UPDATE - We put together a step-by-step business plan for productized service business! It also covers all the ideas from the podcast, so you can turn everything you heard into a clear, actionable plan. 6,000+ words & 22 pages - with charts and graphs and other ideas to launch in this space. And it's only $19 if you use 50% off promo code SERVICE right here: https://buy.stripe.com/28E14o3tB3xC4823Qb3842x ━ Check out my newsletter at https://TKOPOD.com and join my new community at https://TKOwners.com ━
I sat down with Brett Williams from DesignJoy and we talked about how he built a one man productized service business that grew into a massive cash-flow machine. He broke down how he launched with no audience, made around $10,000 on day one, kept his full-time job until the business was doing $80,000 a month, and eventually grew it to peaks of $200,000 a month with almost no overhead. We also talked about what makes productized services so powerful, why design and branding matter even more in the age of AI, and what kinds of agencies people can start right now using skills like copywriting, short form editing, thumbnail design, and AI-first development. Brett also shared his favorite AI tools, his best prompting advice, and why speed and simplicity are winning more than ever.
Do about a million dollars a year just myself. These are low effort, low cost, low risk startups
that you can just spin up a website tomorrow and then you're set. Which is a huge opportunity right
now because AI is so dang good that you could take something like that and scale it to the
moon and back. Everyone needs it. Usually you get instant customers with it and it's totally free.
On a Friday, I had the idea. On a Saturday, I built the thing. On a Sunday, I launched it on
product time. And on Monday, my life was not ever the same. Did you have an audience then?
No. No audience. Nothing. I was completely anonymous. Nope. Nope. I think it was about
$10,000 the first day. It was side income for me. I held my full-time job until I was making
$80,000 a month and then I quit. $80,000 a month? Holy crap. Yeah, and I was making $75,000 a year
at the other job. I kept a full-time job for four years while running Design Joy.
So you were making more per month in profit than you were per year from your full-time job.
Yes, I was. Yeah. Oh my gosh. What is your net profit per month? I think the peak was $200,000
a month, but it usually averages somewhere in the $60,000 to $80,000 a month. It's probably
about $1,000 a month in expenses. Oh my gosh. That's wild. What productized agencies right now
you see out there in the world that are very successful that other people could copy?
Yeah. Is it okay if I get two? I want as many as you got. Okay. I think there's two primary
ones right now. Meet Brett. He's making about $80,000 a month in net profit all by himself,
no team, no contractors. This guy launched an agency on a Sunday and had $10,000 in recurring
revenue by Monday. No audience, no followers, nothing. He was a nobody in a town of 500 in
the Midwest. Not just that, but he kept his day job for four years while running the business,
making more per month in his business than he was making per year in his day job. Today,
he works 30 hours a week and his clients pay him six grand a month before they ever talk to him.
We're going to talk about this thing called productized services and not just his,
but other productized services that you guys can launch in all kinds of different industries.
Enjoy. One of the important things, noting who might be watching this is I don't want this
conversation to totally be focused around just design. These principles and what I've done here
can really be applied to any sort of mainly digital skill, whether it's marketing, copywriting,
maybe you're running social media for a brand or something like that and it's small.
Any of this stuff can be technically what I call productized, which just basically means
selling it off the shelf. Instead of filling out a form on a website, getting a quote,
jumping on a call, having three or four meetings to discuss pricing and proposals
and estimates and all that sort of stuff, it's just you go to the website, you buy the service,
you pay for it then, and then you can consume it. I like to call it lowering the TTV, which is time
to value. When you purchase, getting that thing back as quickly as you possibly can.
In my case, it's design work. In somebody else's case, maybe it's a blog post copy or something
like that. I think it's the model that serves that best in terms of lowering that TTV, which
I think is key more now than ever. A lot of companies value speed over quality, as weird as
that is to say. So that's where DesignJoy comes in and this whole model really, I think, flourishes.
So I run what is now called a productized design agency. So basically selling design services
via subscription. I guess one of the more notable things is I do it all by myself.
So I've been running it since 2017, do about a million dollars a year via that just myself.
So there's very little cost involved other than the time that I invest in the startup.
Yeah, so it's what I do. I do design work for software product, branding, you name it, I do it.
I keep a certain number of clients at a time on basically a retainer and that's that.
Okay. So 2017, what prompted you to launch this?
Yeah. So a company that I worked for, they were hugely disrupted by Uber and that whole boom,
it was a transportation company and I saw the writing on the wall. So while I knew I had time,
I started building something on the side just as a safety net. And actually, story goes,
on a Friday, I had the idea. On a Saturday, I built the thing in Webflow. On a Sunday,
I launched it on Product Hunt. And on Monday, my life was not ever the same.
That's the way it went. It was that quick, just like a few days.
And then did you have an audience then?
No, no audience.
Nothing.
I was completely anonymous. Nope. Nope. No audience. I literally shared my story on forums.
That's it.
Yeah. Explain to people what Product Hunt is if they're not familiar.
Yeah. So Product Hunt, it's primarily for tech products. There are some physical products on
there, but essentially what you do is, it's called an upvoting system. So every day, the list
refreshes and you could submit your startup to it, your product, whatever it is. And in my case,
I had to get friends and family to upvote me. And the more upvotes you get, the closer to the top
you get. And the closer to the top you get, you might go on the homepage. And there's people like
me who are product obsessed. We like to discover new products. We like to figure out what's next
and what's new. So we go to Product Hunt as consumers to check out what the new products
are. Well, if your product happens to be there, you get quite a bit of hits. Usually you get
instant customers with it and it's totally free. You just have to have the network of people to
upvote it. In my case, it was my personal network was my friends and family I went to school with
on Facebook. That was all I had at the time. And I got to a million dollars a year that way
before I even started tweeting. I started tweeting four or five years after I built DesignJoy.
So how high up on Product Hunt's list that day did you reach?
Number four.
Yeah. Number four. How many were you competing with that day?
Total numbers from that day, the unique website visits were about 36,000, which for me from
starting just the day before was insane. To your page?
To my page, 36,000 unique visits.
All spurred from your just friends and family from a hometown, Newburgh, Missouri, 500 people.
Yeah. Yeah. 36,000 uniques. I can't remember quite the exact figures on revenue. I think it was about
$10,000 the first day was bygotten what you would call reoccurring revenue. It's a little early on
the first day to call it reoccurring because you don't quite know how it's going to go, but it's
technically reoccurring. So about 10K. And then I rode that wave for about a year at that level
because it was side income for me. As the story goes, I held my full-time job until I was making
80,000 a month and then I quit. And I was making 75,000 a year at the other job.
Oh my gosh.
They kept giving me raises and raises and raises to get me to stay. So I was like,
okay, at this point I'm only working. I did most of my DesignJoy work during my meetings that I
had at that job. So I was managing both very simple. I took on as much as 50 clients at a
time at DesignJoy. So one client at my full-time job is literally peanuts to me. So it's super,
super easy. I can get it all done. So yeah, I kept the full-time job for four years while
running DesignJoy.
So you were making more per month in profit than you were per year from your full-time
job that you held at the same time.
Yes, I was by about 5K.
Oh my gosh. Hey, YouTube's fine, but what about when you're driving around? What about when you're
mowing the lawn? You need my audio podcast, The Kerner Office. Just look up TKO Pod in Apple,
Spotify, or wherever you get your podcasts.
Anyways, I quit my job. And then two months later, revenue doubles from 80 to 160K. And that's when
I had my first like, oh crap moment. What do I do here? How do I manage this? And then I had
breakdowns. I had price increases. Every time I would double my price, I thought, this will
finally curb the demand, but it just exponentially increased it because it just put me into another
category, which is a lot of things people don't understand about pricing that I had to learn.
And then yeah, so I stayed at that for a while, worked myself literally into a hospital,
almost to death. I was dying. I was working so much. And then, yeah, right now though, I've
gotten my systems faster. I've gotten better at what I do, more selective with my clients.
And so now I have a very easy life. I work less than a full-time designer.
So what were you charging per month per client those first few years?
Ooh, good question. So I started out at a measly $449 a month, day one.
Wow. For unlimited design?
Unlimited design, one at a time. I hit my first big wave, that increased to $849 a month. And
then I went from that to $1,299, to $2,500, to $3,200, to $4,000, to $5,000, to $6,000,
all the way up to $8,000. Then I dialed it back a little bit. So every time I felt like I was
hitting a wall and wanted to shut things down, my gut reaction was just like, I'm just going to
increase prices and see what happens. Yeah. How many hours a week are you working?
Now it's about probably 30. Okay.
And what is your net profit per month nowadays?
So it varies. So one thing about my business is MMR, monthly recurring revenue, it swings a lot.
I can have $50,000, $60,000 a month swings. And one of the reasons is because you have churn,
you have people pausing. So on average, it's about $70,000 to $80,000 a month.
My total expenses outside of taxes, which I won't include in this conversation because they're
pretty astronomical in my sense, because I don't really have many expenses. Outside of taxes,
it's probably about a thousand a month in expenses. Oh my gosh. Yeah. That's wild.
So you have $40,000 a month, you have $120,000 a month, but you're netting out, averaging out
$60,000 to $80,000 a month, which is basically net profit.
Yep. It's been as low as $40,000 per month. It's been as high, I think the peak was $200,000 a
month, but it usually averages somewhere in the $60,000 to $80,000 a month, but you'll get those
peaks and valleys. As a design agency owner, what does that actually mean? What are you designing
on a regular basis? Yeah. In my case, I do a lot of branding. So logos, just the overall aesthetic
of what these brands look like. I do a lot of websites, design and development of these websites.
I do a lot of social media stuff. I do a lot of presentation decks. You name it, I pretty much do
as long as it's digitally focused. So how would a productized service like yours,
specifically a design agency, how does it differ from traditional design agencies?
Yeah. So in a lot of ways, right? So if you just want to talk pricing, for example,
traditional design agencies are relatively very expensive because there's a big head count
involved, right? So when you have a website project, you're not only paying the salary
of the person actually working on that, you're also paying the PM, the project manager, the CEO,
the CFO. These companies, they have to pay these people. So their prices are significantly higher.
In my case, it's just myself. And in other cases, companies like DesignJoy, it's like a few people,
right? They're very small, kind of intimate agencies. So pricing is usually much more
affordable. Speed is the big thing. So with traditional agencies, if anyone has ever hired
one to embark on, whether it's a brand project, a website project, these things take weeks to
months, in some cases, to years. In my case, if you were to subscribe to DesignJoy, you would have
let's say a homepage back tomorrow. So it's very fast, no contracts, no calls required.
You could just decide to sign up today and get designs back tomorrow. So it's like Fiverr,
but on steroids and actually good work, right? So that's the main thing. And it's a subscription.
So one of the big things right now is the economy, right? So it's like people are
very hesitant to hire full-time people, especially in the tech world. So they like to use contractors.
But even then, as long as you don't have work for them, you're still paying them.
With a model like DesignJoy, you can subscribe and pause anytime you want to. Meaning if
work dries up and you don't really have a need for it, you could just pause your subscription,
you're no longer paying for it, and you come back whenever you do need more support. So
it's very flexible in that regard as well. Now, I'll ask you a question that you'd be
completely biased on answering. On these vibe-coded apps and websites that kind of all look about the
same generic, could that be not fully, but largely fixed with better prompting, screenshots from
other websites that have good design? I know the answer to that could put you out of a client,
but is it not prompting, or is it just like AI is just not there yet, and it maybe won't be for a
while on the design side of things? Yeah. I may not be as biased as you think here,
because I'm not so naive to think that everyone requires top-tier design to be successful,
right? The majority of people out there could get by with launching with something made from
Claude, right? It's good enough. Where I like to exist is the tier above that, where it's like,
we're not fine with just being good enough. We want to separate ourselves apart from everyone
else doing this, and everyone that will come behind us and copy us. The only way to really
do that is to layer on a design layer onto that, that separates us from a branding perspective.
Branding is still just as important, if not more important, than it ever has been.
More important, I think, yeah. Yeah, and so the second you build a successful
product, you have to assume that there are 300 people behind you that could spin up something
like that within a week. If they have a team, probably days, right? Yeah.
So what is it that you're going to, what's the moat that you're going to build around that product?
One of those moats is distribution, of course. You know, you understand that.
More so than ever. Yeah.
Because now it's easy to build the thing. It's harder than ever to sell the thing.
Exactly. But moat is obtainable. You can hire distribution partners, things like that,
but it's still a huge moat and a huge advantage. The other advantage is branding and design. So
those two things, I think, will be increasingly more important.
I'm having this realization as we're talking that design has never been more important, you know?
It hasn't. No.
Distribution has never been more important because it's so easy to build the thing. And
you and I, we go to websites every day that all look about the same. And you're like,
let me guess, Lovable, let me guess, Windsurf, let me guess, Replit, Claude Code,
you name it. We've all seen it. And so now it's more important than ever to stand out.
If you look at the number of websites that are live on the internet, it is like
parabolic these last three years. Yeah.
Because it's so easy to make a website. And when you think of, you might agree, disagree with this,
two of the companies that have really, they call it design-led growth. Is that a thing?
I've heard of that. Yeah. Yeah. I can't speak on it, but I've heard of that term. Yeah.
Yeah. I mean, it's Airbnb and Apple, and maybe even more so Airbnb because
the two founders were designers at heart. And I feel like the best products, the best digital
products are incredibly complex under the hood. They're incredibly simple on the surface,
right? That's in my opinion, that's what good design is. And that's what Airbnb is.
It looks so simple, but there's so much going on there. And I hate it when people bring up
Craigslist as an example of like designs overrated, look at Craigslist.
Or Amazon. Yeah. Either one.
Yeah. Dude, Craigslist died. Like Craigslist is dead. Go look at similar web. It went from
hundreds of millions of views a month to like four. Facebook marketplace ate their lunch.
It's like, that's not a good example. I think, yeah. And then Apple and Tesla, right? Those are
design-led companies. And it shows in their market cap. They're trillion dollar companies.
Not that correlation is causation, but I'm terrible at design. And it's something that
I've always neglected because I'm very much like done is better than perfect, move fast and break
things, like ship it, sell it before you have it, all that. That's my whole thing. And so by the
time it comes around to design, it's like, eh, we're working. It's like, yeah, it works, right?
Yeah. But it's like, you just never know how much better it could work with better design.
I'll ask you a question that's probably one of the most frequently asked questions you've gotten
these last three years. On one hand with AI, it's like we can vibe code, we can prompt anything,
we can design anything easier than ever before, right? So that would be a notch against you,
against your business model, right? On the other hand, like we just talked about,
these things need to be designed. They need good design, at least after a certain point,
most websites will never get to that point because most businesses fail, which is fine.
Sure.
So it's a rising tide situation, right? Your TAM, your total addressable market has gotten
bigger than ever. But at the same time, the tools have gotten more user-friendly than ever. So
for your business, net of net, has this been a net good for you, this AI these last three years,
or a net negative?
For me personally, again, just speaking anecdotally, it's been definitely a net positive.
It's a net positive for two different reasons. It's a net positive in the sense that
there's more businesses and startups being built more than ever before in human history. A day is
equal to six month period, a couple of years ago. So there's more need, there's more entrepreneurs,
there's more things being built that require design more than ever before.
Guys, TKOwners.com, that's my community where people are building businesses. I do AMAs,
Q&As every week live. You can ask me anything you want, you can have accountability partners.
It's about a thousand people in there building, starting, growing businesses.
Check out TKO, as in the Kerner office, TKOwners.com.
The second way, and obviously design comes into that, so I get to kind of ride that wave.
The second way is just my internal abilities have grossly expanded in terms of what I'm able
to produce internally, and how quickly I can produce, and the type of assets that I can
produce now that I would have had to turn away because I didn't have the talent on the team.
The team is me, I didn't have the talent, but now AI does enable me to do a lot of those things.
And then when it comes to product work, man, the way that my system works is,
there's not, like I said, there's not that high-level synchronized communication that's
available. So when I embark on a huge product design request, and I mean a piece of software,
maybe it's like a medical software that's very intricate, and it does oftentimes require a lot
of conversations and a lot of requirements gatherings, I don't need to do that anymore
because I plug in what it is they give me. I work with Claude to come up with all the
requirements, all the features, all the flows, just myself in isolation from even the client.
And then I can just bring that, pull that into Figma, and then just start designing.
It does a lot of the UX thinking for me, so it just allows me to speed up things and not have
to really even think that much.
Yeah. What AI tools are you using in your business right now?
Yeah, good question. Mid-journey a lot. Again, I'm a designer, so image generation,
video generation is huge for me. So I use Mid-journey a ton. I use CRIA. CRIA is like
one of my favorite ones because it plugs into all of these other AI model, image generation,
video generation models. I use that mainly for that. I do a mix of those for image generations.
For coding and prototyping, I primarily will use Claude. That's kind of my go-to right now,
but again, you ask me tomorrow, it's probably going to be something different. So it's just
the way it goes right now with those. I use AI inside of non-AI products, like Photoshop and
Loom, even a bunch of stuff. There's all these AI features, but those are the primary AI products
that I use currently.
Okay. For the non-designer, someone who's a design illiterate such as myself,
what AI tools do you suggest they use if they can't afford you?
You can get away with using ChatGPT for almost everything, or Claude. I think Claude is better
if you're looking to not hire a designer. Claude, there's MagicPath, there's MagicPatterns.
Those three are probably the best at getting you the closest to what a true designer would output.
They're not quite there yet, but again, they're good enough for probably most people
watching that don't have high design standards.
What about NanoBanana?
For image generation, there's two things here. Image generation is either you're looking to
generate artwork, vibey, super aesthetic artwork that you really want, a really unique style.
I like to play a lot with Renaissance-themed artwork. I do a lot of that sort of stuff,
kind of like Shopify's latest page that they did. MidJourney is great at that.
NanoBanana Pro is better at specific things. If you're looking to generate a model holding
something, go NanoBanana Pro all day long. It's so freaking good. I would say NanoBanana Pro for
90% of people listening.
What if someone's vibe coding a website or an app in Repl.it or
Claude Code, any of those vibe coding apps, and they really wanted to have amazing design?
What should their tech stack be? Should they go find apps and websites that they
find personally beautiful, screenshot it, send it over there? What kind of prompts should they
put alongside it? Should they ignore Repl.it for design altogether and use something like CREA?
What would you do?
Claude is the best. What I often do is I do the simple approach where I will screenshot.
Let's say I want to add a section below the top section on my landing page,
and it's just like testimonials or something like that. I'll oftentimes just screenshot
a testimonial section that I like and reference that inside of Claude. Say,
referencing the screenshot that I've attached generates a testimonial section.
That gets you like, it's going to be okay. That's what I would do. That's the simplest
approach versus explaining how to connect Claude to Figma and do all that. It's going to be over
people's heads.
Can you give me the names of some websites that you design that you're particularly proud of,
just so we can show people what you consider to be good design?
Origin Finance is one that I've recently done that's really big.
There's others that I wouldn't necessarily classify them as the best designs,
but they work for what they were built for.
Is it this one here? Useorigin.com?
Yeah. Again, design is completely subjective. This isn't necessarily what I would consider
the pinnacle of design, but it's clean and it works and it converts really well.
I've done other things like the T-app, the dating T-app, if you've heard of that.
T-E-A?
Yeah. Did you ever hear about all that?
No. I've never used a dating app in my life.
Well, I know, I know, but it was all over X because they-
Right here?
Yeah.
Oh, this is like a gumroad vibe.
Yeah. These guys, I designed this literally three years ago and crinkets. I thought it was
dead as all get out. Then all of a sudden, it's number one on the App Store, it's number one on
the Google Play Store, but they had a data leak and it was not good. I was not a part of it,
but they hugely successful app, just beat everything by a super wide margin.
Did the data leak basically kill them?
It did on the App Store, on the iOS App Store. They're still on Google Play,
but they got millions of users on day one. It was everywhere and I was the guy that designed it.
That's cool. Did you design the app too or just the website?
Yes. The brand, the app, the whole shebang.
BuyMeACoffee, I've done a ton of work for. I'm sure you've probably seen the BuyMeACoffee buttons
around for creators and things like that. I've done a lot of work for them.
Oh, that's great. They're all so different, but they're objectively beautiful.
That's the thing. A lot of these projects as a designer, it's really hard because you are
oftentimes very constrained.
Any other prompting ideas or prompting hacks when it comes to design that you would recommend?
The biggest prompting hack that I can give to people, and this may sound very common sense
to some people listening to this because they're like, obviously, I already do that,
is to use ChatGPT to write your prompts.
Whether you're prompting within ChatGPT or any other LLM?
Exactly. Especially in other LLMs. If I go to make a video in Runway, which Runway is a video
generation AI tool, or if I go to generate an image in MidJourney or an image in Nano Banana
Pro, the first place I start is typically in ChatGPT. I'll say something to the effect of,
hey, write me a prompt for Nano Banana Pro that does this X, Y, Z. Very generic.
And what it spits out is very specific. It's better than what I could write as far as a
prompt goes. It used to be like we had these prompt engineers, that was going to be a job.
And now AI prompts better than people can. So lean on, use AI to do the AI stuff that
you need help with. You shouldn't have to sit there ever and write your own prompt,
unless it's just something really quick or whatever. But yeah, use ChatGPT to write your
prompts. Why is ChatGPT better at writing prompts than, say, Cloud? Which most people,
I think today would argue that Cloud is a better LLM.
So it depends. If you ask these people, it depends on what they're doing. So if you're
let's say Bolt, which Bolt is a text-to-app platform where you can just say, create this
app or whatever, kind of like Cloud is. So if you're writing something more technical,
I will use Cloud for that. Or if I want to write a script for my Webflow site,
like a piece of code that does this special hover effect on this element, I'll use Cloud for that.
If it's something more of like image generation or more creative or just
personal use even, like I want AI to analyze a tax document or whatever it is,
I've just personally anecdotally found ChatGPT to be better at writing more conversational stuff
versus Cloud being very technical. So those people that usually say Cloud's better for this,
they're usually much more technical use cases for that. But they're both, I mean,
honestly, they're both good. You can't really go wrong. It's just a personal preference.
Leo Dion Well, to me,
there's two big unlocks there in what you said. I've been doing that for a while. Ask it what
to ask it. Cool. But like, A, I wasn't using ChatGPT to do that. I've been using Cloud to
ask Cloud what to ask Cloud, right? So that's an upgrade. B, I haven't been telling in the LLM,
in the prompt, I'm going to do this in NanoBanana. I need you to give me a prompt for Replit,
for Lovable, for Cloud, whatever, because the LLM knows how those LLMs want to be spoken to,
right? So theoretically, it should give you a better prompt that's more befitting for that LLM,
right? It will. An example is MidJourney, and MidJourney uses parameters. So there's style
parameters, there's ratio parameters. So ChatGPT will know that. You can also, one unlock too,
is setting up a rule sheet in ChatGPT so that anything you prompt inside of this project
will reference a rule sheet. So when I'm generating runway videos, runway takes into
account the type of camera you want to use, the type of zoom lens you want to use, the type of
cinema color grading. So I can feed that into the rule sheet. So anytime I prompt a runway thing,
it'll reference all that and include all of those parameters inside of that. So it just
expedites the process and gets you a much more precise, fine-tuned prompt for whatever tool
you're using. I love that. Listen, I need more people like this to interview on my podcast.
So if you know of someone with a side hustle or a business that's unique and cool and super
profitable, email molly, M-O-L-L-Y, at cofounders.com. That's one word, cofounders.com.
Molly at cofounders.com. Tell her your story and we'll give you 100 bucks if we end up
interviewing them. What is your productized service shiny object right now? What service
out there you think is ripe for the picking to be a productized service that you're too
focused on to chase after that you think other people should chase after? Yeah, is it okay if I
give two? I want as many as you got. I think there's two primary ones right now. The one is
more general. It's AI first design and development. That is the big thing, right? Because it's easy
for us technical people to sit back and say, just use cloud. It's super easy. But the second you
have to connect that to a database or the second you have to connect that to Stripe or the second
you have to actually host the thing, most people who are non-technical, that's where they stop.
Because it's easy to build it, but actually distribute it out is a whole other thing,
right? So a lot of these companies will like are right now off menus one of them, for example,
that they've transitioned and pivoted to something that's like AI driven product design and development.
So let us run cloud for you and then let us distribute it and let us build the thing and
let us host it. Let's take care of all the complexities. I think that's a big, big thing.
The other thing on a bias front would be branding. So not product design, not website design,
just branding. That is if I were to create another agency, given my skill set today,
that would be, I think, where the next big thing will be inside of actually the design
industry specifically is branding. So those are the two biggest.
I love it. If you were to go with a branding agency, let's say I'm not inherently creative.
I don't have good taste. I don't have an eye for design. Is AI to the point where we can lean on
that or with some good prompting, Claude can give us something to at least start with? Claude or
anything. So first off, I hate this. I hate this word. If you're on X a lot, you hate it too,
if you're in my world, but taste, right? If you don't have taste, you're only going to get so
far with prompting, right? So if you take somebody with actually good taste and good taste is
obviously subjective, but there's generally this in the higher tier of these industries,
there's agreed upon what good taste is. It'll get you there for something where AI is helpful
is photography generation. So you don't need to go out and hire photographers anymore. If you
have a, let's say a cosmetic, a cosmetic product, like you can get super, super high quality
photography done with AI, just general background images and supporting assets for blog posts and
illustrations and things like that. Iconography AI is fantastic for that as well. So I like what
I do is I will bring on a tech company. I'll assign an aesthetic to them. Let's say it's a
imagery with their product on laptop. And that's kind of like their aesthetic, right? It's like
this old age, kind of whatever neo-Renaissance I can set up a system for them in AI to be able
to replicate and build images out in that style for them to use on their blog posts.
And that's all very easy to do. So AI for that sense is really good with the branding side
in creating this cohesion between assets. Yeah. But it's not so good at other things.
Now break down a little bit more, the AI agency, like what, what would that look like?
You own that agency. I come to you. What need do I have and what solution?
Yeah, absolutely. So let's say like, for example, you want to build out in your case,
I'll give you a specific case. Like you want to teach other aspiring YouTubers how to do
what you've done, you know, get hundreds of thousands of subscribers. You want to build
this very comprehensive system with resources and connecting class members and things like
that to each other. You could go to Claude and do that, right? Like you could maybe get something
done, but like, you're going to probably hit a point if you're not super technical and you've
never actually launched a full product yourself with development, you're probably going to hit
a point where you, you just get stopped or you get in these like vicious, endless loops of like
trying to have Claude troubleshoot. You're just burning through tokens and spending so much money
on tokens, right? All of these things happen all the time. You don't hear about it, but like
everyone runs into this issue. That's non-technical. Instead of you doing that,
why don't you just hire someone that they have, they have actual developers on staff.
When they actually need developers to plug into the AI, they can do that, but they can,
they know how to leverage AI to get that thing up in like a month, right? Or like a couple of weeks.
For you, it's like, okay, money's not really an issue. You're like, I might as well just have
someone who knows what they're doing, who knows how to leverage these tools at the right time,
at the right place, in the right way, but also has the resources in house to be able to
have people that oversee this to make sure it's all done up to par, right?
Security's up to par, payment management's up to par, all that sort of stuff, and I have a resource
to go to if something breaks. That's where you'd hire somebody like that. I think that is the
future of agencies, inevitably. It's not going to be individuals building their own software.
I don't think it'll be companies building their own software. I think it'll be hiring someone
that uses AI to get it done at a tenth of the time, but done well, right? That's where these
AI agencies, I think, are going to plug in. They're already popping up everywhere.
What would you be able to charge for either one of those?
Probably a lot, and you need to charge a lot because tokens are not cheap currently,
and I don't think they're going to get much cheaper. That's kind of the unspoken thing
about this is, yeah, it's free, you're doing it yourself, but a lot of people are spending
thousands of dollars a month on tokens and running into the limits and stuff,
so you'll need to charge a lot. If I were to give a range, it would be somewhere,
if I were to do it, it'd be somewhere in the $10,000 to $20,000 a month range if I did a
retainer for people. Yeah. Now, a lot of people don't charge subscriptions,
so it's going to be one-off pricing, which could be anything, depending on the scope of work.
What productized agencies right now are, like, don't tickle your fancy, but
you see out there in the world that are very successful that other people could copy?
So OffMenu is one of them that does more of development and design if you want something
built like that. They're doing pretty well. They've done pretty well. Hunter Hammond runs
them. We're buddies. He's a cool guy. They run a really cool operation. I have a really talented
team. What's his name? You would know who exactly I'm talking about, dang it. But they run one for
copywriting, actually, which is a huge opportunity right now because AI is so dang good at
copywriting that you could take something like that and scale it to the moon and back and not
have to hire too many people. Everyone needs it. It's sort of, oh, Alex Lieberman. You know who
that is, right? Oh, yeah, yeah, yeah. Yeah. So we had a chat. He pinged me on DMs. He was like,
hey, can we hop on? I'm thinking about building a productized copywriting agency.
And then we had that chat. And then I swear, two months later, he was just crushing it with it.
And I think they do more than that now. But there's another one right now called
Viral Cuts. I think they're still around. Yeah, I know them. Yeah.
Yeah. So that's Hunter Hammond's thing as well. So they do it for short, short videos for content
creators like yourself, which is obviously a huge industry right now. So they take your long form
videos and clip them up and cut them into into different short form clips that you can leverage.
And that does really well. YouTube thumbnails. I can't think of the name, but there's somebody
that does YouTube thumbnails as a subscription, which is one of the best use cases for it,
because I agree. I'm like branding, which is really a one off thing. You get to capture that
revenue. Great. But it may not reoccur. Things like short term video, short form videos,
podcast editing or thumbnail design is is a reoccurring need. Yeah. So those those type
of things do really well. And you can keep going on and on and on. But there's a lot of justification
for making 10 to 20 thumbnails per video, because finally, A.B. testing is now natively in YouTube.
So we're making tons of thumbnails per video. Oh, sure. And it never stops. Like we might make
new ones a month after it's been published just to see if we can fire it back up again.
And it was a good business model before A.I. Like I pushed that before even A.I. was really a thing.
But now with A.I., it's even more so like the the the ability for you to scale now is
wilder than it ever has been before. Like if you were to talk about thumbnail design five years
ago, like a team of two or three people could crank out hundreds of thumbnails a day if they
needed to. Right. So these are low effort, low cost, low risk startups that you can just spin
up a website tomorrow, figure out how to get distribution and then you're set. Right. Like
there's service based. I love service based businesses because there's not there's not
that like risk of if it fails. OK, so what is basically just a website and maybe a couple of
processes that I set up? You know, it's not. Yeah. Anything crazy. So, Brett, I think I'm
out of questions, but this has been excellent. Thank you for your time. Where can people find
you? They want to find you. Yeah, just on X. I think it's at Brett to tease from DJ from his
Android. Yeah, right. From DJ. OK. Thanks, Brett. Thank you. If you like this content,
please share with a friend, hit a like or a comment below and we'll see you next time on
the Kerner office.