How We Use AI to Edit Ads, Build Funnels & Analyze Performance
from Marketing Operators
1h 2m 22s
Alicia Bone, a creative strategist at Runneth, shares her experience of using AI as a powerful, non-technical collaborator in marketing. Her journey from graphic design to creative strategy highlights how AI has transformed her workflow by enabling rapid data analysis, automated reporting, and creative output generation. She emphasizes that AI’s strongest impact is in democratizing data access, allowing marketers—regardless of technical background—to analyze performance and make decisions faster. Alicia stresses that while AI can generate content like ads, landing pages, and videos with minimal input, the final "last 10%" of creative work requires human judgment, emotional intelligence, and strategic insight. She argues that the future of marketing isn’t about replacing human creativity but enhancing it through AI tools. Her approach centers on building context-rich systems—such as branded prompts, personas, and data connections—so AI acts as a seamless teammate. Alicia believes that the most valuable skills in marketing remain creative strategy, customer understanding, and effective storytelling, which AI cannot replicate. Ultimately, marketers must shift from technical execution to strategic thinking, using AI to automate routine tasks and focus on higher-level decision-making. This evolution makes AI not just a tool, but a trusted extension of human creativity, especially in fast-paced, data-driven environments.
Welcome to episode 137 of Marketing Operators. Today, we are joined by Alicia Bone,
Creative Strangely Lead at Runneth. I've been working with Alicia super closely on Wynx,
and today you are all going to see what it looks like to work with a top 1% AI-enabled creative
strategist. We talk about how to make AI your most productive employee, where creative work with
AI is headed, and at the very end, Alicia shares her workflows as a non-technical AI expert.
What's going on guys? Welcome to episode 137. We have a really fun one today. It's me. We got
no Connor McDonald today. Mr. Cal Ripkin has missed one. I don't know what he's doing,
but we got Connor rolling back from, I don't know where you were Italy last few weeks. I have no
idea. A fall has ended in the great Q4 lock and is commenced. We have a very special guest, Alicia
Bone, from Runneth. Super excited to have you here, Alicia. We've been working very closely
together on Wynx. Not a lot of people know, but you've been kind of a Creative Strangely Lead for us,
working with us closely, using Runneth, interacting day to day. I'm very excited to have you
here and share a lot of what you've been seeing with AI. Today, we're going to talk about how to make
AI your most productive employee on your marketing team. Amazing. Thank you so much for having me.
I'm a little disappointed that the other Connor isn't here because I was so prepared to make the
joke that I don't know how he doesn't have a litter robot sponsorship yet. He has the litter box
behind him all the time. I can't not see it every single time. That's funny. I think he's too busy
with the David Protein sponsorship behind him. That's fair. I know I get the hype. When I was in
New York, that was one of the first things I did was run to a corner store and get some David
Protein bars because I don't sell them here in Canada. I was excited. Yeah. I think Alicia just
opened up a whole new world of potential consumer brand sponsors for us. We're really listening.
Let's go. It's not just the sass tools anymore. If David was going to sponsor, they would have by now.
It's every episode thing. Awesome. Super excited to have you. Yeah. I would love to just get a quick
bio of you. Quick background, how you got to run a through you are what you've done.
I found out when we started working out, you actually used to work at Pella with Matt, so
would love to get your backstory. It would be cool. Before I even worked at Pila, Pila and Lomi,
I was selling digital templates online during COVID and that was great. I was running a small
business online for a few years, but I really burnt myself out. It was super successful. I was in
the course community, essentially, but I really just wanted a job again. What I went to school for
was graphic design. What I did at Pila was actually graphic design, but my first thing that they
asked me to do on day one was make a Facebook ad. I had never done that before. I literally asked
what does Rose stand for. They gave me access to motion. Then immediately for the first time ever
as a graphic designer, I'd seen numbers associated with my designs. I just became obsessed with ads.
Even though I was Pila's graphic designer, I very much was wearing the creative strategy hat
and was anxious for that role. Maybe like eight months later, I went to an agency. As a designer,
again, just to get my foot in the door, got promoted to creative strategy lead within a few months
there. Then a couple years after that, I just had maintained a really good relationship with the
motion team at the time with Evan. They reached out and they were like, "Hey, we're hiring for this
role," which was called performance creative solutions consultant at the time, didn't last long.
Yeah. Now I work for Motion aka Runneth, which is so full circle because I learned everything
about creative strategy, honestly, from the motion community. Then Fuller Circle now, as part
of our partnership on Wynx, you're working with Matt from Pila. Yes, it's so cool. I actually sent
him my first winning ad. I don't know if other creative strategists feel this way, but I remember
my first winning ad. It was a carousel ad with the cow case on it. I sent it to him and I was like,
"Do you remember this? This is my first winning ad?" He just said, "Haha." He's not as precious
about my ads as I was, but such fond memories of learning everything I know. While I was at Pila,
it was great. That's cool. What is your role? I've seen, I don't know if you change it or got
prone to buy. I've seen your role be like creative strategist in residence. I've seen that.
I know that now it's different. Tell me about what your role is at Runneth.
It's changed so many times. Sometimes it changes by the day, honestly, at this point, being
at an AI tech company. I was onboarded as a performance creative solutions consultant, which basically
just meant that I was supposed to consult with brands on the customer success team to help with
retention and client success or customer success. Literally within maybe two weeks, I think the
product team saw my brain and said, "Actually, we need that on our team." I worked really closely
with the product team last year. For a while there, we had a context engineering team. I remember
when Resa said, "You're a context engineer now." I'm like, "I'm not anything with the word engineer."
That's crazy. Last year, my big thing that I worked on last year was AI tags. I actually helped
build all of the workflows behind AI tags coming up with the prompts to tag every single individual
creative, all of the visual format, taxonomies and everything. That was my baby last year. That's how I
really drove headfirst into AI and as much of the engineering as I could on the product team.
Then this year, I'm more on the GTM side helping more with our paid ads, but also just making
content and just honestly, just being a resource internally. I think of myself as an internal
consultant for our whole team. I've been very blessed, I will say, to have been given a lot of
freedom in my role in the sense that I can just play with the tool and explore with whatever I want
and then the product team learns from that. It's really ambiguous. That's cool. That sounds
fun. That's awesome. It is. It's fun. You would say you're obviously creative by trade,
creative by heart, obviously a non-tech person, but you've adopted AI pretty well. Tell me a little
bit more about that because I think that's awesome. Yeah. God, I love AI. It's just magical,
but I don't know if other people can resonate, but I definitely had what I call my AI crash out
end of last year, which was I had built all of the AI tag workflows in deterministic workflows.
So first you do this, then you do this, then you do this, and I had reverse engineered my
process and built it in a linear way. Then when I specifically remember, and I said this in my
talk at Chief Marketer in NYC, Rezzo was like, "Okay, Claude just made this thing called a skill."
So you put this front matter in a text document, and then you give it instructions, and I went,
"Oh, I don't know what's going on here. This is freaking me out." The agentic workflows,
I just could not wrap my head around, and it terrified me, but it really just took
like diving in with two feet and exploring. Evan specifically said to make runneth look cute
earlier this year, like just go play with Claude, play with chat, chat with play with anything
you need to like wrap your head around AI, and just enjoy the process, and yeah, I would manage to
become very AI-pilled without having to actually open terminal, which I feel so insecure to say,
but that's just how it is. I just live in prompts, honestly. I maybe saw you more technical,
like I've done some stuff in terminal, like done all the setups before using runneth and stuff
like that, but I never thought about like my file organization structure, and when we start
working together, you're like, "Oh, here's how," and I'm like, "I just like I completely vibed it,
and I would just go to Claude or whatever, and I'm just like, "Hey, build this for me," but it sounds
like you're very at least thorough, even saying, talking about the front matter, whatever it is,
of a skill. I've made hundreds of skills. I've never once looked into how they're built.
That's so funny. Yeah, I'm very much a person. I like to call myself the worst runneth
customer in the world, because I am so skeptical of everything, and I actually think that that's
like part of my innate talent, I guess, as a creative strategy. I just want to know like how things
work, and like get to the root of problems, and reverse engineer everything. So if someone was like,
"Here, download this skill and just use it," for me, I'm like, "I'm absolutely not going to do that.
Like, I want to know exactly what's going on inside of here so that I can understand what's
happening." Because I'm naturally curious, and I want to learn, but also so that if something breaks,
I can figure out where it broke. Like, if I have no visibility, I think that's why I freaked out
about like, "Agentic AI" is just because I felt like I had less visibility and control. So controlling
my file system and understanding how that system works for me is like my way of having control
over how things work, and then I can just like double check AI as much as I can't like to ask.
Like, which skills did you trigger? Why did you do this? Like, asking it questions until I trust it,
and it really is just like me building trust with AI. Yeah, that makes sense. For me, like, some of
the concepts, like when I first heard skills or projects, like I don't get them, they don't resonate
with me until I do it. I'm a person who has to like do it, but it's just like playing with it,
not the files for sure. Connor, where are you at in like your AI journey? Like, have you had AI
psychosis ever? I love watching people when they get AI psychosis. Like, how much are you using it?
Obviously, you have a much larger team, but I know you guys are using it for some stuff, but where
are you at with your journey? Every brand I talk to has the same problem right now. You need way
more ads and you need better ads. We have agencies, creators and AI that all help with the volume,
but then there can be a week of back and forth. There's a lot of logistics and a lot of people in
the way before any of it is even ready to launch. At Wings, we skip all of that. We have runeth that's
helping with everything. Runeth makes almost all of our statics directly in Slack. They come out
launched pretty much one-shoted, uploads them, analyzes them, dissolves it. Runeth also helps with our
tribe. We're getting so many videos, but we want to make sure that they are appropriate and hitting
the right things. Runeth has been trained on all of the feedback and comments that I've given to
tribe creators, so it is briefing them. It is QAing, all of them, and it's messaging them feedback
based on what's working in our ad account. Runeth has started building all of our landing pages for
us. A lot of people are talking about having ad and landing page compliance together. Runeth has
made that possible by allowing us to ship really good landing pages very quickly. Runeth has
a homework. It has all of the context of Wings and our performance. If you want to check it out,
go to runeth.com and start your 14-day trial. Tell them Cody sent you so they can hook you up.
Me personally, I'd say like 90% of what I
I'm using it for is data analysis,
'cause in my role, it's like I always say
it's like the 108010 rule.
It's like you're very involved in the first 10%
of an initiative and giving the ball rolling,
and then it's like the team is more of the hands-on keyboard
kind of getting the bulk of it done,
and then you're kind of coming in at the end,
and that last 10% is like part review, part analysis,
and I think that's most of where I'm at with AI,
and the team's using it, we can speak to some of the way
that our team is using it for creative strategy,
specifically in production and stuff like that.
But for me, and I think one of the topics that we have in here
is how do you onboard AI and do an org?
It's really about connecting it to all of your,
like giving it all the context, right?
Like connecting it to all your data sources,
uploading brand guidelines, design guidelines,
stuff like that, so you can really have all that context
in one place.
For us in the last six months, that's exactly what we've done.
We now have, you know, MCPs and connectors
to every single data tool that we have,
which has just made the ability to go and do analytics
and have these dashboards talk to one another so much easier
than it's ever been before.
So I'd say for the most part, those are the types of things
I'm prompting it, like, hey, I know we launched this ad
in meta, and I wanna hear about how it's going,
and like, just like instead of going into north theme
and Shopify and motion, just like prompting it right there,
and having them all talk to one another.
So that's where I spend most of my time
is in the analytics bit.
And then like, a lot of, hey, I think we need this,
like this is a whole in our current workflow.
I think that we can solve this with AI,
and then like making sure, like I'm not necessarily
going to be the person to own that,
but like making sure that whoever should own that
is like prioritizing that.
So for example, I feel like we're building this like,
I've seen some stuff that London is here.
- Exactly, that's what I was about to say.
It's like London absolutely cooks.
We've built this like, I should say we he,
has built this like hex-clad OS,
which is basically, it has a lot of things,
but one of the core use cases we had right away,
as we were making this, this like monthly marketing
performance report every single month,
that was really robust.
Like, would touch on every single like blended business
performance, and then every single channel.
So in the last six months, what we've, what we've said is like,
hey, we can just connect all these platforms
to our cloud enterprise account,
and we can build this dashboard and have it be dynamic.
We don't need this to be like a manual data poll
every single month.
So that's like one of the examples where I was like,
hey, this is a big time stock from our team.
I think we can build this up in a more dynamic way,
and that's what he's really been working on
in the last six months is like,
and now we have like a very robust,
like you go to this master dashboard,
and it has a section for our entire blended business,
every single channel, affiliate, paid, retention,
like edge mesh data, like every single data point you want,
with the ability to filter by timelines,
and the amount of time that is gonna save our team
is like astronomical, and it's gonna make our decision-making
better, 'cause it felt like every single month
we'd get to the month's end.
You'd be like, all right, time to make the report again.
It would take two weeks to make the report.
It would take another week or so to like go through it
and like make conclusions based on it.
And then by that time, you're like,
you're onto the next month.
So just like we were never like doing it fast enough
for it to be like as useful as it could be.
So stuff like that is while like making sure
the team is focused on the right kind of like dashboard
buildouts and use cases when possible,
but I just think like the democratization of data
and the ability to analyze and make conclusions
is it's so empowering, 'cause like you were talking about
at the beginning of the episode of the issue like,
you're not technical, I'm not technical,
I'm not a developer, I'm not a designer.
I'm not even that great in like jockeying spreadsheets.
I'm okay, but I'm not fantastic.
So the ability, like there's really no excuse anymore,
not to be a data analyst.
I think everyone should be querying,
making the connections and just making like at least trying
to make these conclusions on your own.
And if you're, I think that like, you know,
Peter from David talks a lot about how if you're a founder
and you've like pick up the phone to make an adjustment
on product, like you have to be really close to product.
I think the same is true about being a marketer in 2026
and not being able to like be in and analyze
and make conclusions about the data.
If you're a marketer and you can't review data,
make conclusions based on it in 2026, you're cooked.
Like there's really no excuse anymore.
About two years ago, there was excuses like you'd still have
to know how to work spreadsheets, run pivot tables,
you'd have to know how to like navigate these platforms.
But you really don't anymore.
You just need to know which questions to ask.
So I think that's like my biggest thing is like,
if you're a marketer, I don't care if you're a brand side,
performance side, I don't care if you have like,
no data background, there's really no excuse anymore,
not to be able to navigate it and make conclusions.
And that makes you so much more valuable in any org.
I don't care where you are.
So that's like my, that was my like big aha, you know,
when you could connect all these tools
and have them talk to each other
and just do all your data analytics from a single place.
You know, we're not going between eight platforms anymore,
which is exciting.
- Oh my gosh, there's so much unpack there.
First of all, you know, we want to know something funny
is that when I came to motion,
I did not know what a pivot table was.
Like that's, and I think that's like the definition
of the problem that motion was solving for creative people
is that I never had to know what a pivot table was
to understand data.
And I feel like that's still true.
Also, for what it's worth, I really, I'm like not surprised
that you're like, oh, analytics and data,
that's where it has the biggest impact for me.
'Cause what I really noticed talking to like hundreds
of teams last year when I was doing product research
for AI tags was that there's very much like two camps
of people, there's a people who come from the creative background.
And we would show them like a top ads report
which was very, very visual and they would get it.
But as soon as we showed them what we call
a comparative report, which is like charts and graphs,
they would just absolutely lose their minds.
And I'd be like, okay, what background did you come from?
And they were always videographers, designers, copywriters.
But as soon as you showed something like that to a media buyer,
they wanted to see all of the data.
And I really do put people in camps.
Like when I talk to like Brandon from Harry's, for example,
he's very much this like media,
I call them like the media buyer bros.
Like he talks in numbers.
And if you see what he builds compared to what I build,
what I build is so visual and what he builds is so data centric.
And we get to the same conclusions.
It's just in two different ways.
And it's so interesting to see like the way that people approach AI
and the problems that they try to solve.
It really does feel like those two,
those two different like directions, if that makes sense.
>> I fully agree, but I think what's also cool is
these tools now and like the way a company can set up a cloud,
it allows you to cross that bridge so much easier.
So like if you're a numbers guy or girl,
and you're like, hey, I think in spreadsheets and pivot tables
and aggregate numbers, which is more me.
I'm not fully on that way, but like I'm somewhere in the middle.
But I can then take that, that spreadsheet or that data,
pump it into cloud and say visualize this.
And now like someone that thinks more in visualization
can understand the just and in reverse.
Like you could take your like visualization of the data
and say, hey, make this into just more of like
fewer numbers, spreadsheet.
And I think that's also cool that like that you used
to just not be able to like connect those two brains together.
And now that like there's just like easy two way doors
into both, which is exciting.
>> Yeah, I agree.
One of the first things that I actually built with AI,
I called it my ultimate report.
Like I always had this one criticism about motion reports,
which is that you had to like have individual reports
for different campaigns.
Like you had to set up your filters
and then save them and lock them.
You couldn't like stack them on top of each other.
So I would go into my like weekly creative meeting
and open like multiple different reports
for different campaigns.
So the first thing I did was like stack all of those reports
on top of each other and then integrate
all of the other data tools.
Like so we measure lead quality in HubSpot for example.
So I was able to merge those two things
so that now our creative retros could be like much more involved
with the growth managers, like the people
who are looking at the numbers.
So it's just like a different experience
and a different capability that we didn't have before
to have this like one view that we could fully customize.
Even on the call you would have questions
and we would just be able to ask in natural language
or question and get it back.
Like we could never do that before.
It was the first place that I went.
Yeah.
>> We talk a lot about CAC on this show.
You can't control what meta charges you,
but you can control what happens
after someone lands on your site.
That's where our sponsor after cell comes in.
Helping you make every visitor worth more
from cart offers and one click post purchase upsells
to subscription upgrades.
And here's the best part.
Rock thanks turns your confirmation page
into another revenue stream with cross brand upsells.
With Earn and Save, those earnings also offset
your after cell subscription dollar for dollar
and you get to keep your entire payout.
Head to aftersell.com/earn-save
to see what after cell could unlock for your business
and how you could get the entire software
for zero dollars a month.
I think there's obviously levels
to it like different organizations are adopting
AI at different speeds.
And we've talked about a lot.
It's like very hard to do in larger ones.
And like this is no shade at all.
But like hex class probably not the furthest long
'cause you guys don't have to be.
You have budgets, you have teams, you have resources.
Like you shouldn't be, you know what I mean?
Versus like other in a spectrum
all these like new drop shipping brands,
wings stuff like that is like,
well hey, if I have this idea, I don't have anybody else,
you know, like I got to do it.
And so there's I think different levels, right?
Of like adoption.
And I think like the first level
where everybody should start is like,
let's just get access.
Let's get all of our data and let's have it save us time.
Let's have it analyze stuff.
And then I think if we're talking about turning it
into an actual employee and like sharing some stuff
that it can do is how do you have it go and do stuff for you,
you know?
And it's just getting better at such a rapid rate.
So like, 'cause I think that's,
That's like one role.
It's like data analyst, super helpful.
Obviously you can write copy and stuff like that.
And I tweeted last week and it was a very controversial topic.
I didn't think of like, I was like,
hey, the only skill that's gonna be left is copy.
And it was like a throwaway tweet hot take.
Like I don't actually believe it.
But my current observation of where I think AI is going
is all of the technical skill set
is being weathered away.
For example, like I'm not the most technical,
but I got a Mac mini over here.
Instead of her, me is like spent a lot of time
doing context engineering and stuff like that
and secrets management.
And like I think there was a little bit of like arbitrage
on that not saying it like helps performance.
But like you can do stuff with AI
if you're, you know, it used to be technical.
But now either these labs or right with like news coming out
and Grockba or run it, right?
Like they've kind of taken the technical stuff away.
Like I spent so long when setting up links
before using run-off building a claw code static ed system.
And I can't tell you how many times it broke
and I'd have to be like, why are you doing this?
Let me look at the skill or whatever.
And finally got it in like a pretty good spot.
But like it takes a decent amount of time
to build and train these skills.
And there's, you know, APIs, you got to manage
and stuff like that and you're pushing and pulling.
And then you guys throw run-off and slack
and you brief it on some statics.
And I was like, it just works, you know?
And so I think that technical side is going away
whether you're using an off-the-shelf thing,
using a specialized agent like Reneth.
And so it's like, all right, what is left?
Because it can build landers, it can do all that stuff.
Like it's like, to me, it's copy,
which is more than copy.
It's like strategy, it's creativity.
We were at the Reneth event.
You guys are doing these really cool events.
I was at the New York one, I spoke at that with Grace Clark
and you know, Jamie Ganon.
And my big observation from that,
'cause it was that the creative people
are the people who have the advantage now.
People create people that are adopting AI
because like, I'll get, I'm not a creative person.
I'm like, I'm also not a data analyst.
I'm very in the middle, right?
Meaning I don't have skills either of them.
And I'll get to a point where I'm like designing something.
And I'm like, I actually don't think
that this is a cloud limitation.
I actually think this is a code limitation.
Like if I knew how to prompt it better, if I knew,
and I want to share some of the specific examples
from the event, but that's kind of my current take.
How do you feel about that as a non-technical creative person
who's using AI a lot?
- I mean, obviously it would be very self-serving
and me to say, I totally agree.
That's great.
Yeah, 100% I have the advantage
because I am non-technical.
But to be honest, I wasn't sure.
Like six months ago, three months ago,
like Resa RCEO has always kind of had that perspective.
And listen, there's been so many times
where Resa said something and I'm like,
"Oh, you're out of your mind, you're at the zoo."
And then like six months later, a year later,
I'm like, "Ah, Resa was always right."
So I just, I trust his vision entirely to be quite honest.
And so I was banking on that and trusting him
that I did not have to learn how to open terminal.
And so far, it's paid off.
I will say like, people look on me,
like at me on the internet or our customers even,
will message like the CSM's and sales
and be like, "Alish, you did this thing.
Can I do that with Renith?"
Like, people look at me and see me as someone who's AI-pilled
and who knows how to use AI really well.
And it's so interesting that I, like I said,
like feel insecure that I feel like I don't know
how to use AI very well because I don't use the tools
that I see people using on X.
But I just have to remind myself that like,
like even yesterday, my friend Rachel sent me
a carousel on Instagram.
She's like, "You should start doing carousels like this."
So I go, "Renith, make me a carousel like this."
And it made it in one shot.
And the reason why it made it is because of all the work
that I did behind the scenes of like building that brain
of having the skills and the prompts in the context
behind it to be able to just build something in one shot.
So in that sense, yeah, like,
I absolutely have the advantage
because I'm able to like build up that context engine
is what we sometimes refer to it as,
which is basically just like that file system essentially
that enables me to ask something so simple
or if I don't get what I want from AI on the first shot
then I'm able to work with it
and ask the right questions to get myself there
because I have this like vision and this idea
and the ability to use human language in a way
to get to the end result that I'm looking for, you know?
- Yeah, I think it was like Andre Carpathi like a year ago.
It was like he's a guy who created vibe coding
but I think he said like English is becoming
the most popular programming language or something like that.
But yeah, I know 'cause there's people that are still
like trying to prompt engineer and stuff like that
and I think they're not quite getting it
and it's context engineering and like building
all of your skills, right?
And so like for example,
Jones Road just signed up for Runathon adopted it
and you know, they were asking like,
all right, like what are you using for?
And my big thing is like what I've learned from using it
and using other agencies like there's levels to it
and like start with personal context like absolutely.
Whatever you're using for context,
ever using HQ, if you're using Notion,
if you're just using Runathon GitHub
like give it context as your company brand
just like everything that you have.
And I like to think of AI as like,
what would you onboard a human with?
Like if you're gonna give them a Google Drive folder
like you need to know your brand, deck and tone,
you need to know your pillars,
you need your products, you want your customer intelligence
system that has reviews and all that stuff, right?
The one of one of my favorite people
'cause I'm actually inspired by a lot of the creative
type store doping AI.
This guy, Emma, I don't know his last name.
Used to run, it's like little planes as his agency.
Used to run like, it was like gin lane
and then pattern brands.
They do like brand design work,
but they're doing it very agentically.
But it's like you would never be able to tell.
It's like really beautiful stuff.
And so he talks about that
and he had a really good long Twitter thread.
We could try to put it in show notes
about how like your Figma brand deck
is now just a markdown file.
And so like whatever you would have in the old world
of Figma brand deck, Google brand deck,
you're all of your tokens for your design system,
everything just gets put in a markdown file
and then you just have to give your agents access to that.
That's like step one, I think.
If your CFO gives you $10 million for Q4,
where do you put it?
MetaSys prospecting, last clicks as branded search
and outdated MMM says TV.
Most marketing operators have more measurement than ever
and still no decision.
More data, more dashboards, more complexity.
I know we've been here at Hexclad.
We have MTAs, hold out tools, MMMs
and you often look at all this data
and you start to analyze it,
but then you don't make that final step,
which is what's the decision?
Like what are we doing because of all this data?
That's where Architect houses new AI agent comes in.
It turns measurement into weekly recommendations
grounded in incrementality.
It factors in your launches and promos,
recommends where to move budget and tells you why.
Thrive Cosmetics got $700,000 in annualized incremental revenue
from just three weeks of architect recommendations.
Go to house.io/operators to start getting recommendations
from Architect and optimizing your Q4 ad budget today.
And it's interesting to speaking to such a wide variety of teams,
like small teams, big teams with big budgets, little budgets.
Some of them have all of those things
that they're able to just dump it into a VM and they can go.
And then other ones, it quite literally lives
in their human brains.
And actually the biggest blocker for them
is trying to figure out and unravel why they make decisions
so that they can train an agent
to be able to make decisions like them
because they don't have all of those things documented
or even worse, if the team is a little bit fragmented
in the sense that this person over here
thinks that these are our personas,
but this person over here is defining them completely
differently, then it's really hard to enmesh,
especially with a tool like Runnith or even Mews or Gropbot,
where one of the benefits is being able to work as a team
like you have to get that alignment
and you have to be okay, like build the system together.
Even in the Winx VM, honestly, like Cody,
sometimes you ask Runnith a question
and it answers you a little bit differently
because you merged in a bunch of stuff from HQ
and I merged in a bunch of stuff from myself.
So like you asked it for founder ads the other day
and you didn't even realize that it gave you
my curated set of founder ads,
which it could have also just like script,
inspo, and found you some.
Or sometimes it will write you a script
and it will use some of the customer language
and the TikToks that I put in the brain
and you're not even realizing it.
And then other times like I'll ask it a question
and it will give me like stuff
that you definitely brought into HQ
and I'm like I didn't put that there.
So I think that's also part of the process of a team
is like you have to unravel how your team works together
and how your team makes decisions
to be able to build up all of that context
and to make sure that you're not like confusing the agent.
The same way that you'd probably confuse a person
if two different people went to that person
and said these are my personas.
No wait, these are my personas.
What's that person gonna do, you know, get confused?
So.
- Yeah, no, I know.
So yeah, you have to onboard it like you had a team member.
So I think that's number one.
Number two, number two is what Conor Ones said about connectors.
So where APIs, MCP, give it access to tools,
to pull stuff in, to do it.
And again, like these are both areas
where I think that runneth makes it really easy
instead of having to do all of this stuff.
What, you know, you set up like obsidian for all this stuff
and then like I can't tell you how much time
I've wasted on API management, stuff like that.
And then it's like, all right, like what do we do?
Like I was, I was telling Heart Jones where I would like
start with that stuff, start with context,
start with the like get the data stuff.
And by the way, like you got to be really careful with data.
Like I know that that you guys have done a really good job
of that, but like there's so much bad data
that people are pulling in with AI
that they just don't even realize it.
You gotta have, there's a ton that goes on behind the scenes
with runneth and with other tools that are done properly
that make sure it's cleaned and getting proper data in.
'Cause these APIs get really wonky.
The LLMs need it formatted in a certain way.
way. It's super boring, but just want to get that out of the way. But then it's like all right,
like let's start with the manual stuff. Can I pull these data requests a lot quicker than having to
go in and build this dashboard, you know? And then it's I think everybody should be more ambitious.
And so my input to her, I was like, start with these things. Name me conventions. Boom,
like do that, right? Like there's a lot of low level media buying work that has to happen,
especially when you're launching hundreds of ads a week in day. You know, we used to have a
full-time overseas person who just did, you know, project management, naming conventions, you know,
keeping boards up to date, like all of that manual stuff. Start with that. And I find run it
does it better, right? We agreed on taxonomy, naming conventions. It has UTMs, like I don't
spend any time thinking about that. And then I feel like I'm a pretty AI ambitious person.
I know what it can do. I'm pretty knowledgeable at it. I really know how that run is could upload ads
that well. Like I thought there were API limitations and stuff. And I think it was either you or
Reza was like, no, you should try it. Like mine blown because I spent two hours wings on one Saturday
in like August uploading ads. And it was the lowest experience ever. And so like that is,
that's like to me the next step. So I context connectors, administrative work if you call it.
Is that fair? Yeah, I think so. And so actually something that we also worked on really heavily
with run it. I was very much like a run it hater for a minute there early on because our data
for for our ads looks very different than like your typical like wings like a supplement brand
or even hex clad. We're not using row as we're using four different KPIs. And we look at like
secondary metrics. And we just look at our data so much differently because we're generating leads
or event registrations or or forums or something like that. So right away, like not only did we
want to make sure that the data was clean. We wanted to make sure that the agent like understood
how to use the data. That was also so important. So that's also like the human aspect to you
because you could also just plug into it, plug in a data source. It could tell you something.
And then you don't even realize it's wrong because you don't understand how to correct it,
which is so important. But yeah, yeah, 100 percent. And then after the administrative stuff,
where are we going? Like where is where is AI today? And I totally agree about Reson. I used to
think he was crazy. And then it's like, he was right about this. I'm just going to trust him next time.
Where is it today with like actually producing the work, right? And let's call it making
ads, making ads, making funnels, like growth marketing. And where is it going today? Because I feel
like, yeah, where are where is it going in the future? I wish Reson was here so he could tell you
because he'd know. I know that today, yeah, today, today, it can do stuff that like I never imagined
possible. I think if there's one thing that everyone just needs to do a better job at,
especially if they're like, they feel like they're behind down AI is just ask the AI, like how
do you, how do you do this? Can I do this? So many people come to me or we'll comment on my Instagram
posts or we'll come to a CSM and say, can run us do this? Ask it and ask it how. And it will,
it will figure it out. That's how I figured out the run us get edit videos. I said, everybody says
that they're editing videos and clogged. Can you edit videos? And then it did. And I didn't have to
think about Python or any of these things that or any of these GitHub reposal people are downloading.
But yeah, at this point, like yesterday, for example, actually, you know what, Cody, you mentioned
months ago, I feel like on a podcast where you're like, I want to get to the point where we can
come up with an idea, we can make a landing page and we can make like a base set of statics
and launch that within two hours. I genuinely think that if you did that with wings right now,
it would do it. Oh, absolutely. Yeah, yeah, it's so crazy. Like yesterday. But it can do
way more than that. It can do it. I think it can do more today. Yeah, it can also launch the ads.
And it can do the name. It can write the copy. It can do it can do the whole thing. Even yesterday,
I made myself a landing page and launched it with GitHub behind a run with URL. I asked it to
script me a YouTube video. I went and filmed the YouTube video in 15 minutes and then it edited
the YouTube video and then it wrote the copy. It like, it did everything for me in a day. And it's
just like every single day, it just gets more magical. So in terms of like what it can, like,
there's nothing that it can't do at this point as long as you know how to like prompt it correctly.
But yeah, it's just a matter of like what you're told. Like it absolutely, it can, it can make an ad
for you a hundred percent. But then I, and then it becomes a question of like, where do the
humans come in? Because it can make AI avatar ads. But one thing that Dara did really well in her
talk at chief marketer and in NYC was she had some tips about like how to make AI avatar ads that
look like real humans, right? Because even that is like not a hundred percent. Like you can still
clock it a little bit. But yeah. Every Q4 brands pile money into meta compete against each other
in the same auction and battle rising ad costs where your customer's attention goes beyond just
scrolling on Instagram, holiday movies, TV shows and football. What if your brand showed up outside
the advertising rat race where the entire family is watching? That's where neon pixel comes in.
Neon pixel gets your e-commerce brand onto premium streaming networks, including HBO Max,
ABC, Paramount Plus and tons more. They make sure ads only show up for the households you want to
target. And they connect ad views to site visits and orders. So TV actually earns its place in your
performance budget. CTV through neon pixel is a big part of our media budget all year long at
Hackspad, but especially during BFCM. Last year in 2025, we scaled our CTV budget with neon pixel
287 percent in November compared to October. But the window to launch before Black Friday in
Cyber Monday is closing. So go to neonpixel.co to get a free CTV plan for Q4.
So I agree. I could take it further. So statics like absolutely can do some really good statics.
You guys just just showed people on next last week that run of the community landing pages.
And I didn't know that I was very impressed because again, I've spent a ton of time in cloud
encodax building these skills. And you know, I've talked about a lot and it just it just worked.
And that was like I've had so many aha moments with run of because like I've heard about people with
like instinct or whatever and they got it. And it's just like it just works. You don't have to train
it. And that's like people's like magical moment. And that's like how I felt. So landers definitely
don't you can give it an inspo lander and be like, hey, make it for us. And again, it's all about
having that context system because you have to have your design system with your fonts and tokens.
It's got to have, you know, the personas and they'll pull in all that stuff. And you can say, hey,
make, you know, make, make the seven reasons why using this as an inspo. Here's a spatula ad.
So absolutely. You can, again, it takes time up front to get your context in there. And then it's
boom, make this funnel, launch it for me, like hands off. Very simply editing definitely.
This is like how fast it's going. So when Astra came out, I saw people using Astra to edit videos,
using the DaVinci Resolve MCP. And so I started doing that. You got to build skills, right? It's
not just kind of one shot it, but you can build skills. And then now you don't need the MCP,
just opus 5.5, just edits videos. Again, Python, you don't have to do it. Alex Cooper,
I think put out a good video on this, like done. And then I, I, yeah, I, I'm trying to push my
limits as well, right? Like all these drop shippers and stuff like that. They're just like
doing a lot of AI ads. I am not great at making AI videos. And when I saw Dara's talk,
I was, and Jamie's, I was so impressed because I was like, okay, like it's actually not an AI thing.
I think there were a lot of marketers in that room that did not know that that video was AI.
Oh, yeah, even myself, like too, I was like, is this just because it's on like a bigger screen
that I can't like notice this? Because usually I'm pretty able to clock it right away, but I didn't
catch it. But she, she shared some tips after the fact too, that was like, make sure they're very
short sentences. Like they're, there's just some tips to make it feel a little bit more real.
And as soon as she said that I was like, you're so right, because once you see someone talking
for like 10 seconds, by that time they've like, they've just done some movement with their face
or something where you're like, this person looks a little off. You know, but if it's a quick
sentence, so there's lots of like tips to make that, yeah, it's pretty incredible. How far it's come?
Yeah. And so my takeaway, the reason her ad was so good, right? So some of her tips were, yeah,
don't, you know, cut everyone in half seconds, get a reference, build a character she and reference,
have, she prompted what the hands did, right? There were several things that they did. And plus,
it was really great creative strategy, just understanding of the persona and customer that like,
they weren't technical things. It wasn't like, oh, I used this tool or this API. Her or her team knew
how to prompt it because they actually had the creative background. It wasn't because they were
using different model or API. Like everyone is kind of using the same models now and using a
Higgs field or a seat dance or whatever. And then same thing when, when Jamie again and presented,
like, she was showing different prompts and different outputs. I'm like, and I felt like this,
like, I'll, I'll want something and I just can't get, I don't have the lexicon or the vocabulary
to prompt things a certain way. And be if I want like a certain reference or mood board or something
because I just don't know. And a really good creative person with AI is going to be like,
I want it on this camera, this lighting setup. And they can like recreate these vibes and setups
they want because they actually have the skill set behind it. Yeah. You know what's crazy too,
is like, you can just ask AI to write prompts to achieve what you want to. I feel like references
are so important for anything. Like if you give AI a reference, it will just use that reference
and it'll look incredible. And then the edge is where you like layer on top of that to be like, no,
but I want it like this, like you said, like on a camera. But also if you just like,
if you just whisper flow and yap at it to be like, I don't know how to describe this,
help me describe it. It might be like, oh, what if you said it was on this camera? Like it can help
you get there also, which is pretty incredible. Yeah, you can. I just think you can't replace that
like last 10%, which makes a huge difference. Is that like technical understanding and like the
expertise? And I truly think that's who's going to succeed in this new era. And I have so much
much respect because I know.
a lot of creatives are anti-i, but like people like yourself,
people like Dara, Orrin, you know,
people like him like that are really like leaning into it
and allowing themselves to be more creative
and produce output.
So I think that's where it is.
Where is it going?
Where does Reza talk about?
Where it's going?
Where do you guys think it's going?
Where is the future?
- I don't know if I could put it into words.
Like Reza has to be honest.
I mean, Reza really believes,
yeah, that like the tools are the easy part.
The skill is the important part.
And actually we see this like,
I think motion is partly responsible
for kind of flooding the industry with creative strategists
who there's this like this rhetoric
that's going on online right now
that it's really hard to find a good creative strategist.
And I really feel for people learning creative strategy
in this age of AI,
because for the first time they can just ask an LLM
to like help them along.
But what it truly takes to become a good creative strategist
is the reps is like add breakdowns, add tear downs,
making ads yourself like really like developing
the direct response skills,
understanding a hook, writing briefs for creators.
Like there's so much that goes into learning creative strategy
and it can feel so easy when AI helps you along.
But there are like some innate skills
that a creative strategist really needs to be talented.
And I think we're seeing right now
that like for the amount of tools that exist
and for the amount of work that AI can do for you,
people are still trying to hire creative strategists.
And it is very hard to find a good creative strategist.
So like there's obviously the skill that's still in demand
that is exactly what you're talking about.
It's like understanding why things work,
understanding how to prompt an LLM,
understanding how to like work with an agent,
understand just understanding the principles
of marketing honestly.
That was one of the reasons why they called it run-eth
for example, is like a reservoirally believes
that like the principles of marketing
don't really change.
So like the the old style name was
where that inspiration came from.
And it's so true, like the marketing is the hard part.
Once everybody has the tools, it is the marketing 100%
and like the skill set.
- I couldn't agree more.
I think that's where my tweet was like copyright
and it's gonna be the only skill that is left.
Like I don't actually believe that, it's like a tweet.
But it's I think that like the principles of it
'cause all the technical stuff is going away.
There's no dev stuff.
It's just how can you understand your customer?
How can you write effectively to them?
How can you know?
Yeah, like the LLMs don't know what a good hook it.
Like you can train it on it, but I just think.
Yeah, I think the more strategic stuff
and you know, like they're not gonna come up
with these like great campaigns and you know,
like even research.
I've been doing a lot of research manually
and because I just I think a lot of that stuff
is really important.
I am personally and I'm sure it'll get better.
And again, this is controversial.
Some people like I think it's pretty bad at writing copy.
I think it's I think it's pretty good
at writing average copy.
Yeah, I believe that like the 108010 rule applies
to almost anything like I think about that
about design all the time coming from a design background.
Also, we still have designers and they still like
our website runnith was built by runnith,
but we have designers who are working
on like the second iteration.
That's like the final iteration.
And I feel like that about anything I do.
So like if I give runnith like the idea for a script,
it can take it from 10% to 80% or 90%.
But then I still have to do that like final polish
that it so that doesn't have that like AI taste, you know?
And I think that's true about a lot of the creative aspects
of it, it was like there's something on top of it,
the human element that you have to add,
where like maybe you can get away with the AI version of it,
but it really is at like last 10% that's so important.
Most brands treat customer support as a cost,
but with AI, every support conversation
is now a chance to increase retention,
recover revenue and grow your LTV.
Rich panel doesn't just use AI to answer tickets,
they allow operators to run revenue experiments
inside your support.
You can test different winback offers on refund requests,
subscription save offers during cancellations
and product recommendations that turn chats into sales.
Plus they guarantee you're going to save money.
50% of your support volume automated in 30 days
or your money back, that's a pretty good offer if you ask me.
So at richpanel.com/demo,
they'll build an AI support team live for your business
and show you how it can drive incremental revenue for you.
Obviously, it's getting better, like again,
it's editing these videos.
It's not fully one-shoted, but it's doing that.
So like, mate, who knows?
It could be a model really so away
from being really good at the stuff,
but I don't think, and I saw a good thread on this.
Also, I think people have to understand the reason why
and I would love your take on this.
But number one, it's all about like what's a verifiable domain
versus not.
Like math and science, it's very easy
'cause they do reinforcement learning on these things.
And so it's either binary objective, yes or no,
they can train models based on those things.
Creative stuff, copy is very subjective, very hard.
And then number two, where are they financially incentivized
to do this training?
Probably less on copy, less on these things,
but coding, obviously a huge one, math, science, big one.
So that's why I think the models are not as good.
And why, if anything, they've gotten worse at copy
as the models have gotten so much better.
It's been like a, almost a necessary trade off.
That's interesting.
That definitely sounds like a CEO question to me for sure,
but yeah, yeah, it's interesting 'cause like,
there are things that we have to teach the agent.
Like, if you're building that landing page,
who are we building it for?
And how do we know what copy to write on the landing page?
And how do we know what assets to put
and which order to put it in?
Like, why doesn't it know that out of the box?
But it can do other things out of the box.
I don't know.
I guess I've never really thought about it that way.
Yeah, that's interesting.
'Cause I do have to write all these skills, yeah.
Yeah, but no, like I know a guy, he's got, you know,
not a huge brand, but it's kind of solo operating,
maybe one person, a brand.
And he's, he's running ads for it.
He's building landers, stuff like that.
And he's a creative person.
He's got a brand agency, it's like a short form creator.
I'm like, that's a perfect person
to be like solo operating a brand right now.
Because the technical side all kind of is handled.
He can open PR as do his websites, stuff like that.
But you know what?
It's not gonna, it's not gonna create content.
It's not gonna write and copy.
It's not gonna, you know, it's fine with design.
And like, so I think that's a perfect person.
So like, personally, if I was a creative,
like I would, I would be learning these things
and it's like the best time.
And I know a lot of creatives are very scared of AI,
but I actually think the ones that are embracing it
are crushing it.
- Yeah, I agree.
And that's where it becomes like,
what is the thing that I work on today?
Like what is the skill that I build today?
Do I learn how to use cloud code
or do I like break down 20 ads
and learn more about hook writing or script writing?
Or do I learn how to write a creator brief
or do I learn how to work inside of tribe?
Like what is the skill that I build today
is the big question, right?
And I think Reza and the runnith team
would argue that it's the second one.
It's the marketing skills that you want to build.
And like of course, you want to learn AI,
but what does learning AI actually mean, you know?
I'm not really sure if there's like a definition to that.
Like really, it could be as easy as opening up a tool
that eliminates all of that technical component to it.
Like news or a rock bottom runnith
where you can just learn how to have a conversation
with an agent.
Maybe that's what learning AI actually is now.
- Well, yeah, 'cause like again, you're, you know,
not in terminal stuff like that.
So you're not learning AI in terms of setting up
all of these complex agents,
but you're using runnith, you're training runnith, right?
You're bringing your skills, your expertise into it.
And then you're using it like a teammate, right?
Like, do you have any direct reports?
- No, I don't, not many people at runnith do.
(laughing)
- But you have runnith.
- I do have runnith, yes.
I think of runnith very much as like my direct report,
a hundred percent, yeah.
And I do, I build up, I have multiple runniths.
I have like my own runnith that I use for like content creation
and building up my own like repos and that type of thing.
And then we have a team runnith
that I've trained on our ads essentially.
And then that's what, we have someone
who helps like manage our creators.
We have another creative strategist on the team.
So what I really do is just like build up the VM
with all of our like skills, the personas
that we want to target our goals, that type of thing.
And then that becomes an asset for them.
So in that sense, yeah, it kind of is my direct report.
(laughs)
Yeah, I mean, I think about 'cause I don't have a team,
right, I have one assistant.
And so it's kind of like, all right, I want to get this done.
I got to, I got to go to an agent to get my landing page done.
If I want to make ads, like we have one designer,
but it's like, I'll go to runnith and design it.
If I have a video, it's almost easier for me now
to just toss it in and be like, hey, here's, you know,
an obviously there's skills that are trained on hooks,
but I find it so cringe when people on X are like,
here's my agent org chart, but like it is really going that way.
And I think agents can really handle a lot of the work
if people push it in our ambitious to do so.
- I feel the same way.
I'm like, what do you mean this agent org chart?
Like to me, that just looks like you're replacing people
and that just gives me the IK,
but I have an agent org chart now.
And I wouldn't have told you that a month ago
before I joined using Grogpop.
But now I'm like, I do have three agents
and then a bunch of different tools
and they all do different things
and they all work together and I can map it out for you.
And I like love managing that system.
It's so fun for me.
Cody, can I ask you a question about what you just mentioned
about like you made this comment about,
you're really the marketer in the room, you have a designer,
but like going back to the video editing bit,
so you've created a scale for how you want,
how you want your videos edited in cloud.
And can you just speak to that workflow?
So like you're getting presumably you're like
briefing in a creator, you're getting a bunch of raw assets back
and then you're taking those raw assets,
uploading it to cloud and saying,
hey, use this video editing skill to edit this
into like an actual ad that I'll launch in the account.
Like can you just speak to that process?
I think
I think it's helpful to be like, all right, step A, B, C, D, E, and I would love, especially
for like, I know what you guys do, like, like, pure, like, fully AI-generated video, but
I also know you guys are doing more and more, like, UGC from creators.
So, like, it's not fully AI power, like, there is a human being that is actually shooting
a video, and then you're taking that raw content and you're using this skill to get the
ad ready.
So, could you just, like, speak to what that process looks like?
Yeah.
Of course, great question.
So, I built one skill that has different workflows in it, and I think this is actually an
important thing.
It's like, like Alicia said before, like, ask the AI, can it do these things?
Or how it can do it?
Like, I'll use Westper Flow a lot, and I'll just say, hey, here's what I want to do.
Here's what you have access to.
What's the best way to do this, right?
And for example, I was like, 'cause I don't care about how the skills were set up.
Should it be one skill?
I want to do partnership ads, which is a different editing style than, like, direct response
UGC or, like, VS cells.
Should it be, here are examples of each of them.
Should it be multiple, right?
And it's like, all right.
It said, do one skill with different pathways, and then I just have to tell it, hey, here's
a partnership ad, right?
And then you have to train it.
And again, it's not, there's definitely stuff you could train.
And so the same way you train an editor, right, like, again, stop, you know, I think people
have to stop expecting AI to just one shot and just be like, oh, it can edit, let me just
give it an edit.
It's like, no, you have to, you have to really, you know, train these things up.
And so let's do partnership ad.
I give it that.
And the same thing I would brief an editor on, I just Westper Flow it.
Go to the creator's Instagram, find how they normally style their videos, the font that
they use.
It's supposed to look native.
And so you're just, you're just using the point.
I'm not, not that I'm an expert, but I'm using my marketing knowledge of what's going
to work for a partnership ad to edit these videos.
You know, I want minimal b-roll, I want it, minimal edits on these things, organic.
I want, I always want three hooks.
I always want their, because they send the Ross, I want three hooks, safe zone.
So whatever it is to train that and get in that's going on, not necessarily the individual
brief.
Right.
Yeah, yeah.
I do, to start the processes, I'll say, here's what I want to do.
I have this footage, it's a partnership ad.
I want to train, I want to edit these videos and I want to train the skill.
And as I'm giving you feedback throughout the way, I want you to save that.
And at the end of the session, I want you to propose how we will edit the skill for
this.
And so these skills are not like a one shot one and done.
It's like, it might be over the course of five different partnership ads that we're editing.
It kind of refines.
Because sometimes you do a skill and it, it makes one thing better, but something worse.
And so that's kind of like the process for it.
Alex Cooper did a really good YouTube video on this.
You know, he's got, he runs Parker and I great.
But he is having his editors, edit and opus instead of Adobe Pro.
And they are leaving their comments, all of their comments that they give to human editors.
They're just leaving them in a, in a software and a dam and training that to the model.
And just being like, all right.
They're the most common things we're saying.
And if you, you know, 10 other editors or strategists are saying, hey, you need to do this
with a B roll.
Like, all right, cool.
How do we make that a rule?
How do we make that a skill?
So I would just think about it the same way that you would, you would teach people, but
I think you have to be a little bit more explicit about like, okay, now I want you to remember
this and save that.
At Jones Road, we needed thousands of new ads every month, daily emails, lots of creative
as well.
We needed great people that had really high standards, but we're a bootstrap company on a budget.
And that is why we were propelled to recruit our team for us.
They are specialists in recruiting marketing and operations people from Latin America, and
they work with the best brands and agencies in Ecom.
They charge a one off fee once you hire their no recurring charges or ongoing fees.
They also offer a three month guarantee on all new hires, plus you can get 10% off your
first hire when you go to we are propelled.com/operators.
Yeah, that's basically how I did it to you.
Like I started with here is I started with my TikTok videos, I was like, here's what
my TikTok videos look like, and then also here's a few videos with effects or with animations
or what have you that I've not been able to pull off yet, because I just haven't had
the time, but I'd like to do them.
So first, like reverse engineer, all of these things that I've shown you, like breakdown,
the pacing, the fonts, the colors, the exact animations, the style of them, like the way
that you zoom in and zoom out and where you cut and where you might trim and like just reverse
engineer that and then I, I don't know, I think of things in like my own creative way.
So I was like, build a skill that we're going to call style file and that's like the style
of my TikToks and then build a separate skill that goes through the process of editing.
And that skill, what it tells the agent to do is just like, first do this first cut where
you take the transcript and you tell me where you're going to cut, pick the right take
and build me that first raw file where you can show me like the base, the baseline of
the video, so that I can just make sure that the pacing and the cuts and the structure
are good. And then we'll layer in the animations and the effects and then we'll layer in the
text and the sound effects. And then finally, I will say, okay, it's approved and then it
will export the final file because that's the other thing too is when you're managing
this, like using a lower res version of the video to do that initial editing, just gets
you to move a little bit faster. So there's a bit of like the style and then you can have
like different style files, right? Like here's one for a yapper, here's one for a problem
solution, UGC, here's one for a street interview, here's one for an animation, but then like
the process for me is still the same. And I would do it the same like the same way for
anything, just like reverse engineer something and then build on top of it.
So there's two bits here, you're building the scale, refining it along the way. And then
once that's dialed, not to say that you won't adjust it at some point because you will,
like you're going to get an output that's going to make you think, Oh, I don't want that
output, but if I adjust the scale this way, I won't get it anymore. But now, now all you
have to do is create the raw footage, right? Because now you're saying, Hey, I have this
skill belt. It's going to edit this exactly how I want it. So now like you said earlier,
you go and shoot a 15 minute YouTube video. That's the creator component. That's the part
that AI cannot do yet. Yet, maybe it will be at some point. And then you're pumping that
raw footage and saying, Hey, go make this whatever ad that you've already built to skill
on. And then in, I don't know how long that it's actually taking to edit it and like get
that output. But you go back to a Cody said, it's like, it just puts so much power in the
hands of the creator that can shoot engaging raw content and not have to go and spend either
hours editing themselves or hours passing it to an editor and going back and forth. And
like, we've all felt those pains of like, wow, why did this? Why did this landing page
design take a month? You know, yeah. And you know, what is someone who's who content
creation of runnes and Evan specifically, my manager at run it, like wants me to make
more content. One of my hang ups about it is because I'd have to send it to an editor
and I would get it back and be like, Oh, that's not like quite what I wanted. Like I don't
maybe it's just because I come from that design and create a background that like I had
a vision. And something about the ability to edit with AI has unwatched that for me where
it truly I can just work with AI and have like full control of the process, which has been
so great. And yeah, it does. I will say it will take you a lot longer the first time you
do it. And then the second time you do it, it might take you a little bit longer. And
then the third time you do it, it should take way less time because it will learn over
time. And yeah, you just have to have to ask it like once you're done with your edit and
you're like, I'm happy with this. You just ask, what did you learn from this conversation?
And how can you update the skill to improve it? And then every once in a while too, this
is just like a file management thing. I like to say like, look at this skill and find
any places where you're like where you're contradicting yourself where you can clean things
up where you can make things more organized, just in case it's like layers in too much. But
yeah, it really is just just learning over time. But it's pretty, it's pretty incredible.
Yeah, no, I've seen, I've seen, again, people like Aura and then other creators even
on Instagram are like super excited about it because I think for creators, correct me
wrong, like editing is like not the most fun part for most people. And so people are
able to create a lot more and create a lot more content because they're like, hey, I can
just shoot and yeah, and just send it over. Which, which is, again, I think a really
cool thing and probably really freeing. Yeah, it is like very, yeah, very freeing. I
mean, I had like a YouTube video and three videos get edited for me yesterday while
I was working. It was insane. Yeah, I have so many videos right now and even landers
where it's like they're 80% there and 9% that way there, like even these editing skills.
And I have the thing that is holding it back is me going in and training them. Like for
example, for these VSLs, it'll pull B-roll from an asset library or create some AI stuff.
But like it's not doing a great job of it, right? Like you all know a really good edit
where you have a really good sound effect. The visual is really a tight edit at the
right time, but also showing what's being talked about. It's not going to do that out
of the box, you know, until the models get better. And so you have to go in just like
you would a human editor and be like, switch this one out for this. And the more specific
that you can be with those. And then it'll learn and train it. So there's a lot of leverage
that goes into it, but it just takes a lot of time. Interesting. And you said it's like
it's not pulling B-roll the way that you wanted to. Is it about like not pulling the
right clip or is it about not using the clip effectively? Both. Both. So sometimes it'll
be like, hey, what's shown on screen? Like doesn't really match the emotion of the voice
over. Yeah. And then sometimes it's like, yeah, I want to add a sound effect or I want
it's just not as tight. It's not as like, yeah, so it's a little bit of both. But it can
do it if you train it on it. Yeah, that's the other part too is like asset management,
tagging, summarizing, like how it can understand what's inside of that B-roll and how you classify
it. Like if you said go go pull me an unboxing clip and you have hundreds of them and a few
of them are unboxings, but a few of them aren't like tagging is a great way to like get you
one step out of the way there. So at least it can like filter through things. So there's
just like so many levels to like what it's doing. You love tagging. I love my tag. It helps
so much.
Even like I have my own custom, like I said,
when you asked for founder ads in the VM,
it gave you my curated founder ads.
One of my favorite workflows that I have
is every single time I see an ad or a piece of content
that I love, I just say drop this in my repo,
drop this in my repo, drop this in my repo,
and then run it this train to give in a naming convention
actually in the file that names it
after the visual format that I would use,
the season, whether it's a creator content
or a brand piece of content, and to have that built up
so that now I have this like asset library
of when I need an example for something,
it knows how to go fetch them,
and I have like my own custom inspo board,
but that's like one of the things
that I've built up for me as an asset.
- Yeah, yeah, I agree with that totally.
- So the way that I actually muse
has changed everything for me about this one,
because I have private Instagram collections,
and I also have private TikTok collections,
so hopefully maybe TikTok
can come up with their own version of muse
to see my private collections,
but I always grouped them by like,
I would have a collection called Hooks,
a collection called Formats, a collection called Yappers,
I even had a collection called Winks
where I was like bookmarking things,
but then they would just like die in my Instagram collections,
the same way that if you have like a million screenshots
on your phone.
So I would send the links and Slack to Runnith,
but now muse can go through my collections
and send them all to Runnith,
and then Runnith can tag them,
and I don't have to touch anything.
I just, I don't even have those individual collections
that are labeled anymore,
because Runnith can just look at the ad
and understand how to label it.
So I just, I just favored it,
and it gets sent there like every day at midnight.
- Cool, I think we're probably at time.
This was an amazing episode.
I enjoyed it.
I could chat about AI and Creative Strategy all day,
but we'll call it there.
Thank you so much for coming on.
This was really fun.
It's been great working with you.
Where can people learn more about you?
Where can people follow you?
You mentioned all the content you're creating,
which is awesome.
It's really great.
I know you've been active on X,
so tell people where they can follow you.
- Yeah, my username on TikTok and Instagram
is alisha.md with the dot,
and then everywhere else it's alisha.md spelled out,
not that I'm a doctor.
It's meant to be a markdown file joke,
but that's my handle everywhere.
- Is it on for Instagram as well?
- Yeah, Instagram, TikTok, YouTube, X,
and you can find me on Instagram.
- We got a white list of ads for links through that.
- Yeah, we should.
- Okay, that's like a great white hat way
to make it seem like you're a doctor,
but can't get enough of that.
(laughing)
- Feel free.
Use it to your advantage.
(laughing)
- Awesome.
Well, thank you for coming on.
- Yeah, thank you so much for having me.
Podcast Summary
Key Points:
Alicia Bone, a non-technical creative strategist, has deeply integrated AI into her workflow by building contextual systems of skills, prompts, and data to enhance creative decision-making.
She emphasizes that AI’s greatest value lies in data analysis and performance insights, allowing marketers to democratize access to data and make faster, more informed decisions.
A core principle in her approach is the "108010 rule"
Alicia highlights that AI tools like Runneth enable marketers to bypass technical barriers by providing intuitive, natural-language interfaces that connect data sources and automate reporting.
She argues that creative strategy remains valuable because it combines deep customer understanding, storytelling, and prompt engineering—skills that AI cannot replicate.
The most effective AI adoption starts with onboarding tools to brand context, data sources, and team alignment, ensuring consistency and reducing confusion in creative outputs.
AI is increasingly capable of generating full campaigns—from ad copy and landing pages to videos and funnels—though the final human touch ensures authenticity, emotional resonance, and strategic nuance.
The future of marketing lies not in replacing human skills, but in empowering marketers with AI as a collaborative teammate, focusing on strategy, creativity, and customer insight.
Summary:
Alicia Bone, a creative strategist at Runneth, shares her experience of using AI as a powerful, non-technical collaborator in marketing. Her journey from graphic design to creative strategy highlights how AI has transformed her workflow by enabling rapid data analysis, automated reporting, and creative output generation. She emphasizes that AI’s strongest impact is in democratizing data access, allowing marketers—regardless of technical background—to analyze performance and make decisions faster.
Alicia stresses that while AI can generate content like ads, landing pages, and videos with minimal input, the final "last 10%" of creative work requires human judgment, emotional intelligence, and strategic insight. She argues that the future of marketing isn’t about replacing human creativity but enhancing it through AI tools. Her approach centers on building context-rich systems—such as branded prompts, personas, and data connections—so AI acts as a seamless teammate.
Alicia believes that the most valuable skills in marketing remain creative strategy, customer understanding, and effective storytelling, which AI cannot replicate. Ultimately, marketers must shift from technical execution to strategic thinking, using AI to automate routine tasks and focus on higher-level decision-making. This evolution makes AI not just a tool, but a trusted extension of human creativity, especially in fast-paced, data-driven environments.
FAQs
Non-technical marketers can use AI tools by focusing on building context—like brand guidelines, customer data, and creative strategies—so AI understands their needs. They don’t need to write code; instead, they prompt AI in natural language to generate ads, landing pages, or reports, trusting the system to deliver results based on their established creative direction.
A creative strategist guides AI by setting clear objectives, defining customer personas, and providing creative direction. They validate outputs, ensure brand alignment, and add the final human touch that AI lacks—like emotional resonance and strategic insight—making them essential in the marketing process.
AI can generate content quickly and efficiently, but it cannot replicate human creativity, emotional intelligence, or deep understanding of customer behavior. The best results come when AI supports human creativity rather than replaces it, especially in the final stages of campaign development.
AI connects to multiple data tools and platforms, pulling real-time performance data to create dynamic dashboards. This allows marketers to analyze campaign performance instantly, identify trends, and make faster, data-driven decisions without manually switching between spreadsheets or reports.
First, build context by uploading brand guidelines, customer data, and past campaign performance. Next, connect AI to key tools via APIs or MCPs. Then, define clear use cases—like ad creation or reporting—and train AI with real-world examples to ensure it produces relevant, accurate outputs.
The final 10% involves human judgment, emotional nuance, and strategic refinement. While AI can draft content or generate ideas, it’s humans who add authenticity, tone, and creative intent that make campaigns truly impactful and resonant with audiences.
Chat with AI
Loading...
Pro features
Go deeper with this episode
Unlock creator-grade tools that turn any transcript into show notes and subtitle files.