. I was going to ask about the branding and the design. It's not going to do a great job. Like the pages look HTML. They look good and they use the brand font and they use like the brand colors and stuff. But it's very HTML. It's like there's like another step that I feel like Instapage and some of these other page builders can do where like, you know, like they can have like more on brand design.

8:04So that's like the one caveat. If you're an extremely brand sensitive client, maybe these pages aren't great. Or they can serve as like a learning opportunity where it's like, hey, let's just launch this page for a short period of time, get some page data back to see if this is outperforming our PDP, and then take those learnings and then integrate into the PDP. But I mean, it took me about six hours to just like vibe code this page. And that would have taken a week. A hundred percent. You can get that down even in the future as you build it more. Like we've been talking about altering the landing page for having funnel congruency since this agency began and long before.

8:45But it appears to be a real, it has, I'm hearing it more and more in the ether of people really understanding that you've got, you know, in this age of Andromeda, you've got all these different customer archetypes you're trying to tell stories to and, you know, fill out the picture of the product experience. yeah that you want to be able to have funnel congruency and actually have landing pages for them and and so the ability that you can you know have spin up pdps and different landing pages so quickly is going to be beneficial the question is how long until it's all in the algorithm in a way right where it's like landing pages can be spun up on the fly for certain people at certain parts of the journey sort of automatically i mean at that point you still have the issue with assets until that kind of is fixed where, you know, like I said, clients are giving us just like a slew of stuff.

9:31And it's like kind of up to the human in the loop to have the taste to be like, okay, like we should put this video in the hero and we should have this video down here or this image here. And I feel like if you can, I guess, give all the raw materials to it and you can trust it, then like that's probably a possibility where like page, even from like a copy perspective, if it's just like switching the headlines around or copy or just like testing it like like maybe meta is just testing that would be really cool but i don't know i don't know how far along that is but and then the other cool thing is that there's a there's a clod connector this is like a real secret this is like this is some valuable stuff there's a clod connector called microsoft clarity and it's free and you can you it's free heat maps and recordings for your pages so you can put microsoft clarity in And then you can connect the clod to Microsoft Clarity via API.

10:21And so that then the same clod that built the page can then go in and analyze all the page data, the heat map data, the attention data, the scroll data, the scroll depth data, the heat, like the actual recordings. It can like Microsoft Clarity is a part of Copilot. So it's like doing its whole AI thing. and so then you can then get it to go there and you can come back and you can say okay based on all the data you have now let's create a v2 and then let's split test this v2 and so then all of a sudden you know that was also a bottleneck for having just like large quantities of pages was in the interpretation of the data like for me to go in and look at a heat map for this page and then like you know um take the time to like digest it and then like understand where the next steps are going to be was like it's a lot so now all of a sudden you've got the Claude that's connected to the motion that's looking at the motion to create the pages to the pages to the it's like so cool man and then to do the data and then coming back and then i mean i don't know if there's any like split testing software out there that has a intelligence oh really yeah intelligence we're we're working with them potentially on an upcoming thing we're doing but they're they're great and they yeah they they specifically do they're an ab testing platform that's their whole thing and i think uh so then you can do api calls from Claude to that then all of a sudden you've got like, hey, Claude, go check out Clarity for the heat map data.

11:40Go check out this for the split test data. The other thing that's not really connected is Meta, but maybe there's still opportunity there. I know that they've kind of been changing things. I think Meta did open up their API for Claude, but I haven't looked too far into that yet. Speaking of which, we sure have on the D2C side, my co-founder has been absolutely skunk worksing on Claude's connection to Meta. And it's early days, but we have our first AI team members, Claude, sorry, not Claude, obviously Claude, but their names are Gary and Blanche. So Gary is a media buyer and Blanche is a creative strategist.

12:15And, you know, we're just giving them knowledge bases. What's interesting is it's like people debate all the time about best practices and I'm not, when it comes to Andromeda and what you're supposed to do and the gem and the lattice and how, okay, this does that. So you're supposed to, and so there's always an online discourse about it. It's interesting just to talk with Gary and Blanche about like why it's doing what it's doing, like really understand its logic based on analyzing creatives. And it's really, you know, we have our creative analysts telling us which of these are too much the same and which ones we should be testing.

12:47And it's quite an interesting experiment to run. And we're getting so far fairly comparable results than we were getting manually. It's an interesting one for sure. It's research, right? Like as an organization, it's just, I think it's just something that behooves you to check out and figure out because even if you don't know how to do it, you just ask Claude how to do it. And it tells you how to do it step by step. You have no excuse. You don't. You just take screenshots and it's like, what do I do next? I was trying to figure out how to launch these pages with Cloudflare. And it was, I was just like, oh, I just like took a screenshot and plugged it in.

13:20And it was like, oh, that's what you got to do now. It's just, it explains it to you. So if you ever get stuck, just screenshots. That was my number one. It's funny that like, it's the most fundamental thing about launching a website, but it's all those like back end things that I just didn't have any idea how to do. And they'd always stopped me in certain ways, but it's just like, Oh yeah, it's super simple. And you know exactly what to do being told step-by-step. So are you, I'm just curious, is he using claw? Is it claw bot or like open claw or is it more clawed desktop? It's clawed desktop running on its own laptop.

13:50Basically multiple laptops are running our creative strategists and our, and our media buyer at this point. Yeah. But it's just giving us every 15 minutes, we're getting an update where we've just got it all, you know, or it's looking to update every 15 minutes. It gives us two daily updates. You know, right now it only answers to Jeff. He won't open it up. He doesn't want to muddy the water. So we can't talk to Gary and Blanche yet. But yeah, we've got a system now where we're just sort of dumping creative into a Slack channel. It's being analyzed and then it determines whether it goes into the...

14:18It's giving us an opportunity to really just set things back to basics in a lot of ways. Just like pick our best conversion action, run as broad as we can, have a testing campaign, have a scaling campaign and have like clear rules for when one goes to the other. So far, so good. Yeah. So cool. So cool. I need to get on a separate call with you guys and talk more about that. Yeah. Jeff, I'm sure Jeff will be doing some lunch and learns about it, but it's, yeah, it's interesting research. What other integrations are you having some success with? Well, I've been playing around with Claude Design.

14:48So Cloud Design kind of came out since the last time we talked, and I've been playing with it a little bit. I think I'm still in the infancy phases of kind of like getting valuable outputs from it. But I have played around with it and tried to create some design systems. And it's pretty cool for those people who haven't played with it or tried it. You can go to your Cloud account in your Chrome browser, and there's a little design menu that you can click on. And what you can do is you can like upload Figma files. You can upload raw assets. So I experimented with uploading some emails. So I just went on to the email team.

15:22I was like, give me like your top performing emails. And I brought them in. And then I created a design system based on those. Because those are really nicely made, really well designed, stuff like that. And then was able to then, you know, it takes you through the whole thing. It's like, does this look good for, you know, bubbles? Does this look good for CTAs? Is this how you want your headlines? Is this how you want your sub headlines? Is this how we're going to do the backgrounds? Are these the colors you want to use? It goes through all that. and you can, you know, you approve or deny until it kind of like works through everything.

15:50And then it creates this design system. And then from there, you can just start creating emails, creating whatever you've like used in this design system. So pretty amazing. And it's pretty cool with like the connection with Figma where you can just go into Figma and download a Figma file and then upload it to Cloud Design. So like I said, not something I've like really gone too deep with, but I can see it as like a really fast prototyper or like a last minute, like if you need something last minute, just to have your client's design systems all kind of set up there. And then I think for me, what I'd love to do is, like I was saying, the HTML sites that Cloud's building, it would be cool to build them with Cloud Design so that then all of a sudden, like you can start implementing those more branded clients that are more brand sensitive.

16:36You can start actually like using Cloud Design to run those. So yeah, that's been another one that's been really interesting over the last little bit. Did you mention Higgsfield there? I just, I missed. Did you mention Higgsfield or is that a separate one? That's a separate one. That's a separate one. So have you heard of Higgsfield before? I haven't, no. So Higgsfield is like, it's really like a hub for creatives. It basically is like, you can go on like VO3 and make videos or you could go to Kling and make videos. But Higgsfield has all the models within it. So you can like create, you can use the new Seat Dance model.

17:09You can use Nano Banana Pro. you can create 4k images you know for a long time i was just using photoshop and using generative fill with gemini nano banana inside of photoshop but you don't get high resolution images and it's it's really clunky to use but then hicksfield is a little bit better at at um at all that stuff so you can i mean they have like a marketing section where you can create like ai avatars for ugc and stuff like that. Create UGC videos. So really cool. You know, put the product in, put the image of like a UGC creator and then get like AI, you know, an AI version of an influencer holding your product and stuff like that.

17:48So Higgsfield is really interesting. But the cool thing that they've recently come up with is another connector to Claude. So now you can have Claude go and generate assets for you within Higgsfield. So you can be a creative strategist where you're saying you've got your project file that understands your client. It's got all the branding information. You can put in ads, you can put in old ads, you can connect it to Motion and stuff like that, right? Then you can have this Higgs field connector. So then all of a sudden, all these tools are starting to connect into the central database. And then you can say, okay, now I've gone and looked at Motion at my top ads.

18:19Now I want you to go to Higgs field and I want you to create 10 raw assets, not ads, but just raw 4k assets with Nano Banana Pro for all the different avatars that we're trying to reach with Andromeda or whatever. And it will go and it will prompt all the, you know, all the image prompts. And it'll have like, you can give it, you can give Claude the PNG of your product image. And it will go and it'll communicate with Higgsfield and then it'll drop it in a folder on your computer. And you can open it up and it'll have all these images in there. So like a really amazing connector. And then you can take those images and then you can have different workflows where, you know, you, it will then create big scripts and it will connect so you can storyboard things with video and stuff like that.

19:00It's really quite amazing. There's a lot of great videos on YouTube about it, but I think that's the next thing that I'm kind of like leaning into a little bit is seeing how you can leverage the connection between Claude, your project folders, and your clients and Higgsfield. But one caveat is just that you can blast through a lot of tokens pretty quickly. So, and if you don't know how to use Higgsfield first and you start prompting Claude to go into Higgsfield and do stuff, you can run out of credits, I think, quite quickly. So you kind of want to understand how the workflow in Higgsfield works, you know, what you can get, what you can do, the video side, the image generation side.

19:37And then once you kind of understand how it works, then you can start like prompting Claude. And then there's a lot of great guys on YouTube who have actual like skills built. So like it'll keep your Claude at least kind of contained and not go all out wild on generating stuff. Super cool. Great kudos to the branding team at Higgsfield because the Higgs field is the field that CERN discovered, which is the reason why any of us have any mass at all. It's the energy that connects the entire universe, supposedly. It's the reason that we have mass instead of nothing when our atoms are actually 99.999 % nothing.

20:12Wow. Yeah. Very cool. We learned some abstract physics today. Very cool. We can close this with the closing part of Jeff's speech. I didn't realize this also. This actually went in our agency newsletter, our agency community newsletter that you can actually sign up for soon on the agency, joint agency website, but it's a good way to end it. We're talking about Gary and Blanche, our two AI media buyers and creative analysts. They live in a Slack channel and talk to each other. They actively run one of our internal meta accounts for DTC newsletter. It's early, but it's working well. The fascinating thing is that out of the box, it could do things, but it didn't really run the account like I would want it to.

20:49So almost all of my time has been spent distilling that about how Pilot House Media buys, training that into Gary and Blanche. That's been the hard part. If you're expecting AI to just work out of the box for your agency, you're probably not going to get very far. But if you actually start integrating it in a meaningful way, you've got to get the saw out. You've got to get the hammer out and some nails, some connectors, and you've got to start building something unique for you and your business. Which is, it's a very empowering human statement, right? Even in this world, and if you build a system that frees you up from things, it's going to allow you to do other things and be more creative and innovate further.

21:27It could be a real accelerant for us. Yeah, I've found that I'm more creative or that I'm focusing more on finding good ad examples or actually doing the Reddit research on new audiences and stuff like that. I'm having time to do that sort of thing, as opposed to just sitting, like, grinding through tasks or, like, you know, that sort of stuff is kind of all off-boarded. And I heard another really great thing to echo what Jeff was saying is you have to treat it like it's an employee, like you're hiring an employee, and that, like, an employee would come to you and you would need to teach them how to media buy a pilot house, and you would need to teach them how we, you know, build stuff.

22:06But the cool thing with AI is that it never forgets, or it's going to do it the same way every time, as opposed to an employee that might forget a principal or they might, you know, you'll get lazy and sometimes. So that's the cool thing. But I think that is the key thing is when you're, if it's not giving you what you want, maybe ask yourself, am I really teaching it? Like I would teach it an employee. A hundred percent. And remember that its output really is raw input in a lot of cases. And you really have to make sure you take out those M dashes. Yes. Take out the M dashes. Take out the M dashes.

22:35And it's not this, it's that we talk about this all the time, but that it, and chat GPT just can't chat GPT cannot get out the, it's not this, it's that it's not just this. It's really this. They cannot get it. It's out of its brain at all, which is kind of really interesting. I wonder if it's just because it's, and I was thinking about this today. I'm like, I wonder if there's a gap because all of us techno enthusiasts are so dialed into this, but I wonder if there's a gap right now with, for the general population. It hasn't really, it's still just extra persuasive copy and those willing to embrace the M dash during this short window until everyone realizes it might actually have an advantage because it's one of the most useful things in all of punctuation that we are banishing to the annals of history.

23:16Yeah, that's true. It is. It's a beautiful piece of punctuation. You got to use the semicolon now, which nobody knows how to use. Nobody's using it right. It's a real grammatic snafu that we've, that AI has put ourselves into. So that's one downside for sure. Yeah. And I think too, like chat, I'll just say, maybe chat gmpg has like a an issue where so many people use it and they don't use it in in like a work environment they kind of just you know a lot of people are just using it on a day to day and so it's maybe like training itself on what it thinks people want and because so many people are just using it for day-to-day stuff it's not really refined in that way whereas like say something like claude it's just like so used for enterprise that it's just like a little bit better.

23:58It's training on like the positive reinforcement that it's getting. And so it's kind of like separating itself from chat GPT in that way. It just reminds all of the writing. I just hear Russell Brunson. I hear Russell Brunson writing landing pages for like he just this. It's very persuasive, just the short punctuation declarative statements. I think there's so much like classic conversion copywriting that goes into the way it natively speaks at this point because it's that it is the best way to persuade people, quite honestly. So I'm interested to see if that evolves or adapts or if it just continues to work well as it has for a hundred years.

24:31All right. Well, we will catch back up with you soon and we'll compare notes on what we're doing with Claude. Thanks for coming on, Brayden. Thanks, Eric. Glad to be here.

Read the full transcript

24:45Thanks for listening to today's episode. If you're not getting the DTC newsletter, you can subscribe for free at directtoconsumer.co. And if you want to learn more about Pilothouse's all killer, no filler services, take off to pilothouse.co. I'm Eric Dick, and this has been the D2C Podcast. We'll see you next time.

From the publisher

Subscribe to DTC Newsletter - https://dtcnews.link/signup


Braydon's back on AKNF with the most tactical AI-for-agencies episode we've recorded.


Eric opens with a Jeff Shannon line worth the whole listen: AI right now is a giant stack of two-by-fours that everyone got handed for free. By itself, it's not a chair, it's not a house, it's not a sofa. The value shows up when someone actually builds something with it.


Then Braydon walks through what he's been building.


Inside: connecting Claude to Motion to audit ad-to-landing-page mismatches, then having Claude vibe-code a new PDP in HTML in 6 hours instead of a week in Instapage. The Microsoft Clarity connector that nobody's talking about (free heatmaps, free recordings, API access). The Higgsfield connector for generating raw 4K assets through Claude with Nano Banana Pro and Seedance. Why Claude Design is worth experimenting with for brand-sensitive clients. And a peek behind the curtain at Gary and Blanche, the AI media buyer and creative strategist Jeff is running on DTC's own Meta account.


Plus: why the em-dash is dead, the semicolon problem nobody's solved, and the actual reason Claude reads cleaner than ChatGPT for enterprise work.


If you've been "playing with AI" and want to actually build something with it, this is the episode.


Catch the DTC and Pilothouse crew at The Whalies May 19 in LA.


Timestamps:

00:00 AI Is Raw Material

02:36 Why AI Needs Human Builders

04:18 Claude Building Landing Pages

10:02 AI-Powered Heatmap Analysis

16:36 Higgsfield + Claude Creative Workflow


Subscribe to DTC Newsletter - https://dtcnews.link/signup

Advertise on DTC - https://dtcnews.link/advertise

Work with Pilothouse - https://www.pilothouse.co/?utm_source=AKNF613

Follow us on Instagram & Twitter - @dtcnewsletter

Watch this interview on YouTube - https://dtcnews.link/video

More from The DTC Podcast

All 164 episodes
Ep 613: AI Is a Stack of Two-by-Fours. What Are You Building With It? (Plus Meet Gary and Blanche)The DTC Podcast · 25 min
Listen in VO