AI DTC WTF

How should a creative team use AI for creative strategy without losing the original ideas that make ads work?

Use AI to compress the non-creative work that surrounds ideation, not the ideation itself. A creative strategist typically spends 40 to 60 percent of the day on coordination, recaps, and admin; cutting that to 20 or 30 percent buys back the hours where original ideas actually come from. Feed the model a living brand context document so it can draft, expand, and fact-check briefs and scripts, but keep the originating idea human, because an LLM predicts from rules and past patterns and will always regress toward what has already worked.

In this episode, Will Sartorius and Matthew Gattozzi of Goodo Studios break down how to use AI for creative strategy without losing the original ideas that make ads work. Matthew explains why a strategist should hand AI the coordination and admin instead of the ideation, how to build a brand context document that actually improves your outputs, and why he cancelled his ad analytics software and built his own reporting tool instead.

We discuss:

  • Where AI delivers real ROI in a business built on physical production
  • The "Matthew tokens" model for deciding what you personally should spend time on
  • How to cut non-creative work from 60% of a strategist's day down to 20%
  • What belongs in a brand context document, and how to get AI to build it with you
  • How to sync context documents across a team so they stay living, not stale
  • Why the best ideas come from being out in the world, not behind a laptop
  • When a creative strategist should break the rules their own context doc sets
  • Why he replaced his analytics SaaS with a custom tool, and what it actually tracks
  • How to find net-new angles that will never show up in your account data

Join the AI DTC WTF newsletter: https://www.ai-dtc-wtf.com/aidtcwtf-newsletter This episode is brought to you by Runneth from Motion: https://runneth.com Learn more about Matthew Gattozzi Goodo Studios: https://www.goodostudios.com/ LinkedIn: https://www.linkedin.com/in/matthewgattozzi/ Site: https://matthewgattozzi.com/ To connect with Andrew Foxwell reach him here Andrew@FoxwellDigital.com To connect with Will Sartorious DM him here https://x.com/will_sartorius To Connect with Thomas Moen DM him here https://x.com/thomasmoen To learn more about Foxwell Founders and conversations like this one, go here: www.foxwellfounders.com Learn more about Foxwell Digital: https://www.foxwelldigital.com All episodes: https://www.ai-dtc-wtf.com AI DTC WTF is the podcast for people who work in DTC, use AI, and want to know what is actually working right now. Every week: tactical tips, strategies, and ideas you can implement in your AI workflows to make your brand or agency more money.

Chapters 00:00 Why AI can't break creative rules 01:12 How Goodo Studios combines creative strategy and production 03:20 Where AI fits into creative strategy, and where it doesn't 05:02 The "Matthew tokens" model for protecting your best thinking 07:04 Cutting non-creative work from 60% of a strategist's day to 20% 09:04 Why the best ad ideas come from living, not prompting 11:07 How to build a brand context document that improves AI output 13:09 Using AI to draft, expand, and fact-check ad scripts 17:50 How to share and sync AI context docs across a team 20:16 Let AI interview you to build your creative filters 22:30 What taste actually is and how you develop it 27:23 Separate context docs for statics, UGC, and studio video 28:25 When a creative strategist should break the rules 31:40 Why he built his own creative analytics tool instead of paying for SaaS 35:29 Tour of the custom Meta creative reporting dashboard 38:34 How to find net-new angles the data will never show you 43:41 Why AI is retroactive and can't generate new ideas 45:43 Building a hooks library from saved ads 47:43 Getting your team to actually use the AI tools you build 49:46 Parting advice: finding harmony between AI and creativity

What is AI DTC WTF?

Your Home For Tactical Tips, Strategies and Ideas for Your DTC - AI Workflows

1
If you work in DDC and you use AI

and you're wondering

what the F is going on every week,

this is your podcast.

The AI to see

F podcast

your home for tactical

tips, strategies, and ideas

that you can implement right now in your

AI workflows to make your brand or agency

more money.

Welcome to another episode of AI fuck!

Here we have Matthew.

Matt, thanks for joining us.

Thanks for having me. Will appreciate it.

Long time friend.

Matt has obviously built

a really incredible studio

that works with e-commerce brands.

Maybe you can just give us

a bit of background

on what you've built and sort of

how you got there.

Yeah, so I started getting the studios

four years ago,

and really what it's all about

is combining

strategy and production together.

Traditionally,

those two things are fairly separate,

which I think creates bad incentives

for both sides of what's being created

and why it's being created.

And so

my goal was to make sure

that I was building a studio

that really took into account

performance and business goals,

because if you could understand

why you're creating something,

it would change how you go and create it.

We work with many different brands

across different industries

and other companies as well,

and really

what we're providing them

is kind of like a full stack

of creative, right?

So we're doing the in-person shoots.

We're also doing

creator content,

static ads,

edit iterations, all of it, as well as

landing pages and

looking at the data.

So we

kind of

we kind of do a lot

within the creative sphere.

And, you know,

I think when it comes to AI,

really excited

to kind of talk a little bit

about what we're doing here,

because I think the way I look at

it is like I am very much

a creative person.

I don't consider myself technical.

My brother is the software engineer.

I was always the artist.

So very two different minds, right?

But it's been really cool

to kind of

see how

I myself can actually build some things

and use it

so that we can make better creative.

Right?

Because that's what I care

most about

is making the best creative

for our clients.

And how do we use

AI and be thoughtful about it.

So yeah,

that's that's a little bit

about what we're building. Nice.

And when I think about you

and your agency, I think about,

you know,

you're thinking about lens type,

you're thinking about

where are we going to be doing this shoot

for this product,

or how can we source

the best actors or talent?

So it feels almost like the most human,

or rather,

is one of the most human agencies

I'm aware of.

So I'm curious, like,

what has been your sort of

journey with AI

and where have you actually found

like ROI there?

I think it's a great question.

I think that, you know, obviously,

you know, it's not like

we're replacing our video

shoots all of a sudden with AI video.

And I actually don't think

that that's really

what's going to happen.

But as you said,

there's a lot of human components

to how you get a production done.

And there are a lot of things

that people don't really know.

But the reason

why it's really hard

to replicate what we're doing is

because there is a lot of crappy

hard work that has to get done,

and those things can be

you got to coordinate

a bunch of people's schedules

in a physical location, okay?

When you're dealing

with the physical world,

which is what we do.

We have to deal with things

like a lot of communication

across many different people.

You have to deal with shipping.

You've got to deal with a lot of things

that you don't really think about,

but it really stacks up.

And when you start to scale

an organization

where you're doing

multiple shoots a week,

we're doing 2 to 3 shoots a week.

That's a lot to be planning out,

not only week to week,

but month to month.

And so

not only do you have a lot of like, human

like elements

that you have to coordinate,

you have a lot of ideas,

a lot of writing,

a lot of things that have to get done.

The way I look at AI is not necessarily

that it's a replacement,

but I think about what are the most

if you are like as a human, as me,

I have Matthew tokens, right?

Like, I've got what my what I'm best at

and the efforts that I'm going to do

if I can not spend my tokens on

writing a recap slack message, and

I can use my tokens to be more strategic,

now I'm getting more value

out of my tokens while offloading,

you know, the smaller things to that.

And I look at that across

my organization.

I look at my creative strategist

and I say,

I don't want them to necessarily be like,

not sink or not do writing.

But there's a lot of parts of the job

that are coordinating.

There are things that just take

time that you've got to do,

and if we can compress that time for them

on some things,

it opens up

more time for other things, right?

So if we can compress

their communication time

writing slack messages,

recaps, follow ups, etc.,

and now spend more time being like,

what's the best idea?

Or my producer,

if she can answer all of her emails

a lot faster

now, she can think more about

what's the best talent,

what's the best location for this place

to make the best video as possible.

So it's very much about

not usually replacing,

but more about where

can you compress

certain things

so that you can

spend more time creatively.

I think the thing that

when many people they

when you're not doing creative work

full time,

you think, oh,

I just want to do

if only I could be

a videographer full time.

If only I could be whatever, full time.

That's usually kind of how people feel.

And then they actually maybe work

hard enough

and they get good enough to be in

that full time role.

And then they realize that,

well, actually, 40,

maybe even close to 60% of that

time is actually

not even the creative work itself.

And if we can actually, like take down

that 4,060%

of non

creative work to

3,020% and then expand it to hey,

now you're spending

80% of your time, 70% of your time

on actual creative work.

The value to the client is way

higher, as well as the value

to them as a person

and what they get to do as well

in their career.

And so these are the things

that I'm really thinking about.

And then I'm thinking about

how do these tools and how do we use

AI within the organization

to do that thoughtfully?

And I don't think that like,

we do anything

like super crazy or super fancy,

but I think it's enough to get that alpha

that I think most people probably want.

And I think there's

some really practical things

you could do.

There's some cool, complex

things you can do,

but there's a lot of like simple things

you can do that

if you can just get

most of your team members to do that,

like you're you're

kind of ahead of the game.

We all hear a ton about AI.

We all hear about what it can do

and how amazing it all is.

And I don't think that there's

anything incorrect about that.

But this episode is brought to you

by an AI brain

that is absolutely incredible

for your marketing team called runneth.

It's from motion.

And therefore, you know, it's awesome

because it's from motion.

But really, what runneth does is connects

every single function

of your marketing department

into one place.

It can do so many different things.

It can build you ads.

It can verify

that your brand looks proper.

It can verify

that what your top creators were

and you can ask it,

you can have it live in slack,

or you can have it live as a CLI

so you can interact with it

any way you want to.

The thing

that's really insane to me

about motion

is how absolutely intelligent it truly is

and how it works cross

functionally across the team.

There was recently

a marketer

that Reza talks about in the landing page

on run,

which you should definitely check out,

where basically she went on vacation.

The CMO did and said

if run says the ad is okay to run,

then it is okay to run.

So she uploaded all the context

for the team

they ran it into, Run It and run

said, yeah, that ad is totally on brand.

It's absolutely crazy

the number of things that this can do.

And I think the fact

that we all live in silos

and the fact that we all live separately

is really, really challenging.

And one of the hard parts about AI.

So that's not what this does.

This brings it all together

with shared context,

shared resources,

and it also looks at thousands of ads

every day.

So it's always up to date on

what actually is the best performing ads

that it's seeing across motion.

And so that's another whole thing

that it does.

It's really crazy

because instead of one more dashboard,

you get a teammate in your slack

that knows your performance data cold.

As I said,

you can ask it really anything

and answers instantly.

Weekly recaps top creatives.

Which creators your audience trusts.

It learns your brand,

your playbooks,

and get smarter every day.

E-commerce Adidas

teams are using it

to turn weekly reporting

from a two day job

into a two minute conversation

head to runneth and see it for yourself.

And it also has a 30

day money back guarantee.

So if you need a cheat code in e-commerce

and to see with your ads

and your ad creative

and your marketing team,

check out runneth com

30 day money back guarantee.

It's absolutely incredible

what this thing can do

and I encourage you to give it a shot.

Got it.

That makes a ton of sense.

So just it is still done.

What you've

said,

the idea is like you're trying

to just free up your team

to do

creative work

and effectively offload

a lot of the admin to AI.

Is that right?

Yeah, I would say that.

I think like there's I don't think that

AI can right now just I,

I think there's a lot of value as a human

to be able to,

especially in creative work.

So I'm speaking

kind of more for,

for, for our role and how we,

what we do for brands.

But I think that when you think about

what makes a great strategist

in general is

when somebody can take something

that they see inspiration,

they take that,

and then they're able to relate

that back to the brand

and how they're going to use it.

So maybe they are on YouTube

and they see this really cool

format of another video about whatever.

We ask somebody 72 questions.

Oh, we could do a 72 questions

video, right?

And it's like, you know,

AI's doesn't just like say like

you guys should do a 72

question video, right?

Like that doesn't happen.

Like the ideation part

doesn't happen that way.

And so like

we need to actually create more space

so that like

our just can just exist in the world

because it's like by existing,

by doing things, by being out and about,

they actually are,

that's where they're going

to get their best inspiration

because the best strategies are can like

take things

and then put it for the brand.

Okay, I saw this

and now I'm going to take that, I'm

going to bring it to this brand,

I'm going to do all these things.

I'm going to take this format

onto this style.

Wow.

I love how they like,

explain this thing on this TV show.

Now I'm going to take this

and do that or like whatever. Right.

And so I think that like

if I have my strategist

having to like be behind

the computer 24 over seven

and then I'm like,

why aren't you making great ads?

It's like,

that's probably

that's probably why they're not existing.

They're not living. Right.

And so I'm thinking a lot more about

how do we free up their time

so they can live and exist

so that they can come up

with the best ideas possible?

That makes sense.

So let's pull in that

thread a little bit.

So let's just assume

your strategist

found a 72 questions video.

Walk me through end to end

what that process would look like.

And where does AI

sort of fit in throughout.

Is it you know,

is it helping to brief and if so,

what does that actually look like?

Is a strategist

going back and forth with AI tell brief.

Is it, you know, for revisions?

Is it you know for editing.

Like where exactly is

I fitting into your workflow

that is streamlined it to this to,

to a point

where you have freed up your strategist

to do like the messy human. Yeah.

So I think that every shot

just on my team

has a bit different of a process,

which I think is good because I think

like they people think

and work in different ways.

So some people might really have

like more of the visuals first,

or some people might have more

like the copy kind of first.

But really

the process

is kind of the first thing,

which is like one we have already

have a document that's for the brand.

And I think that like

this is a very simple thing to do,

but essentially

to get the best outputs of AI,

you need good inputs, right?

So having some sort of

context document is extremely valuable.

And always updating

that context document is valuable.

So if if you are a brand yourself

listening to this,

make your own document about that

and you might wonder, well,

how do you know what you want in there?

Well, there's some things that are

I don't want to say proprietary to

to to good.

But the things that we think

about in our creative strategy,

we have included.

But there's a lot of other things

that are very like

basic and simple,

like reasons by hesitations to buy

things that like,

you know,

I think everybody would want to have

like a cheat sheet, if you will.

If you're wondering how to do that,

you would just go

just ask AI, like,

if you don't know how to do something,

just ask AI.

Like,

if I need to give you everything about

our my business, what do I need to do?

How do we need to build it right?

Like so.

Always ask

AI how you can get more out of AI

and like it will help you do that, right?

So just has this idea there.

Like this is really great.

I want to do this. All right.

That already is a very messy process

because there's probably

a lot of self-selecting of what

not to do. Right.

So that's something

that's like not being done in AI.

And like, again, everybody's

going to have a little bit

of a different process

and now goes into

kind of like the writing of it.

Let's say we're like,

okay, we're going to do this idea.

This is kind of

what it's going to look like, whatever.

Then it's

like sometimes people are writing like,

what am I trying to just

she loves going on walks

and kind of like writing down notes.

So maybe she's like putting down notes,

kind of like fragmented.

So not like completely writing,

but being like,

I want to talk about this.

I want to talk about this.

I want to talk about this right

then maybe goes into AI

and then it's like

has the context document.

We also have another context document

that's about like script writing.

So if it's a more traditional

kind of like UTC, that might work.

If she wanted to do something

that is like more of

like the 72 questions,

she might take that and say,

like, we're going to do a video

about 72 questions, adding that.

And then maybe she's like

speaking on like,

I want these things,

this is what I want to hit on Da.

And then like maybe editing

and working off of that, like that script

that's coming up

and then being able to say yes or no,

change it.

Sometimes they'll take it off

and write it on their own

and then put it back into AI as well.

So it's kind of like like CI less

so about like this needs to complete it

from end to end,

but it can get that jumpstart

to get the juices flowing.

And then saying, oh, we could do this

and changing it.

So there's a lot of like messiness

within it. Right?

So it's not like trust

whatever comes out or whatever.

But I think playing with it

as best as possible.

Another way of doing it would be

you do

create the

initial shell of the ad itself.

Like, I've even done this recently.

I actually write out the script.

I then put that script into

AI to kind of like

buff it up a bit, right?

Okay,

that claim based on

the context document is like,

not maybe

I wrote the percentage wrong or whatever,

and it's going to fix it for me.

It's going to edit some of those claims.

It might add

in a few additional reasons

to buy or whatever.

So like sometimes

you might have that core idea

and kind of

exactly what you want to hit on,

what you want to see,

you plug it in

and then like

it can expand it from there.

So cool.

There's a few ways you can use that.

But I think that, like

the most important thing for anybody

to practically be able to kind of like,

right, would be

you need to spend time having like,

what is good writing for you?

Like what's your style

and things like that.

You need to have like core, like,

what is my brand?

What is like the reasons why?

And whatever you find most important

because that is how you're going

to be able to get kind of better outputs

that you can work with.

So those are kind of the process.

But the last thing I'll say is

this is that if

you are compressing your time,

you can actually spend more time

thinking like,

what's the best way of writing this?

What's the best visuals?

So you can let your

you can let your process be messy

because you have more time

to really flesh out these ideas

and really think about.

Yeah, okay.

So I think there's a few important things

you touched on there.

So it sounds like this context

document is sort of your foundation,

your bedrock.

And so you're making that individually

for each brand.

But presumably there are also elements

that are congruent

throughout each of these documents.

And it also sounds like this context

document is also a bit of a fact

checker for you as well. Right?

Like you are using it at the front end

and the back end.

So how is that context document

evolving

and how

are you actually tactically using that.

Like are you just like

is it a markdown file.

Like how are you

cross-referencing the context doc

as you are building these briefs?

So I think I've mentioned this to you

in our group chat,

but we've been using this thing called HQ

and it's from Jacob, Jacob and Team.

I think it's like Indigo, Indigo HQ.

I'm probably not doing my best to

to I'm not sponsored, so whatever.

But what is what's it really is good at?

And I think this is actually

what's most important

is that one of the hardest things

about using AI

within an organization,

outside of yourself,

is sharing

those context documents,

as well as always updating them.

Right.

So the point of HQ is that it

is your context layer.

Okay.

So good Studios,

we have all of our documents in there.

Okay.

What's cool about this

is that it is stored locally and you're

working on cloud code or Codex.

It doesn't really matter.

You can

you can decide whatever model you want.

So that's another value

is that as models change you can.

Oh I'm going to jump to Codex.

But I don't have to

like ship off all of my Claude documents

and projects over there.

I can just plug one on top

or change it out. Right.

Because you own that context layer.

Now that's like stored locally,

but the app or the software

or whatever

allows you to sync it up into the cloud

and then sync back.

So if I change something about

the document, it sinks up into the air.

And then the other person on my team,

it sinks to their their desktop as well.

And so they are also getting that update.

So it makes it really easy

as opposed to being like

hey guys, updated this.

Make sure you know

you've got the correct markdown file,

all that stuff right.

And so that is a really huge gain

because I think

team context documents

is really hard to like do.

So like as I see outputs

I might be like like

I feel like I need to work

that copywriting document again

and I might rework it,

I might add to it or whatever.

It also allows other people

to add to it, change to it as well.

So it's a living,

breathing document versus

always having to reset it.

So that's practically

how we're sharing things and doing it.

I think that like

the the

best thing you can

do is, is spend the time doing the setup.

Right.

So like spend the time

really thinking about what is

what are your filters,

what do you think is good writing?

What do you think is good statics?

Like all of those things,

I think most people just start.

They just keep creating.

They're just like, I gotta make more ads.

I gotta

and it's like, pause for a second.

Like what's like,

what do you really think is good?

And think about that

a bit more and,

and be thoughtful about it.

And it's worth spending.

Like you could spend just one day

like is really

you don't need that much time

to set a lot of these things up.

It's been one day and like just talk to

AI and be like, I like this.

I don't like this, I want this.

I, you know,

and like just really shaping it.

And again, like,

I am a non-technical person,

but like asking AI like,

hey, this is the outcome that I want.

How can I get that outcome?

Like what documents do we need?

And like,

can you build me a plan to do that?

And it's like, yes.

And then it will interview you all.

Like it'll be like,

here's all the questions that we need

so that we can build this document.

And you're like, great, okay, cool.

So you're still

putting yourself in it, right?

You're not just being like,

well, AI gave me this.

This is what it is.

It's like you can shape it,

but you've got to spend that time.

But it's not like you got to spend

like 100 hours to do this.

Like spend a day, you know, and just yap

into the into Claude and have it

ask you questions about your expertise

or whatever.

And it can really shape that document

and get you like a pretty good output,

way better than what it was before.

And then you're not repeating that.

And then if

you do have multiple team members

and whatnot,

because we were sending

prompts back and forth,

we had like a full doc, like,

like with a table and being like,

copy and paste this prompt into this.

And then we were trying to like do

Claude projects and all that stuff,

but it just was like so difficult.

Now it's like way easier

to have this whole like, HQ setup.

Got it. Okay, cool.

So what I'm gathering from what you're

what you're outlining here,

is that you're having AI interview

you into building this context document,

and you're effectively

uploading your taste,

if you will, into the AI.

And I think that word taste

is obviously in the zeitgeist right now.

It's you go to New York, San Francisco.

That's all I can talk about is test,

test, test.

Alex taste.

So like, how are you?

How are you sort of thinking about like,

that human taste as you are

briefing content.

My other question is,

are these context documents?

It sounds like you may have

multiple, right?

It sounds like maybe there's one

that is very focused on

you as the tastemaker.

And then maybe there's another one

that's focused on the brand

specifically, like

how are you

sort of thinking

about these different verticals

as you're building

your, your briefs, your assets, etc.?

Yeah. So

it's always funny

when people talk about taste

because I'm like,

most likely

the people that talk about taste

don't have the taste because.

That's all these Silicon

Valley bros being like, yo,

it doesn't have taste.

Yeah, exactly.

So,

yeah.

So anyway, let that let

that sit for what it needs to be.

But jokes aside, I think, you know,

when I had just mentioned about, like,

a person

being able to

see something in the world

and then find

how it can come back

and be a creative idea,

that's actually

where the taste really exists.

And you can only develop taste by like,

experience, unfortunately.

Like it's

not something you can just like,

I just have it.

It's like,

no, like you kind of need to like watch

a lot of content and be like,

what's good, what's bad and just know it.

Right.

And like,

you can refine that palate,

if you will,

of what's going to work

and being able to be creative.

Like I actually think that's like,

you know,

creative director tends to be this

like very ominous, vague term.

And it can be from like

a title perspective.

But I think that's like,

you know,

you think about the people

who are putting together

like beautiful Pinterest boards.

You're thinking about people

who are grabbing inspiration

all the time.

Like,

those are people that are really,

really good

at being able to generate ideas.

So I think that, like,

your ideation should really come from you

as a human.

And I think that's where it's like,

I want to compress

so much of other things

so that they have more time

to just sit and exist

and try to find interesting things.

Like if I'm thinking

about a lot of my client conversations

that I've had in the last

two days, it's Wednesday.

So Monday and Tuesday

I had a conversation

with a telehealth company

that we work with

because I had another friend

tell me about something

with their health,

and it made me think, oh,

that's interesting.

Like, that's an interesting trigger.

Let me talk to them about this.

And then like, hey,

that might be something, right?

That's just me

existing, being friends

with another person.

I'm not asking them

so that I can get some like internal info

so that I can make them ad,

but it's just like

I existed in the world.

But because I'm somebody who's like,

that's interesting.

Like, I think humans are interesting

and I love how what's motivating them.

And like that motivated

that guy

to go get blood work

and go get testing and stuff.

That is actually something

that we need to figure out, like,

what is that motivation?

Maybe there's multiple reasons, right?

And I'm starting to ask that question.

If I was behind a computer

and not out to dinner with my friend,

I would have missed that.

So like and that that idea

probably is going to unlock

like a lot of money from my client.

And so you think about like,

oh, well, is that work?

Is that it's like,

I guess, you know, like,

was I sitting behind

making the best brief ever? No.

But the best briefs come from like,

great ideas and connecting it

and understanding people.

And you can't understand people

if you're not with people like that.

I think is just like the heart,

like the funniest thing.

It's like if you're grinding all day,

it's like,

no, you should probably

watch some Love Island

because most people

are watching Love Island.

And if you

watch Love Island, it's like crazy, like

you're watching people in real time

be motivated by actions

and completely complex ways.

It's like it's insane, right?

So I say all that because like,

that is where you can really get

great ideas

and thinking about things is

like existing.

So that's how you can build

your taste, your palate

and like all that stuff.

Now going back

to like the context documents though,

which is like

we have different

you have the brand context document.

That is true. Right?

So you want to say I,

I know that

like whatever,

I want to talk about these things

for this brand,

maybe the theme, maybe whatever.

And then I want to put it in this format,

72 questions,

because I saw this from Vogue.

And then

then you've got the different

types of kind of writing documents.

We have one for statics,

we have one for UGC briefs,

and then we have one for studio videos.

And they all have different elements

because those we shoot

those kind of differently.

The hardest thing

in the most important thing,

I think

to note

whenever it comes

to how you use AI with creativity

and specifically with creative work,

is that you are giving it rules, right?

The context document says

this is how we want to write

static briefs, right?

A great creative

strategist

is someone who knows

when to break those rules.

And so 72 questions probably would break

those rules

of what a studio vision

video typically is for us.

And so it's probably not ideal

to use that context document

because we're really doing something

completely different. Right.

It's not as much direct response.

We're going to be,

you know, asking questions or whatever.

So like knowing

when to break rules is important.

And I think it's

the hardest thing about AI is because

llms are rule.

It's rule based, right?

It's predicting the rules

and what you think the word should be.

And so

it's not very good at like breaking out

of those rules.

And I think that's

where the human element is, is like

if you have

as much human parts of your ideation,

you know, like,

like how to get something

really interesting

that taste, if you will,

and then you should have

individual types of things,

like I like to have

my static adds a certain way.

We have certain rules

of how we want to write things,

design things.

Same with UGC, same with studio.

But I think over time,

like we might flesh that out even more.

We might have like a, you know,

maybe we build out

like a specific style of video.

It has its own context document.

That way you're like,

hey, write it this way or whatever,

but for right now, that's what we have.

But I think always know

when to break those rules is,

I think, important.

So I think if you have that human element

in the ideation phase

that actually will prompt you to be like,

no, we're breaking the rules.

Like we need to do something

wild or very different.

And I think the best content is when

people know when to break those rules.

And they do do that. Right.

So the best examples that we love

in advertising tend to break the rules.

And yet we try to create

very rules based content.

So I think just being careful of of that

you need both.

But if you're like, wow,

I really wanted a very different idea.

It's like

it's probably not going to come

from your context document output.

So got it. That makes sense.

And I

don't want to belabor this point to much.

I don't want to move on

because I know you're building

something that I would love to see,

but it feels like

maybe there's like a streamlined approach

if you're doing something sort of,

you know,

let's just say you're coloring

within the lines,

you know,

you've done this a million times.

You have this context document.

The beginning part of your process is ID

come up with the idea

on my own,

cross-reference the context document

to make sure

the brief looks okay,

write the brief and refine it,

and then come back at the back end

and have the kind of like document

review it.

And sure,

like all my stats are correct,

maybe beef it up a bit

and then put it into production.

If you're thinking sort of like,

you know, coloring outside the lines,

thinking outside the box,

you know, insert your metaphor here.

You are saying, okay,

maybe I actually don't

want to use any context.

This is going to be

entirely human process

where AI actually doesn't even

necessarily make sense to be involved.

Yeah.

And then you could

after you build it in a more human,

you could run it through

that context

document of the brand and say, like,

are we missing anything or whatever?

Like, like,

is there other hesitations

we need to hit on?

So like, you don't always have to use it

AI to help you create or like right.

It can help you edit as well.

So I think those are

those are different ways to use it.

But I think, you know,

you definitely want to have,

you know, space and time to create.

Yeah, that makes complete sense.

And honestly having that as a fact

checker, that's pretty brilliant.

I feel like we should

probably be doing that more,

just like

checking stats

and just like

maybe this stat is better than that set.

Okay, cool.

So you've been sort of talking

about maybe like built a

AI tool sort of similar to motion

and maybe

like it's a little bit more personalized

to your, to your studio.

Maybe you can share that

and maybe talk through it.

So I think that we're in an interesting

time within,

I think, SaaS and I think that,

you know, SaaS is

they are a business, right?

They have their own business.

I find that

within kind of software itself,

they have to kind of

build for the middle.

So

there's always times where you're like,

if only I had this one feature.

Right?

And then they ask people

for features and stuff,

and they start building those features

so that you feel like complete.

But it's like, man,

sometimes you get into a program

and you're like, this is so great.

I love this program.

This is awesome.

Maybe it doesn't have everything,

but we make it work, right?

It's like out of the box. Good.

And then like two years later,

you're like, yo, what is this?

It gets.

The expansion of product is just insane.

You're like, I don't need all of this.

I didn't ask for this.

And now you're charging me

for AI credits.

And I personally think your stuff sucks

and you're making me pay

like, ten x that, like, no way.

Right now,

hopefully nobody sponsors

this this podcast,

but I've not been as public about this.

But I will right here.

So you'll get the hottest of takes.

I think that if you look

at the major players of, of analytics,

their model of charging is egregious,

if I'm being honest.

Because if I am a and see who's

working with a brand

and spending millions of dollars a month

not all on my ads,

but just millions of dollars

a month on meta,

which we have multiple brands doing that.

You are charging me

thousands of dollars to get some analytic

like an analytics tool

when I could just do a Facebook

ads report. It's like insane, right?

So like,

I at a certain point was like

I was looking at like,

if I'm looking at motion, I'm like,

I get a data analytics team

with how much you guys

are going to try and charge me.

That doesn't make sense to do, right?

Like I might as well just hire

like a brilliant, like,

you know,

whatever mathematician at this point,

which I, you know.

So that being said, I was like,

there is so many things that like

we could just like simplify

if we could just have like these 2

or 3 things like, it's like,

this is all we really need.

So I spent the time trying

to build something of equivalent to it.

Now, is it the most perfect thing now?

Does it get the job done?

Yes.

Does it save us thousands of dollars

in software? Absolutely.

And I think what's cool is that, like,

there have been features

that I thought would be helpful

and then they've become not helpful.

So we just delete it. Right.

And I think that's what's really cool.

And I think potentially one day

SaaS will be able to have like

you get the out of the box,

you know, 80%.

And then you can code by the code,

the 20% that you need.

But as of right now,

that's not how it works.

So you kind of have to do it yourself.

So I'm going to share my screen

so you can kind of see it.

But basically some of this will have

to be blurred just because of

confidentiality.

But I will

share my screen

so you can kind of see it.

But this is called Verdana.

Cool.

So basically

what you'll see is

like you have this overview.

You're able to see like

kind of high level last 14 days.

Like you can pick your time. Right.

So like it's just the Facebook.

You're just plugging in the Facebook

API into this

and you're just grabbing that data

in and it's syncing. Right.

So I can kind of

see how things are improving versus

the prior period. Very simple.

There is a creative.

So this is like your actual

like table view.

So you're able to kind of

this is our custom report that we do like

these are this is what we look at right.

Like nothing too crazy here.

We actually have we we

this is a newer client

so we haven't tagged them yet.

But we do have tags of like by

is this a founder ad.

Is this an actor. Is it no talent?

The type of hook,

the style and things like that.

So there's a lot of

like tagging that we do,

but you're able to like

see the ad and then you can click in

and then it will load

because you'll

actually play the video itself,

which is pretty cool.

And sometimes it doesn't fully load

because of the API,

but it always has like a

a preview link.

But you can see like

we've got some really cool stuff

here, right?

Like you can see the retention curve,

which is pretty cool.

Basically everything

you would need to know about this.

So like there's some other finer

details in here that's pretty cool.

But like I think overall

like what I love about

this is the fact that like

what matters most to

to a strategist is like,

can I just quickly

see what's

working in the account and what is the

like, co-create the whole rates,

like what's actually working.

And it doesn't need to be that complex,

right.

Like I think that like,

you know,

people are like

and then you're going to connect

all the AI and the LMS

and like all these things

and these software companies

are doing like all of this,

like analytics and stuff.

And it's like,

honestly,

if you have a good strategist

and like

they can just view these things

like it actually is, it simplifies it.

And I think oftentimes

we're trying to

almost overcomplicate it and like do this

like hand wavy stuff when it's like

does this add work yes or no.

And if it's working

like what are some elements as to why.

So yeah,

I think that people just over complicate

some things that I think,

you know, just make it

way harder to use their application.

So we are really just focused on like,

what do we need

so that we can get back

to being creative as much as possible?

Yeah okay.

That makes sense.

So tool like this

it seems like commensurate with motion

atria etc..

And you know, I guess one question I have

and this isn't necessarily related,

and obviously

you probably built this

with cloud code or Codex.

So this is more of a question

for creative strategy in general.

These tools are very helpful

with iterative learning.

Going in looking at

oh, this is at a great thumbs up, right?

Like maybe we should use this hook

for a different persona or whatever.

When you sort of think

like, you know,

not to insert myself at all, but like,

you know,

I generally like to think of

iterative content and sort of net

new content like net new testing.

So when you're strategists and you said

if a good strategist comes

in, I will say caveat

finding a good strategy

is really fucking hard. Yeah.

Yeah.

It's it's way easier to code stuff.

It's a fine great talent right now.

So very very true.

So this is great iterative learning.

How do you take an analytics tool

that you built with AI

and turn that into testing

entirely new tests,

entirely new personas,

angles, things

that aren't

necessarily going to

be evident in the data?

Yeah.

I like made a post to you about this,

like especially with reviews,

because a lot of people

will say like, oh, look at reviews.

And it's like, well,

if you're advertising a specific angle,

you're most likely going

to attract a certain people.

Like it's just going to kind of

like reinforce that for you.

I like this is where I like

I genuinely like hate to say this,

but that that's

actually where the taste comes in.

Like that is actually where

like the person comes in

because you have to be always open

to understanding.

How can you open

and expand the aperture of the funnel

right then?

That comes from new styles, new personas,

like people you want to talk to.

So like that's the strategy part.

Like like yeah. Like that's it. Right.

I think that when you have that new idea

where AI can help you

is maybe start to start to expand it.

I'll give you an example, which is that

we're working with this one brand.

It's a home goods brand,

but the it's a weighted blanket.

And so it's like okay great.

Like way to blanket there

science behind that.

There's a health component to it. Great.

Now like we started to expand.

Well how do we talk about this.

Not just saying like hey

everybody needs great sleep okay.

That's true.

But now we're saying,

okay,

well what about people

who are neurodivergent

and they feel overstimulated

and the weight actually

helps them with that?

Now we're doing the research

to like figure that out

if that's true or not.

And then we're doing that

and then we're thinking about like, well,

how can we break this down

into other health components?

Somebody on dialysis,

somebody on blah, blah, blah, blah.

Right.

And like it just now,

you're now you're generating like,

oh, you could do this, you could do this.

And now maybe you're using

AI to like be like,

are there other studies out there that,

you know, say weight is helpful, right.

Like,

but that initial prompt of asking

AI that came from a human

being thoughtful, right.

And being like,

how can we do something different, right?

And so I think that like that is where

the human aspect is.

So you can expand it.

Now where the tool can come in

is is like,

let's say you have an idea, you now go

and make a few statics,

see how that works.

Are you getting,

you know, good click through.

Is that expanding that like

are you seeing good early signs

so that you feel like I can de-risk going

and making a big video about whatever,

you know, an idea that you had, like

it's riskier

when you're expanding off stuff

that you don't know. Right.

So like that's

where an analytics tool can come in

to say like, hey, I had this early test.

This is what I was hoping for.

This is my hypothesis.

And now we're getting some early signs.

Let's double down on it

and let's continue on. Right.

And so

I think that's where those two things can

kind of

play in is like data can help refine or,

you know, say yes or no to an idea.

But like that idea,

you want to get something

that new or different.

That's where

like you need

to have

somebody who they

themselves are generative, right?

Like,

you can't hope that the

AI is going to be generative,

if that makes sense. Yeah.

I mean,

I think you touched on a point

that people would just like, don't

talk about

enough is like,

you know,

AI is built on

historical human knowledge.

Like,

if we're not sort of thinking forward,

you know,

these things are going to be

perpetually retroactive

and we're never going

to have any new ideas,

and we're just going to be in this

like negative feedback loop

where every outage is,

you know, the same. Right?

And so like,

I think I think people and listen,

I was in an AI psychosis

state for a very long time

somewhat come out of it.

Welcome, welcome.

Yeah. It's good. It's good to be back.

It's good to be back.

But listen

there's obviously

we've talked on many podcasts

like there's plenty of ROI here.

I don't think AI is going to be this

like fundamental foundational tool

that everyone it's

going to replace all these workers.

I genuinely don't think that anymore.

I think

if I had a better margin account,

I would be shorting space

times over.

But I guess what I'm

trying to say here is that,

like AI is very good

at retroactively understanding

what has worked in your account.

Like Matt

just showed,

like he built his own motion effectively.

And motion is a great tool

that you can use out of the box as well,

but it's not going to be the tool

for you to come up with new ideas.

And I think that is

where a lot of folks get into trouble,

because they are relying

too heavily on a tool

that is looking at history

rather than sort of looking forward.

Yeah, if I could jump in

like I think like

it's like

you can vibe code a dashboard,

but why do you need it? Right.

And like, it's so fun to be like,

I made this, I mean, this, but it's like,

no, no, no, but like,

how are you actually using this thing?

Right.

And that's why I was like,

there were things that I thought like,

oh, this would be helpful.

And then I'm like, this is not helpful.

Like, I'm not even using this like,

get rid of this thing. Right.

So I think that like,

that's why

I have issue

with a lot of like sass

because it's like

they are trying to like feature

Max like the heck out of it to justify,

like we're bringing you

value or whatever.

But it's like most of the time

you don't really need a lot of that.

I think that, like,

what's great about AI

is that

there are some really like core

things that it can do well,

like whether it's it can take an idea

and then like iterate off of that idea

and like do it more.

So if you're like,

I need ten or like this like great.

And like that

honestly is a lot of work

that's like not even fun anyway.

Like for the strategist. Right.

And so it's like, that's great.

Let AI duplicate a lot of stuff

so that you again,

have more time to be like,

what's the new thing

that I can like generate?

And I think that's really cool.

One other thing to think about

in terms of the

like coming up with new ideas

and how you do that also in this tool

and this can be completely shared.

We have this kind of like

we can save ads basically.

So these are some ads

that like we've saved.

And what's cool about it is like

we

have like

it generate the script of the video.

I started getting this bread delivered.

You know you can download that video.

So basically what is foreplay.

What is trea.

And then we have like

like what the verbal hook is.

The on screen hook is

and then like what the visual hook is.

And you can always change this

but like very cool

because like you can also

just like use this as,

like a way of like, man,

I really like the framework of this video

and I want to use it for different,

like whatever.

But like sometimes

you see good frameworks and you're like,

yeah, I want to make that.

And it like makes it very easy.

There's a bit of like an analysis here.

I don't know some of the stuff.

I'm like, yeah, I might as well just like

take it out.

I don't know if we use it completely,

but I think it's really interesting.

And then,

you know,

we have a board you can like

make the board for a client.

So you have that.

So these are like interesting

like elements that I think are like good.

And like

we have this thing where you can like

star the verbal hooks

or like any of these hooks.

And so we have like a hooks library

if you will.

This is like on the newer side.

So like still building it out.

But like if I was a strategist

and we had like 100 hooks, it'd be like,

well this is a great hook.

Like I want to use this,

but instead of it being like,

if you're a cocktail by the pool

type of person, you need this.

It'd be like, if I'm a, you know, ten

step skincare person, like,

you need this or whatever, right?

Like you can start to like,

think about how that, like curation

can help you

with coming up with ideas

and kind of, again, blending things.

And so we think about that a lot as well.

And again,

super simple tool

to kind of like build out and again,

like it's like there's purpose behind it.

The goal is how does this help

the strategist

come up with ideas

and do their job better.

And I think that

that's like that's important.

The the thing I will say is that,

like with anything

that you're building, like

making sure you get team by in

and getting them educated is important.

I find that a lot of leaders are like,

this is so awesome.

And then like,

it's not being used and whatnot.

And so I think, like

just making sure

that you understand

that the adaptation of your tools

and what you're doing

is actually takes a lot longer.

But we find that it's been very helpful,

like traditionally speaking,

because of just how like Facebook

and all that stuff was working.

You know, our editors

maybe not

didn't always have like a view of

like what's happening in the accounts.

And now they can just log in

and see everything.

And so there's been a lot of like value

add from the tool itself.

But like also know that

like you've got to do

trainings,

you've got to walk through people

and you've got to

encourage them to use it

because they may not use it if you don't.

Yeah, I think it's a really good point.

The buy in,

something I'm even guilty of, it's very,

very hard.

It's very, very hard to get your team

fully bought into this stuff.

Okay.

We do have to wrap,

but if I were to sort of encapsulate

everything here, it's just like,

you know, iteration

and building

upon existing winners and concepts.

Use AI, right?

Like you have an incredible tool bank.

You have all this data.

Matthew showed

a really, really powerful tool

they've built internally

that sort of dissects

ads that they pull in for meta.

They utilize that

and as their foundation to build

maybe a hypothesis

that they came up with on their own,

but then let the

AI really,

really help you with iteration,

but free up your own time

to do your own ideation

and come up with sort of.

Net new concepts. Is that Sarah?

That's exactly it. Cool.

All right.

Any any sort of last questions.

Are you reading anything interesting

any any sort of parting words

you like to share?

I think similar to you.

There's been times where I've gone

maybe too deep with the AI,

and I think that, like,

there's there's a balance

in the process of creativity.

And so I think find your final works for

you and that's different for each person.

Spend the time to like build out

the context documents to like do that.

Know that you're going

to have to refine it over time,

but spend the time to do it

because you will get

you will have gains from that,

but also know

when you need to write down,

you know,

pen and paper,

some ideas

and like come up with headlines

and like push yourself for that.

So you know I think try to flex between

varying different,

you know, ways in which to create.

I think a great creative

person knows how to do that.

And so I think like, you know,

find what works for you

to get you the best ideas.

At the end of the day,

if you're in advertising,

the goal is to make ads that win that.

That's it. Right?

And you may not need 100

ads, you may just need a few 1020.

And most people are spending

not enough money

to need hundreds of generated ads anyway.

So you probably should

focus on

less is more,

and most likely that's going to be you.

So I think just thinking about that

and that can maybe ease your process

and kind of build it in a thoughtful way

for how those two things come.

AI and your own creativity

kind of come into harmony

with one another.

You need both of them at this point,

but I think they only work

if you have that harmony

and you can blend it together.

And that's going to be different

for each person.

So, you know,

you got to find what works for you.

But I think spend the time

to kind of find

that harmony is going to be

where you can be the most effective.

Nice spoken from the man himself.

Appreciate you coming on

and we'll talk soon.

Thank you.