The AI System That Built a 1.4M Audience | Sabrina Romanov

What if I told you you could build an

audience of 1.4 million people in under

12 months? That's what this episode's

guest has done using AI. Sabrina Romanov

is one of the best AI people we have

talked to and she's going to give you

her entire AI playbook, including an

endto-end look at her workflow that she

uses to take a single piece of content,

turn it into many pieces of content, and

turn all of that into millions and

millions of people viewing that content

each and every month. We're going to

break down her exact workflow step by

step on this episode of Marketing

Against the Grain.

[Music]

Well, Sabrina, you are on a mission.

You're going to teach 10 million people

how to use AI. And we're going to teach

thousands of people on this show at

least, how to use AI, maybe hundreds of

thousands. And one of the things that we

wanted to talk about today was basically

like how do you go from an idea to get a

basic vibecoded app, but not just to

have an app that like you play with and

nobody ever sees or use, but actually an

app that you can distribute, get out in

the world and generate revenue either

for that app itself or for a business

that might be related to it. And I know

that you have spent a ton of time vibe

coding. You've built a lot of cool apps.

And you're going to walk us through

that. So I'd love to kick off the show

today with just like how the heck do you

get started? If somebody has an idea for

an app, what should they do?

>> Yeah, absolutely. I'll walk through the

entire process from start to end. So,

we're going to walk through building a

small vibecoded app. For example, an

interactive quiz with an ROI calculator

to help you capture leads for your

business. And then I'll show you how I

use content repurposing automation to

create like one original high-quality

piece of content and then distribute it

across eight other social platforms. And

for me that saves a ton of time because

you make one original piece of content,

you pour your heart into that and then

AI and automation intelligently

repurposes it everywhere else in your

own brand voice. So yeah, we'll walk

through that today. Uh and happy to

share my screen to kick it off.

>> Yeah, let's go.

>> So just to prove that this is a really

valuable thing to vibe code micro apps,

this is a sample AI agents and

automations directory that I vibe coded

about 6 months ago and in the past 90

days it's gotten 37,000 visitors. So,

it's probably 70,000 since the inception

of this project. The bounce rate, by the

way, is really high because it's only

one page, a directory listing agents and

automations. Let's say you're interested

in content creation, right? So, you'll

see a lot of different automations here.

For example, this AI agent automates

social media carousels and slideshows.

You can view the tutorial. You can

download the template. And this is just

a really handy resource for folks who

are brand new to agents and automations,

don't know where to get started, and

don't necessarily want to spend a lot of

money building custom solutions yet,

right? They kind of want to tinker,

download pre-built templates, see how

far they can go with those. And so this

little app, so 37,000 visitors in the

past 90 days, and it drives traffic to

my bootstrapped website just by subtle

links here, right? get the best content

creation and social media scheduling

API. A link up here, social media API.

And that is helping me drive traffic to

my own bootstrap startup website, which

by the way has gotten 1.3 million views

just this year through organic social

media content and vibecoded micro apps

like this. So I'll walk through exactly

how to make this. So this was built

completely in lovable. Vibecoded it in

just a few hours. And today we'll walk

through vibe coding like a lead magnet

in Lovable as well. And for those of you

not familiar with lovable, it's one of

many vibe coding tools. I really like it

now because it includes kind of the

backend agents AI. So it'll handle your

database. It'll handle authentication,

things that you traditionally would have

to like set up yourself and then connect

to Lovable, but now you can do it all in

one place. So I'm just going to try a

prompt here. So let's say let's use

Lovable Cloud, which is the new backend

feature, uh to create an interactive

lead magnet quiz with a calculator. And

let's do something related to AI. So

let's say it's a AI readiness and ROI

calculator to help your clients find out

how much time and money they could save

with AI. And just in general, like let's

say we want to ask 15 questions and then

generates a personalized report based on

the user's responses. And we also want

to capture their email, right? And okay,

so this is our prompt and it's going to

take about five minutes to actually run

and complete this first time. Um,

especially because it's scaffolding the

entire codebase from scratch. So it's

thinking about the design. It's thinking

about my prompt, the best way to

implement this, how to architect this,

it's going to ask for permission to set

up the database, etc. And so here it's,

you can see it's thinking. Let me zoom

out, read this more. Okay.

>> And Sabrina, do you come from a

technical background? And how hard do

you think it is to get started on vibe

coding without a technical background

and to actually build something that is

you know robust and usable?

>> Yeah. So my background is computer

science and physics from Berkeley and

right out of college I actually started

yeah I started an AI company right out

of college in 2013 and sold that

actually to Pegasus which is I know for

a long time down the street from you

guys in Boston. And so yeah, I do have a

technical background, but you know, I I

have a for example, I have a women

community of builders and many of them

don't have a technical background and

have started vibe coding. And I think

like your success is all about what

you're starting with. So if you have a

huge scope for the product you want to

build, like usually I see people fail

unfortunately because it's a little too

much to bite off on your very first try.

But something like a lead magnet is

fairly simple and absolutely feasible

today without any technical experience.

plug it into

lead magnets are like little small apps

you use to acquire traffic but then send

that traffic to the main thing you want

someone to buy. So like they're, you

know, related to your core product or

service, but they're kind of like

usually a single use case, a small

little use case that you can get

attraction for and then send some of

those folks to your product or service.

>> Yeah, exactly. And so here, so lovable

is already ready to kind of implement

the backend. So it's going to ask you

for permission to actually like control

the backend. So it'll set up your

database storage. For a lead magnet, you

typically don't need authentication, so

we're going to skip that. Um, you can

also do things like add a chat GPT call

to your app without having to set up API

keys and fiddle around with all of that.

So, they've really made it much easier.

So, just go ahead and click allow when

uh Lovable prompts you and then it's

going to finish all of that backend

setup. They also give you risk

assessments and so they can tell you

when you've exposed your keys, which is

a big thing that vibe coders do. Expose

your AI keys and you can just go get

them on GitHub. And so, I think that's

pretty cool there. They can tell people

like, "Hey, there's a risk we found from

critical risk." You can just click a

button and loable will actually solve

some of those security risks.

>> Oh, I love that. Yeah, that's really

awesome. I know that's been a big pain

point and criticism of Vibe coding tools

today.

>> And so you can see here it's thought

process and I do encourage everyone to

kind of just read through what it's

doing. So what cloud enables is a

database to store your quiz responses.

And you can see here it's actually

modifying your database in real time.

And now it's editing your CSS files

which are your style files. And again,

this initial prompt will take about five

minutes to complete running because it's

building your entire lead magnet micro

vibecoded app completely from scratch.

So scaffolding the design, understanding

what it is you're trying to do, thinking

the questions right, for this lead

magnet, what should these questions be

for this ROI report, and then designing

what that final ROI report actually

looks like.

>> And Sabrina, while it's doing a little

bit of work here, I did have a a

question for you. You're somebody who's

obviously done a good bit of vibe

coding. you have created a bunch of

micro apps and we've seen the success

you've had there. Can you tell people

kind of what the the learning curve is

like? You know, I imagine the first time

you're doing it, you're like, "Oh, it's

a lot of back and forth. It's hours and

hours." But maybe the second time it's

an hour. How do you get better at it and

how fast can it be once you're actually

pretty proficient?

>> Yeah. So, the typical curve that I see

is people are surprised even their very

first time using it how far they can

get. it feels like they've gotten 70%

into building their app, but then that

remaining 30% like the actual complex

parts of your app that make it unique

and different and your vision that ends

up taking 99% of your time. And one of

the things that really differentiates

successful products that I've seen like

going from vibe coding start to finish

is how they approach uh building it

incrementally. So rather than like a

massive PRD that you start with with all

the features laid out and described and

everything, start with something like

really really simple like this lead

magnets calculator and then add one

thing at a time very small incrementally

and also read what it's doing. I

actually commented I think on Kieran's

LinkedIn post earlier today uh this is

related to AI and content marketing but

I see the same with vibe coding. the

people who really succeed use AI to go

deeper into the process and learn what's

actually happening, understand it,

question it, push back, and like keep

iterating and learning. Whereas the

people who tend to get stuck are using

AI to distance themselves from the

process. They're not really

understanding like what's going on, not

willing to challenge it, don't ever want

to open up anything and see what's going

on. And so they're two very different

mindsets and one tends to succeed much

more often than the other

>> because all of these tools whether

you're crafting content, imagery, video,

code, where it starts to get uh harder

is when you need to make edits and

changes. Exactly.

>> And that's where you need actually some

skill on how to work with the AI

assistant.

>> Yeah. Exactly. Yeah. And so, you know,

for vibe coding, I I really do recommend

just start by keeping it as simple as

possible. Like I know everybody wants to

build a really fancy MVP, but even for

my app, Lo, I vibecoded the original

product 100% with Cursor AI and I filmed

it, documented it, and it's

embarrassingly simple. Like you could

rebuild my MVP in 3 hours, honestly,

with vibe coding tools today, one year

later, because they've significantly

advanced. But that's just the starting

point. I try to emphasize this as well

for people who are new to building

product, but like building your first

version and getting it out there is just

the starting point of product

development. It's not the end.

>> All right, Sabrina is giving you her

exact workflow she uses to repurpose

content across multiple platforms. This

is the exact process she has used to

build her own audience. You'll get her

complete automation system and a guide

on how to multiply your distribution so

you can post once and automatically

distribute it everywhere. Get it right

now. Scan the QR code or click the link

in the description. Now, uh let's get

back to the show.

>> Cool. So, here Yeah. So, we have our AI

readiness calculator here. So, you'll

see a preview just like this. And so

it's going to ask us 15 questions. So

let's just go through this and see what

it came up with. So what industry does

your business operate in? So let's just

say tech, let's say 1 to 10. Employees,

familiarity, let's say three. Um, a

little bit maybe chat GPT. Uh, this one

probably most businesses struggle with.

Quality and organization of data

>> for sure.

>> Hours per week, repetitive tasks,

probably more than we'd like to admit.

So let's say 20. Um, customer inquiries

manually, yes, for the most part.

Emails, right? Supporting emails. Uh,

data analysis and reporting, let's say

almost no time with that. Uh, does your

team create content? Yes. Okay. How

ready is your infrastructure for

automation? Probably a three. Uh,

monthly budget for AI implementation,

let's say 1 to 15. Willing to learn new

AI tools, also tough. Let's say this is

a three.

>> This this company's in rough shape.

>> Yeah.

um how but we know it's important so we

know it's this is an eight uh biggest

operational challenge uh I think maybe

slow I would choose between this or

scaling difficulties okay and when do

you want to implement everyone wants it

as fast as possible and the ROI as fast

as possible and then here is our email

capture so we'll just click this and

here's our AI readiness report right so

50% readiness score um annual savings

projected 80,000 yearly and recommended

next steps Now, of course, for your

actual lead magnet, you'll probably want

to tweak the questions, the possible

responses, even the scoring, right? So,

it makes sense given your expertise,

given your client base, given your

niche. But literally in five minutes, we

went from nothing to this interactive

lead magnet that you can now iterate on

and make it awesome for your users.

>> You could iterate on it, then publish it

to the web and start collecting emails,

which is pretty powerful. But I think

the last comment you made, Sabrina, was

really important. I think the best apps

in general are ones that come from a

person or a team with a strong point of

view. So you want to take your expertise

and your perspective and translate it to

the instructions and into the app

itself. Right.

>> Exactly. So use this as a starting

point. And this is why I try to

emphasize in all my AI content like even

if you see me posting chat GPT prompts

like that's just a starting point for

you to have a conversation and go

deeper. It's not the end state. Like the

point isn't copy paste this prompt and

then just launch this as is. It's really

to think about like what do you want

your before and after experience for the

customer to be when they interact with

your lead magnets and then code your

expertise into this app and then deliver

your expertise at scale in a way that

wasn't really possible before without a

lot of investment to build something

like this out. Yeah, it's like

traditionally marketers would have built

a report like a PDF report and give that

away and now you can actually build

these kind of cool code experiences

these interactive tools and that's

really how you have to get attention

today on the web right code has become

the new content really for marketers and

how they actually uh attract people to

the website their products and services.

You can even personalize this lead

magnet. Like if you allow them to input

their company website, you can then like

let's call perplexity and research

everything we can about this customer.

So we can personalize the report even

further. And it already captures email.

So if I go to cloud database, so this is

the database it automatically created,

right? So it's just refreshing here, but

here's the email it captured. So you

don't even have to set up a separate

tool like you traditionally would just

to capture emails. You can integrate

this with an email API like resend if

you want to use this for email marketing

and send them the report via email. So

it's it's incredible what you can do

today. And I'm most excited about vibe

coding for people who have domain

expertise but not traditional technical

expertise.

>> Yeah, exactly. Same here. I think that's

it's an unlock for people who want to

bring their domain expertise to life

through code.

>> Okay, so we got the app. Now let's talk

marketing. This is marketing against the

grain. We've showed you how to build

what we call a basic lead magnet. So

something that's valuable enough to get

somebody's attention, but maybe not

valuable enough that they'd pay money,

but they'll give you some information

exchange. Sabrina, you've learned a lot

this last year in terms of distribution

and marketing with AI and the ability to

get apps out there. Could you maybe

break down for us the top lessons you've

learned over the last year here? Yeah, I

think number one is um many people

misunderstand how I use AI. Um like I

mentioned earlier, I use AI to go deeper

into content creation to challenge me as

a sparring partner to point out blind

spots. I don't expect AI to fully

automate highquality original content

for my personal brand. So that's the

number one learning and kind of

confusion that I'd love to clarify. Uh,

number two, in from what I've seen, the

easiest lowhanging fruit for creators

and businesses is to repurpose their

existing content because you already

wrote it in your voice and your style

and now it's just a matter of like

repackaging it so it fits the format of

each social platform. So, the automation

I was going to walk through today is

something that's literally helped me

grow Instagram from zero to 418,000

followers in the past year in a

completely automated autopilot fashion

just through repurposing my Tik Tok

videos, which is where I focus on

creating high-quality original content.

And I would say the third learning is

that it's it's incredible how much

automation and AI can scale you, but at

the end of the day, it's it's a lever.

And like the function I think about is

like skill times clarity equals the

leverage I get out of AI. So even when

it comes to vibe coding, if you have a

technical background, you will get more

mileage out of vibe coding tools because

you can understand what's happening and

you can direct AI appropriately. You can

stop it from going down a rabbit hole

that doesn't make sense. And then

clarity like are you able to communicate

what it is you want to build step by

step so that AI can actually follow

those instructions. to the extent you

have skill and clarity that's how much

le like leverage you get that's

literally the formula I think about

whenever I think about applying AI to

different activities

>> so this is 8N do you think that that is

a harder tool to learn than like a

lovable or a vibe coding tool if someone

wants to start to replicate this as well

>> yeah I do think naden is a bit harder to

learn um there are other workflow

automation tools that honestly are much

easier to learn I do like naden because

it's self-hosted and open source. So I

literally host this um in my closet in

the server. Um so it's it costs zero

dollars to me and normally the amount of

workflows I'm running would cost like

thousands of dollars per month. I don't

necessarily recommend any for beginners,

but you can build basically the same

thing in Zapier or make.com uh which are

much easier for beginners.

>> Cool. This looks like a cool workflow.

>> Yeah. Yeah. So I'm happy to walk through

it. So basically what it does is every

time I post a Tik Tok video, it has this

RSS feed. So it'll literally be alerted

of my new Tik Tok video. It'll download

it without a watermark and then this

uploads it to Google Drive just so I can

have it for future reference in case I

want to repurpose it in the future. And

then this next step is where all the

repurposing magic happens. So this is

actually my app, Blotato. So it has a

social media API as well as a content

generation API. So you can create

carousels, you can create faceless

videos, etc. This option is just to

repost the video as is to Instagram.

This one is really neat. We'll hopefully

dive into this. This will repurpose that

Tik Tok video into an Instagram

carousel. And so completely different

media format, but same content and

message. And then this one repost to

YouTube Shores, repost to Facebook. This

is a sample human in the loop because I

don't want like all my Tik Tok videos

reposted to LinkedIn three times a day,

right? Like some are not appropriate for

LinkedIn. So, this is showcasing what a

human in the loop approval step might

look like. And then this last section

here takes a Tik Tok video and rewrites

it into a long form thread for Twitter

threads and blue sky. Um, and then this

last one is to repost to Pinterest as

well. But yeah, I'm just showing like a

couple different examples of how you can

repurpose and remix your original

high-quality content into just different

formats, right, that appeal to like

slightly different audiences on each of

the platforms. And so yeah, we'll dive

into this one, turning it into an

Instagram carousel. So basically, this

AI agents will take uh the Tik Tok

video. And this is a super long kind of

like writing prompt here. But the most

important thing is it outputs a set of

carousel quote cards based on the topic.

Okay. So what I'm going to do is

actually just execute all this. So, if

you're not familiar with NADN, I'm just

going to like run this create carousel

and it's actually going to start

executing the entire workflow from the

beginning up until it creates this

carousel over here. I've done a lot with

content and AI in many different forms.

Actually, never a complex workflow like

this. Uh really kind of through building

tools and apps. One of the things I find

is it's really hard regardless of like

if you use your own content or someone

else's content, it still outputs like

pretty AIcentric content. It's easy to

tell that the AI has generated that. Do

you still kind of edit it and make sure

it sounds like you before you post it?

Does the does the AI auto post it?

>> Yeah, it's a I think it depends on the

type of content. So AI avatar videos,

for example, I don't post those as is

without me reviewing and ensuring

quality. Um, the reason for that is it's

hard to get the pronunciation correct

and sounding natural. And so I have

prompts for that. But so for example, if

you're not familiar with making avatar

videos, every piece of punctuation tells

the avatar when to pause. So if you have

a period, it's going to actually be a

longer pause than a comma. If you have a

dash, that's another type of pause. If

you have like a colon, that's another

type of pause. And so um yeah, that's

difficult to get right. So I do

recommend like a human in the loop step

for for that. But like other types of

content like if it's just text then like

I do think AI can get it right if you're

giving it lots of examples of your own

writing and then the original source

content is also based on your own voice.

So this is one for example three ways to

grow faster on Instagram. So this is an

example of like a carousel that was

created with my automation. So the title

here three ways to grow faster. You can

put your CTA and your bio here your

profile picture here and it'll make it

look just like this. like be generous

experiments uh and the CTA here and then

you can obviously post this to uh

Instagram, Pinterest, LinkedIn and then

Tik Tok slideshows as well. And there's

lots of different formats you can choose

from. So like that's just one format,

but if you want to do this kind of like

popular Twitter codec card carousel,

this is another format that you can do

and then just automatically post it to

social media. So in that flow you're

showing us is like it's taking your

YouTube video then it's uploading that

video then is it stripping out the

transcript?

>> I tend to write really uh long captions.

So this particular flow is just basing

it off the captions that I'm writing.

But you can definitely add a step to

call openai whisper and get the

transcript for your video. Especially if

it's a short video like a Tik Tok it

won't cost that much in terms of

credits.

>> And so when it creates the carousel how

does it know what style to use? Does it

just choose a style itself or do you

kind of tell it at some point? Do you

like tell it what style to choose?

>> Yeah. So, if you open this node for

example, create carousel, this list

templates, uh you have a lot of options

here for different templates and the

styles here. And so, within this app,

you can choose different styles and like

preview what they look like. So, this is

the tweet quote card style. This is like

a tutorial style here. So, this is

tutorial carousel with monocolor

background. the previous one was a quote

card with minimal style and just like

different styles here. And one of the

things I'm working on right now is

actually allowing users to create their

own templates, right? So, if you're an

agency and you want to create custom

carousel or video templates for your

clients, then you'd be able to choose it

here from this drop down menu.

>> Like one of the ways that could work is

each of those styles is like a little

agent and then another agent decides

what style to use based upon, you know,

what's been used in the past or data. Is

there any of that going on where you

have a agent trying to decide what style

to use or is it still very much the

human choosing the style each time a

post is made?

>> Yeah, you can definitely do that. I

actually have a sample automation that

shows that. So, let's see. I think it's

yeah, automate Instagram carousels with

AI chat. So, here you actually chat with

this AI agent and you just like finalize

the quotes and then the agent will

decide which carousel to use here. And

by the way, this this one finally

finished running. So this is based on

actually my last Tik Tok video I posted

before this meeting. So over here. So

harsh truths about learning AI that

nobody talks about. Okay. So it pulled

that video and then made this carousel

out of it. Um that's a longer video. So

it just took a little bit longer to make

that carousel. So here it is. And like

again it's a very simple example, but it

it's really powerful because you can

automate or semi-automate this and just

have like a human in the loop step for

that final approval. Like this last step

would just be to fetch the carousel and

then post it to Instagram. But you could

imagine feeding it into Air Table or

Google Sheets, you know, wait until

someone's approved it and then push it

out and publish it.

>> That's awesome.

>> And then this last section will just

like write it into a Twitter thread,

right? So it's it's a very similar to

the first one. It's just the output

looks a little bit different because you

want that initial first tweet and then

your additional tweets after that. Maybe

something that would help everybody

understand because if you're not a

complete marketing dork like myself or

you or Karen, you spend all of your time

on the original content side on Tik Tok,

right? And how many videos are you

posting a day on Tik Tok?

>> I'm posting three a day.

>> You're posting three a day. And you

have, I think, 600 and some thousand

followers on Tik Tok. Is that right?

>> Uh I think so. Yeah. 637.

>> Okay. 637. But you have o over a million

followers across all of your channels,

right?

>> Yeah, right under 1.4. So I'm on every

channel plus email newsletter through

Substack.

>> Yeah. So what I'm trying to break down

is this workflow is kind of like more

than doubled your followers, right?

Because if you're just out there and a

smart person just doing Tik Toks, you

would have done really well, right? and

you have done very well, but you

essentially took the time to build

something that basically more than

doubled the value and output of the

work. Obviously, that's normalized

across channels and there's, you know,

if you were really breaking it down, but

at the most foundational level, you

doubled your distribution through a

workflow that does a good job of

repurposing like your best and highest

quality content.

>> Yeah, exactly. And on that note, like

there are ways you can customize this to

check the performance of your top three

videos, let's say reals or Tik Toks in

the past 24 hours, and only take the

best performing one. You can even scrape

the comments to see what people are

discussing and then convert that into

another piece of content. And so this is

the basis of a repurposing engine that I

believe every intermediate creator or

business should have. It's really

powerful what you can do with AI and

automation applied in this way.

>> Like when you look at this, it's kind of

incredible in that you've basically

built yourself a social media team of

sorts, right? Like um you've built

yourself a basically a little social

media agency and if you were a listener

watching this or you were on YouTube

watching this and you're not as

technical as you because you're

incredibly uh skilled at AI, but you're

also quite technical and you can kind of

see that in the tools you're using. what

would you suggest they use? One of the

lessons seems to be like start with one

anchor channel like what is your primary

channel and then like you know start to

like maybe automate a little bit of but

what would you recommend to kind of get

started if I'm like wow I need to do

this how can I get started with this? So

my advice for beginners uh to be

completely honest is to ignore AI and

automation

which I know sounds which I know sounds

uh backwards but having seen so many

people like start and stop their content

creation journey. I really believe that

consistency and developing the habit of

creating content is the number one most

important thing in the early days. So

many people I can't tell you they just

start. I'm sure you guys have seen this

too. They just start. They're super

excited, incredible momentum. 3 weeks

in, absolutely nothing. Maybe they

revisit it again 3 weeks later, but not

quite serious and they fall off the

bandwagon again. Um, so in having

observed that, I don't want AI to be

another shiny distraction. And I don't

want beginners to treat it as like the

cure all penatia because I'm using AI,

I'm definitely going to succeed. At the

end of the day, it's like about you as a

human, your creative, your thought

process, your thought leadership. What

is the brand you want to build? What is

the voice you want to have and what is

the education you want to share with

people? Those are the things you figure

out in your first 100 days of posting

content on a consistent basis. Once

you've developed the habit, you're going

to stick to it. And how you know that is

it feels weird to not post content on a

day. When I started, I just had a

streaks habit app and I was just like, I

just have to post one thing a day, then

I'm done. Like that is that is my goal.

Um once it becomes weird to not post

anything then I would start looking at

okay how can AI help me scale what I'm

doing that's already working maybe I can

use it just for brainstorming different

ideas brainstorming hooks right and then

I would look at automation maybe six

months or so into the journey when

you've like really figured out this is

going to be my core platform where I

produce high quality original content

for me that's YouTube for long form Tik

Tok for short form and that's it

everything else is repurposed from those

to original sources.

>> You're saying something very important

that I want to make sure people watching

this understand, which is first start

with having real domain expertise,

right? You need to actually be good at

something for AI to be truly impactful.

And so you actually have to be good at

content, really put the hours in, learn

how to do that. But let's say you were a

seasoned content creator, like I'm

watching this, I create a lot of

content. I'm like, I wish I had this

flow, right? And I think uh one of the

things that AI works really well is when

you have a recognizable pattern that

works and AI can replicate that. So if I

have a bunch of hooks that work, if I

have a bunch of like post types that

work, then I can actually start to train

the AI to maybe help me do that for that

channel, then I can basically train it

to actually do that maybe for one other

channel. Like maybe I would take my

LinkedIn content and I would use Hey Jen

and create an avatar for Instagram.

That's probably actually an experiment

I'm going to actually start to do, which

is why that one came to mind because I

don't do anything Instagram and I think

that would be kind of a cool test. But

like go peace meal by piece meal. But

you actually don't start with wow this

AI and automation is really good. I'm

not very good at content, so I'll just

start like

>> put some stuff in it and then like

spraying it out everywhere and just

create a bunch of noise. It's like no

like

>> start with some domain expertise and

then find a pattern that you can teach

AI and then AI can help start to

replicate that in other places.

>> Yeah, exactly. That's how I think about

all content creation. It's like um

reduce the time to finding your

repeatable viral format and that looks a

little different for everybody and it

also looks a little different per

platform. Um, but once you find that,

then you can scale it and replicate it

with AI to other platforms and it will

still involve experimentation. Like

sometimes reposting your Tik Tok exactly

as is to Instagram works. In many cases,

it doesn't work that well and you'll

have better performance with carousels,

which Instagram is pushing really hard

right now, right? And so, but to your

point, yeah, have a basis of like high

quality content that you're working off

of, which ultimately comes from like

your your experience, your background,

your industry knowledge and knowhow.

Yeah, by the way, your Substack is

awesome. I subscribed while we were

actually doing this as well. You've got

tons of great tutorials there. You built

a an app, you sold it for a bunch of

money, you could do a bunch of stuff.

Why did you want to become a teacher?

Like the teaching thing is like

interesting. It's not like make $100

million. It's like teach 10 million

people. You'll probably still make $100

million if you do that, but like I'm

just curious about why you started with

that as a your goal.

Yeah, I mean, so I don't live in a tech

bubble anymore. So, I do think that was

a big part of it. So, moving out of

Silicon Valley to where I live now in

the Salt Lake City area, I love it here,

by the way. But, you know, AI chat GPT,

most people are not aware of what's

possible today. And I kind of had that

light bulb moment when I started

creating educational content about a

year and a half ago, maybe a year ago.

My family visited. They saw midjourney

canvas art on our wall in our home and

they were like, "Wait, what is that? How

do you make that?" And I sat there

showing them how to use midjourney. And

for hours, my mom was like trying to

make the perfect image of a scene from

her childhood in the Philippines of like

a specific island, a specific sunset and

gradients of colors. And she was just

like having so much fun and creative and

just just a big smile on her face and I

was like, "Wow, like I can teach other

people how to do this. I don't think

there's actually that much knowledge in

the mainstream about what AI is capable

of today." That's kind of one thing that

was pushing me. The second one is just

viewing the AI literacy gap again

because I'm not in a tech bubble

anymore. It it is kind of concerning to

me how few people know about what's

possible with AI today and then there's

kind of like everybody else uh not aware

of it or being left out of that

conversation and some of the critical

decisions like we have to make as

society when it comes to how AI should

be used, where does it make sense not to

be used, all of these different aspects.

>> I really love that. I agree. I think

that's a cool mission to not teach the

folks in Silicon Valley how to extract

more revenue from AI but teach everyone

else how to be empowered with AI. I love

that.

>> Okay. Well, Sabrina, this has been an

awesome show because I think what you

have been able to do today is show

anyone who has a strong point of view,

has some expertise in making and

teaching how they can actually do that

at scale in a really productive way in a

way to make money for their business or

or just a way to have an impact about

something they care about, right? And

you're choosing to do that with AI,

which is incredible. And thank you for

coming on the show and doing this. We

will link up your app. We'll link up

Substack and everything below. So,

please check that out. If you have any

questions about how any of this

happened, drop comments below, hit like

and subscribe, and we'll see you real

soon on Marketing Against the Grain.

>> Cool. That was awesome.

>> This data is wrong every freaking time.

>> Have you heard of HubSpot?

>> HubSpot is a CRM platform where

everything is fully integrated.

>> Whoa. I can see the client's whole

history. calls, support tickets, emails,

and here's a task from three days ago I

totally missed.

>> HubSpot grow better.