Latest / AI DTC WTF / How to Use AI for Creative Strategy Without Killing Your Best Ideas | Matthew Gattozzi

Transcript

How to Use AI for Creative Strategy Without Killing Your Best Ideas | Matthew Gattozzi · AI DTC WTF

• 0:03
  If you work in DDC and you use AI
• 0:05
  and you're wondering
• 0:06
  what the F is going on every week,
• 0:08
  this is your podcast.
• 0:09
  The AI to see
• 0:11
  F podcast
• 0:12
  your home for tactical
• 0:13
  tips, strategies, and ideas
• 0:15
  that you can implement right now in your
• 0:17
  AI workflows to make your brand or agency
• 0:19
  more money.
• 0:20
  Welcome to another episode of AI fuck!
• 0:24
  Here we have Matthew.
• 0:26
  Matt, thanks for joining us.
• 0:28
  Thanks for having me. Will appreciate it.
• 0:30
  Long time friend.
• 0:31
  Matt has obviously built
• 0:32
  a really incredible studio
• 0:35
  that works with e-commerce brands.
• 0:36
  Maybe you can just give us
• 0:37
  a bit of background
• 0:38
  on what you've built and sort of
• 0:40
  how you got there.
• 0:41
  Yeah, so I started getting the studios
• 0:43
  four years ago,
• 0:44
  and really what it's all about
• 0:47
  is combining
• 0:48
  strategy and production together.
• 0:50
  Traditionally,
• 0:51
  those two things are fairly separate,
• 0:54
  which I think creates bad incentives
• 0:57
  for both sides of what's being created
• 0:59
  and why it's being created.
• 1:00
  And so
• 1:01
  my goal was to make sure
• 1:03
  that I was building a studio
• 1:05
  that really took into account
• 1:07
  performance and business goals,
• 1:08
  because if you could understand
• 1:09
  why you're creating something,
• 1:10
  it would change how you go and create it.
• 1:13
  We work with many different brands
• 1:15
  across different industries
• 1:17
  and other companies as well,
• 1:19
  and really
• 1:19
  what we're providing them
• 1:20
  is kind of like a full stack
• 1:22
  of creative, right?
• 1:23
  So we're doing the in-person shoots.
• 1:25
  We're also doing
• 1:26
  creator content,
• 1:27
  static ads,
• 1:29
  edit iterations, all of it, as well as
• 1:32
  landing pages and
• 1:36
  looking at the data.
• 1:37
  So we
• 1:38
  kind of
• 1:39
  we kind of do a lot
• 1:40
  within the creative sphere.
• 1:42
  And, you know,
• 1:43
  I think when it comes to AI,
• 1:45
  really excited
• 1:46
  to kind of talk a little bit
• 1:46
  about what we're doing here,
• 1:48
  because I think the way I look at
• 1:50
  it is like I am very much
• 1:53
  a creative person.
• 1:55
  I don't consider myself technical.
• 1:58
  My brother is the software engineer.
• 1:59
  I was always the artist.
• 2:01
  So very two different minds, right?
• 2:04
  But it's been really cool
• 2:05
  to kind of
• 2:06
  see how
• 2:07
  I myself can actually build some things
• 2:11
  and use it
• 2:12
  so that we can make better creative.
• 2:14
  Right?
• 2:14
  Because that's what I care
• 2:15
  most about
• 2:16
  is making the best creative
• 2:17
  for our clients.
• 2:18
  And how do we use
• 2:20
  AI and be thoughtful about it.
• 2:22
  So yeah,
• 2:22
  that's that's a little bit
• 2:23
  about what we're building. Nice.
• 2:25
  And when I think about you
• 2:26
  and your agency, I think about,
• 2:29
  you know,
• 2:29
  you're thinking about lens type,
• 2:31
  you're thinking about
• 2:31
  where are we going to be doing this shoot
• 2:33
  for this product,
• 2:35
  or how can we source
• 2:36
  the best actors or talent?
• 2:38
  So it feels almost like the most human,
• 2:41
  or rather,
• 2:42
  is one of the most human agencies
• 2:44
  I'm aware of.
• 2:45
  So I'm curious, like,
• 2:46
  what has been your sort of
• 2:49
  journey with AI
• 2:50
  and where have you actually found
• 2:51
  like ROI there?
• 2:52
  I think it's a great question.
• 2:54
  I think that, you know, obviously,
• 2:57
  you know, it's not like
• 2:58
  we're replacing our video
• 3:00
  shoots all of a sudden with AI video.
• 3:04
  And I actually don't think
• 3:06
  that that's really
• 3:06
  what's going to happen.
• 3:07
  But as you said,
• 3:09
  there's a lot of human components
• 3:11
  to how you get a production done.
• 3:13
  And there are a lot of things
• 3:14
  that people don't really know.
• 3:16
  But the reason
• 3:17
  why it's really hard
• 3:17
  to replicate what we're doing is
• 3:19
  because there is a lot of crappy
• 3:21
  hard work that has to get done,
• 3:23
  and those things can be
• 3:24
  you got to coordinate
• 3:25
  a bunch of people's schedules
• 3:27
  in a physical location, okay?
• 3:29
  When you're dealing
• 3:30
  with the physical world,
• 3:31
  which is what we do.
• 3:32
  We have to deal with things
• 3:33
  like a lot of communication
• 3:36
  across many different people.
• 3:37
  You have to deal with shipping.
• 3:39
  You've got to deal with a lot of things
• 3:40
  that you don't really think about,
• 3:41
  but it really stacks up.
• 3:42
  And when you start to scale
• 3:45
  an organization
• 3:46
  where you're doing
• 3:47
  multiple shoots a week,
• 3:49
  we're doing 2 to 3 shoots a week.
• 3:51
  That's a lot to be planning out,
• 3:53
  not only week to week,
• 3:54
  but month to month.
• 3:55
  And so
• 3:57
  not only do you have a lot of like, human
• 4:01
  like elements
• 4:03
  that you have to coordinate,
• 4:04
  you have a lot of ideas,
• 4:06
  a lot of writing,
• 4:07
  a lot of things that have to get done.
• 4:09
  The way I look at AI is not necessarily
• 4:12
  that it's a replacement,
• 4:14
  but I think about what are the most
• 4:16
  if you are like as a human, as me,
• 4:19
  I have Matthew tokens, right?
• 4:21
  Like, I've got what my what I'm best at
• 4:24
  and the efforts that I'm going to do
• 4:26
  if I can not spend my tokens on
• 4:30
  writing a recap slack message, and
• 4:33
  I can use my tokens to be more strategic,
• 4:38
  now I'm getting more value
• 4:40
  out of my tokens while offloading,
• 4:43
  you know, the smaller things to that.
• 4:45
  And I look at that across
• 4:46
  my organization.
• 4:47
  I look at my creative strategist
• 4:50
  and I say,
• 4:51
  I don't want them to necessarily be like,
• 4:54
  not sink or not do writing.
• 4:56
  But there's a lot of parts of the job
• 4:58
  that are coordinating.
• 4:59
  There are things that just take
• 5:01
  time that you've got to do,
• 5:02
  and if we can compress that time for them
• 5:05
  on some things,
• 5:06
  it opens up
• 5:08
  more time for other things, right?
• 5:10
  So if we can compress
• 5:11
  their communication time
• 5:13
  writing slack messages,
• 5:15
  recaps, follow ups, etc.,
• 5:17
  and now spend more time being like,
• 5:19
  what's the best idea?
• 5:20
  Or my producer,
• 5:22
  if she can answer all of her emails
• 5:24
  a lot faster
• 5:25
  now, she can think more about
• 5:26
  what's the best talent,
• 5:27
  what's the best location for this place
• 5:29
  to make the best video as possible.
• 5:31
  So it's very much about
• 5:34
  not usually replacing,
• 5:35
  but more about where
• 5:36
  can you compress
• 5:37
  certain things
• 5:38
  so that you can
• 5:39
  spend more time creatively.
• 5:41
  I think the thing that
• 5:42
  when many people they
• 5:44
  when you're not doing creative work
• 5:47
  full time,
• 5:48
  you think, oh,
• 5:50
  I just want to do
• 5:51
  if only I could be
• 5:52
  a videographer full time.
• 5:53
  If only I could be whatever, full time.
• 5:56
  That's usually kind of how people feel.
• 5:58
  And then they actually maybe work
• 6:00
  hard enough
• 6:00
  and they get good enough to be in
• 6:02
  that full time role.
• 6:03
  And then they realize that,
• 6:04
  well, actually, 40,
• 6:06
  maybe even close to 60% of that
• 6:08
  time is actually
• 6:08
  not even the creative work itself.
• 6:09
  And if we can actually, like take down
• 6:13
  that 4,060%
• 6:15
  of non
• 6:15
  creative work to
• 6:17
  3,020% and then expand it to hey,
• 6:21
  now you're spending
• 6:22
  80% of your time, 70% of your time
• 6:25
  on actual creative work.
• 6:27
  The value to the client is way
• 6:29
  higher, as well as the value
• 6:30
  to them as a person
• 6:32
  and what they get to do as well
• 6:33
  in their career.
• 6:34
  And so these are the things
• 6:35
  that I'm really thinking about.
• 6:37
  And then I'm thinking about
• 6:38
  how do these tools and how do we use
• 6:40
  AI within the organization
• 6:42
  to do that thoughtfully?
• 6:44
  And I don't think that like,
• 6:45
  we do anything
• 6:46
  like super crazy or super fancy,
• 6:48
  but I think it's enough to get that alpha
• 6:51
  that I think most people probably want.
• 6:53
  And I think there's
• 6:55
  some really practical things
• 6:56
  you could do.
• 6:57
  There's some cool, complex
• 6:58
  things you can do,
• 6:58
  but there's a lot of like simple things
• 7:00
  you can do that
• 7:01
  if you can just get
• 7:02
  most of your team members to do that,
• 7:04
  like you're you're
• 7:06
  kind of ahead of the game.
• 7:12
  We all hear a ton about AI.
• 7:14
  We all hear about what it can do
• 7:16
  and how amazing it all is.
• 7:17
  And I don't think that there's
• 7:19
  anything incorrect about that.
• 7:20
  But this episode is brought to you
• 7:22
  by an AI brain
• 7:24
  that is absolutely incredible
• 7:25
  for your marketing team called runneth.
• 7:27
  It's from motion.
• 7:28
  And therefore, you know, it's awesome
• 7:30
  because it's from motion.
• 7:31
  But really, what runneth does is connects
• 7:34
  every single function
• 7:35
  of your marketing department
• 7:37
  into one place.
• 7:38
  It can do so many different things.
• 7:40
  It can build you ads.
• 7:41
  It can verify
• 7:42
  that your brand looks proper.
• 7:44
  It can verify
• 7:45
  that what your top creators were
• 7:47
  and you can ask it,
• 7:48
  you can have it live in slack,
• 7:49
  or you can have it live as a CLI
• 7:51
  so you can interact with it
• 7:52
  any way you want to.
• 7:53
  The thing
• 7:54
  that's really insane to me
• 7:55
  about motion
• 7:56
  is how absolutely intelligent it truly is
• 7:59
  and how it works cross
• 8:00
  functionally across the team.
• 8:02
  There was recently
• 8:03
  a marketer
• 8:04
  that Reza talks about in the landing page
• 8:06
  on run,
• 8:07
  which you should definitely check out,
• 8:08
  where basically she went on vacation.
• 8:10
  The CMO did and said
• 8:12
  if run says the ad is okay to run,
• 8:14
  then it is okay to run.
• 8:16
  So she uploaded all the context
• 8:18
  for the team
• 8:19
  they ran it into, Run It and run
• 8:21
  said, yeah, that ad is totally on brand.
• 8:23
  It's absolutely crazy
• 8:24
  the number of things that this can do.
• 8:25
  And I think the fact
• 8:26
  that we all live in silos
• 8:28
  and the fact that we all live separately
• 8:29
  is really, really challenging.
• 8:30
  And one of the hard parts about AI.
• 8:32
  So that's not what this does.
• 8:33
  This brings it all together
• 8:35
  with shared context,
• 8:37
  shared resources,
• 8:38
  and it also looks at thousands of ads
• 8:41
  every day.
• 8:41
  So it's always up to date on
• 8:42
  what actually is the best performing ads
• 8:45
  that it's seeing across motion.
• 8:46
  And so that's another whole thing
• 8:48
  that it does.
• 8:49
  It's really crazy
• 8:49
  because instead of one more dashboard,
• 8:51
  you get a teammate in your slack
• 8:52
  that knows your performance data cold.
• 8:54
  As I said,
• 8:55
  you can ask it really anything
• 8:56
  and answers instantly.
• 8:57
  Weekly recaps top creatives.
• 8:59
  Which creators your audience trusts.
• 9:01
  It learns your brand,
• 9:02
  your playbooks,
• 9:03
  and get smarter every day.
• 9:04
  E-commerce Adidas
• 9:06
  teams are using it
• 9:07
  to turn weekly reporting
• 9:09
  from a two day job
• 9:10
  into a two minute conversation
• 9:11
  head to runneth and see it for yourself.
• 9:14
  And it also has a 30
• 9:15
  day money back guarantee.
• 9:17
  So if you need a cheat code in e-commerce
• 9:20
  and to see with your ads
• 9:22
  and your ad creative
• 9:23
  and your marketing team,
• 9:24
  check out runneth com
• 9:26
  30 day money back guarantee.
• 9:27
  It's absolutely incredible
• 9:29
  what this thing can do
• 9:30
  and I encourage you to give it a shot.
• 9:31
  Got it.
• 9:32
  That makes a ton of sense.
• 9:33
  So just it is still done.
• 9:35
  What you've
• 9:37
  said,
• 9:38
  the idea is like you're trying
• 9:39
  to just free up your team
• 9:40
  to do
• 9:41
  creative work
• 9:42
  and effectively offload
• 9:43
  a lot of the admin to AI.
• 9:45
  Is that right?
• 9:46
  Yeah, I would say that.
• 9:48
  I think like there's I don't think that
• 9:51
  AI can right now just I,
• 9:54
  I think there's a lot of value as a human
• 9:57
  to be able to,
• 9:59
  especially in creative work.
• 10:00
  So I'm speaking
• 10:01
  kind of more for,
• 10:02
  for, for our role and how we,
• 10:04
  what we do for brands.
• 10:05
  But I think that when you think about
• 10:07
  what makes a great strategist
• 10:09
  in general is
• 10:10
  when somebody can take something
• 10:12
  that they see inspiration,
• 10:14
  they take that,
• 10:15
  and then they're able to relate
• 10:16
  that back to the brand
• 10:18
  and how they're going to use it.
• 10:19
  So maybe they are on YouTube
• 10:21
  and they see this really cool
• 10:22
  format of another video about whatever.
• 10:26
  We ask somebody 72 questions.
• 10:29
  Oh, we could do a 72 questions
• 10:31
  video, right?
• 10:32
  And it's like, you know,
• 10:34
  AI's doesn't just like say like
• 10:36
  you guys should do a 72
• 10:37
  question video, right?
• 10:38
  Like that doesn't happen.
• 10:39
  Like the ideation part
• 10:40
  doesn't happen that way.
• 10:42
  And so like
• 10:43
  we need to actually create more space
• 10:45
  so that like
• 10:47
  our just can just exist in the world
• 10:51
  because it's like by existing,
• 10:53
  by doing things, by being out and about,
• 10:56
  they actually are,
• 10:57
  that's where they're going
• 10:58
  to get their best inspiration
• 10:59
  because the best strategies are can like
• 11:01
  take things
• 11:02
  and then put it for the brand.
• 11:04
  Okay, I saw this
• 11:05
  and now I'm going to take that, I'm
• 11:06
  going to bring it to this brand,
• 11:07
  I'm going to do all these things.
• 11:09
  I'm going to take this format
• 11:09
  onto this style.
• 11:10
  Wow.
• 11:11
  I love how they like,
• 11:12
  explain this thing on this TV show.
• 11:15
  Now I'm going to take this
• 11:16
  and do that or like whatever. Right.
• 11:17
  And so I think that like
• 11:20
  if I have my strategist
• 11:22
  having to like be behind
• 11:23
  the computer 24 over seven
• 11:25
  and then I'm like,
• 11:26
  why aren't you making great ads?
• 11:27
  It's like,
• 11:28
  that's probably
• 11:29
  that's probably why they're not existing.
• 11:31
  They're not living. Right.
• 11:32
  And so I'm thinking a lot more about
• 11:34
  how do we free up their time
• 11:36
  so they can live and exist
• 11:37
  so that they can come up
• 11:38
  with the best ideas possible?
• 11:40
  That makes sense.
• 11:41
  So let's pull in that
• 11:41
  thread a little bit.
• 11:42
  So let's just assume
• 11:43
  your strategist
• 11:44
  found a 72 questions video.
• 11:46
  Walk me through end to end
• 11:48
  what that process would look like.
• 11:49
  And where does AI
• 11:50
  sort of fit in throughout.
• 11:52
  Is it you know,
• 11:53
  is it helping to brief and if so,
• 11:55
  what does that actually look like?
• 11:56
  Is a strategist
• 11:57
  going back and forth with AI tell brief.
• 11:59
  Is it, you know, for revisions?
• 12:01
  Is it you know for editing.
• 12:02
  Like where exactly is
• 12:03
  I fitting into your workflow
• 12:05
  that is streamlined it to this to,
• 12:07
  to a point
• 12:08
  where you have freed up your strategist
• 12:09
  to do like the messy human. Yeah.
• 12:12
  So I think that every shot
• 12:14
  just on my team
• 12:15
  has a bit different of a process,
• 12:17
  which I think is good because I think
• 12:18
  like they people think
• 12:20
  and work in different ways.
• 12:21
  So some people might really have
• 12:23
  like more of the visuals first,
• 12:24
  or some people might have more
• 12:26
  like the copy kind of first.
• 12:27
  But really
• 12:29
  the process
• 12:29
  is kind of the first thing,
• 12:31
  which is like one we have already
• 12:36
  have a document that's for the brand.
• 12:39
  And I think that like
• 12:41
  this is a very simple thing to do,
• 12:43
  but essentially
• 12:44
  to get the best outputs of AI,
• 12:46
  you need good inputs, right?
• 12:47
  So having some sort of
• 12:48
  context document is extremely valuable.
• 12:51
  And always updating
• 12:52
  that context document is valuable.
• 12:54
  So if if you are a brand yourself
• 12:57
  listening to this,
• 12:58
  make your own document about that
• 13:01
  and you might wonder, well,
• 13:03
  how do you know what you want in there?
• 13:05
  Well, there's some things that are
• 13:07
  I don't want to say proprietary to
• 13:09
  to to good.
• 13:10
  But the things that we think
• 13:11
  about in our creative strategy,
• 13:12
  we have included.
• 13:13
  But there's a lot of other things
• 13:14
  that are very like
• 13:15
  basic and simple,
• 13:16
  like reasons by hesitations to buy
• 13:18
  things that like,
• 13:20
  you know,
• 13:21
  I think everybody would want to have
• 13:22
  like a cheat sheet, if you will.
• 13:24
  If you're wondering how to do that,
• 13:26
  you would just go
• 13:27
  just ask AI, like,
• 13:28
  if you don't know how to do something,
• 13:30
  just ask AI.
• 13:30
  Like,
• 13:31
  if I need to give you everything about
• 13:33
  our my business, what do I need to do?
• 13:35
  How do we need to build it right?
• 13:36
  Like so.
• 13:37
  Always ask
• 13:38
  AI how you can get more out of AI
• 13:40
  and like it will help you do that, right?
• 13:42
  So just has this idea there.
• 13:45
  Like this is really great.
• 13:47
  I want to do this. All right.
• 13:49
  That already is a very messy process
• 13:52
  because there's probably
• 13:53
  a lot of self-selecting of what
• 13:55
  not to do. Right.
• 13:56
  So that's something
• 13:57
  that's like not being done in AI.
• 13:58
  And like, again, everybody's
• 13:59
  going to have a little bit
• 14:00
  of a different process
• 14:02
  and now goes into
• 14:04
  kind of like the writing of it.
• 14:05
  Let's say we're like,
• 14:06
  okay, we're going to do this idea.
• 14:08
  This is kind of
• 14:08
  what it's going to look like, whatever.
• 14:10
  Then it's
• 14:11
  like sometimes people are writing like,
• 14:15
  what am I trying to just
• 14:16
  she loves going on walks
• 14:17
  and kind of like writing down notes.
• 14:19
  So maybe she's like putting down notes,
• 14:21
  kind of like fragmented.
• 14:22
  So not like completely writing,
• 14:24
  but being like,
• 14:24
  I want to talk about this.
• 14:25
  I want to talk about this.
• 14:26
  I want to talk about this right
• 14:27
  then maybe goes into AI
• 14:30
  and then it's like
• 14:31
  has the context document.
• 14:32
  We also have another context document
• 14:35
  that's about like script writing.
• 14:38
  So if it's a more traditional
• 14:39
  kind of like UTC, that might work.
• 14:42
  If she wanted to do something
• 14:44
  that is like more of
• 14:45
  like the 72 questions,
• 14:47
  she might take that and say,
• 14:49
  like, we're going to do a video
• 14:50
  about 72 questions, adding that.
• 14:53
  And then maybe she's like
• 14:54
  speaking on like,
• 14:55
  I want these things,
• 14:57
  this is what I want to hit on Da.
• 14:59
  And then like maybe editing
• 15:00
  and working off of that, like that script
• 15:04
  that's coming up
• 15:04
  and then being able to say yes or no,
• 15:08
  change it.
• 15:08
  Sometimes they'll take it off
• 15:10
  and write it on their own
• 15:12
  and then put it back into AI as well.
• 15:15
  So it's kind of like like CI less
• 15:19
  so about like this needs to complete it
• 15:22
  from end to end,
• 15:23
  but it can get that jumpstart
• 15:24
  to get the juices flowing.
• 15:25
  And then saying, oh, we could do this
• 15:27
  and changing it.
• 15:28
  So there's a lot of like messiness
• 15:30
  within it. Right?
• 15:30
  So it's not like trust
• 15:32
  whatever comes out or whatever.
• 15:34
  But I think playing with it
• 15:35
  as best as possible.
• 15:37
  Another way of doing it would be
• 15:40
  you do
• 15:41
  create the
• 15:42
  initial shell of the ad itself.
• 15:45
  Like, I've even done this recently.
• 15:47
  I actually write out the script.
• 15:49
  I then put that script into
• 15:51
  AI to kind of like
• 15:52
  buff it up a bit, right?
• 15:54
  Okay,
• 15:55
  that claim based on
• 15:56
  the context document is like,
• 15:58
  not maybe
• 15:58
  I wrote the percentage wrong or whatever,
• 16:00
  and it's going to fix it for me.
• 16:01
  It's going to edit some of those claims.
• 16:03
  It might add
• 16:04
  in a few additional reasons
• 16:06
  to buy or whatever.
• 16:07
  So like sometimes
• 16:08
  you might have that core idea
• 16:11
  and kind of
• 16:11
  exactly what you want to hit on,
• 16:13
  what you want to see,
• 16:14
  you plug it in
• 16:15
  and then like
• 16:16
  it can expand it from there.
• 16:17
  So cool.
• 16:18
  There's a few ways you can use that.
• 16:21
  But I think that, like
• 16:22
  the most important thing for anybody
• 16:24
  to practically be able to kind of like,
• 16:26
  right, would be
• 16:27
  you need to spend time having like,
• 16:30
  what is good writing for you?
• 16:31
  Like what's your style
• 16:32
  and things like that.
• 16:33
  You need to have like core, like,
• 16:35
  what is my brand?
• 16:37
  What is like the reasons why?
• 16:38
  And whatever you find most important
• 16:41
  because that is how you're going
• 16:42
  to be able to get kind of better outputs
• 16:46
  that you can work with.
• 16:48
  So those are kind of the process.
• 16:49
  But the last thing I'll say is
• 16:51
  this is that if
• 16:52
  you are compressing your time,
• 16:53
  you can actually spend more time
• 16:55
  thinking like,
• 16:56
  what's the best way of writing this?
• 16:58
  What's the best visuals?
• 16:59
  So you can let your
• 17:00
  you can let your process be messy
• 17:02
  because you have more time
• 17:04
  to really flesh out these ideas
• 17:06
  and really think about.
• 17:07
  Yeah, okay.
• 17:08
  So I think there's a few important things
• 17:10
  you touched on there.
• 17:11
  So it sounds like this context
• 17:13
  document is sort of your foundation,
• 17:15
  your bedrock.
• 17:16
  And so you're making that individually
• 17:19
  for each brand.
• 17:19
  But presumably there are also elements
• 17:21
  that are congruent
• 17:23
  throughout each of these documents.
• 17:25
  And it also sounds like this context
• 17:27
  document is also a bit of a fact
• 17:28
  checker for you as well. Right?
• 17:30
  Like you are using it at the front end
• 17:32
  and the back end.
• 17:33
  So how is that context document
• 17:35
  evolving
• 17:35
  and how
• 17:36
  are you actually tactically using that.
• 17:39
  Like are you just like
• 17:39
  is it a markdown file.
• 17:41
  Like how are you
• 17:41
  cross-referencing the context doc
• 17:43
  as you are building these briefs?
• 17:45
  So I think I've mentioned this to you
• 17:49
  in our group chat,
• 17:50
  but we've been using this thing called HQ
• 17:54
  and it's from Jacob, Jacob and Team.
• 17:59
  I think it's like Indigo, Indigo HQ.
• 18:02
  I'm probably not doing my best to
• 18:04
  to I'm not sponsored, so whatever.
• 18:07
  But what is what's it really is good at?
• 18:10
  And I think this is actually
• 18:12
  what's most important
• 18:13
  is that one of the hardest things
• 18:15
  about using AI
• 18:16
  within an organization,
• 18:18
  outside of yourself,
• 18:19
  is sharing
• 18:20
  those context documents,
• 18:22
  as well as always updating them.
• 18:24
  Right.
• 18:25
  So the point of HQ is that it
• 18:28
  is your context layer.
• 18:30
  Okay.
• 18:30
  So good Studios,
• 18:32
  we have all of our documents in there.
• 18:34
  Okay.
• 18:35
  What's cool about this
• 18:36
  is that it is stored locally and you're
• 18:40
  working on cloud code or Codex.
• 18:42
  It doesn't really matter.
• 18:43
  You can
• 18:44
  you can decide whatever model you want.
• 18:46
  So that's another value
• 18:47
  is that as models change you can.
• 18:50
  Oh I'm going to jump to Codex.
• 18:51
  But I don't have to
• 18:52
  like ship off all of my Claude documents
• 18:56
  and projects over there.
• 18:57
  I can just plug one on top
• 18:59
  or change it out. Right.
• 19:00
  Because you own that context layer.
• 19:03
  Now that's like stored locally,
• 19:04
  but the app or the software
• 19:06
  or whatever
• 19:07
  allows you to sync it up into the cloud
• 19:10
  and then sync back.
• 19:11
  So if I change something about
• 19:13
  the document, it sinks up into the air.
• 19:16
  And then the other person on my team,
• 19:18
  it sinks to their their desktop as well.
• 19:22
  And so they are also getting that update.
• 19:24
  So it makes it really easy
• 19:26
  as opposed to being like
• 19:27
  hey guys, updated this.
• 19:29
  Make sure you know
• 19:30
  you've got the correct markdown file,
• 19:31
  all that stuff right.
• 19:33
  And so that is a really huge gain
• 19:36
  because I think
• 19:37
  team context documents
• 19:39
  is really hard to like do.
• 19:40
  So like as I see outputs
• 19:43
  I might be like like
• 19:45
  I feel like I need to work
• 19:46
  that copywriting document again
• 19:49
  and I might rework it,
• 19:50
  I might add to it or whatever.
• 19:52
  It also allows other people
• 19:53
  to add to it, change to it as well.
• 19:55
  So it's a living,
• 19:56
  breathing document versus
• 19:57
  always having to reset it.
• 19:59
  So that's practically
• 20:01
  how we're sharing things and doing it.
• 20:03
  I think that like
• 20:07
  the the
• 20:08
  best thing you can
• 20:09
  do is, is spend the time doing the setup.
• 20:13
  Right.
• 20:13
  So like spend the time
• 20:15
  really thinking about what is
• 20:19
  what are your filters,
• 20:20
  what do you think is good writing?
• 20:21
  What do you think is good statics?
• 20:23
  Like all of those things,
• 20:24
  I think most people just start.
• 20:26
  They just keep creating.
• 20:27
  They're just like, I gotta make more ads.
• 20:28
  I gotta
• 20:29
  and it's like, pause for a second.
• 20:30
  Like what's like,
• 20:31
  what do you really think is good?
• 20:34
  And think about that
• 20:35
  a bit more and,
• 20:36
  and be thoughtful about it.
• 20:38
  And it's worth spending.
• 20:40
  Like you could spend just one day
• 20:42
  like is really
• 20:44
  you don't need that much time
• 20:45
  to set a lot of these things up.
• 20:46
  It's been one day and like just talk to
• 20:49
  AI and be like, I like this.
• 20:51
  I don't like this, I want this.
• 20:52
  I, you know,
• 20:53
  and like just really shaping it.
• 20:56
  And again, like,
• 20:56
  I am a non-technical person,
• 20:59
  but like asking AI like,
• 21:00
  hey, this is the outcome that I want.
• 21:03
  How can I get that outcome?
• 21:05
  Like what documents do we need?
• 21:07
  And like,
• 21:08
  can you build me a plan to do that?
• 21:09
  And it's like, yes.
• 21:10
  And then it will interview you all.
• 21:13
  Like it'll be like,
• 21:14
  here's all the questions that we need
• 21:16
  so that we can build this document.
• 21:17
  And you're like, great, okay, cool.
• 21:19
  So you're still
• 21:20
  putting yourself in it, right?
• 21:22
  You're not just being like,
• 21:23
  well, AI gave me this.
• 21:24
  This is what it is.
• 21:25
  It's like you can shape it,
• 21:27
  but you've got to spend that time.
• 21:28
  But it's not like you got to spend
• 21:30
  like 100 hours to do this.
• 21:33
  Like spend a day, you know, and just yap
• 21:37
  into the into Claude and have it
• 21:39
  ask you questions about your expertise
• 21:41
  or whatever.
• 21:42
  And it can really shape that document
• 21:43
  and get you like a pretty good output,
• 21:46
  way better than what it was before.
• 21:48
  And then you're not repeating that.
• 21:49
  And then if
• 21:50
  you do have multiple team members
• 21:52
  and whatnot,
• 21:53
  because we were sending
• 21:54
  prompts back and forth,
• 21:55
  we had like a full doc, like,
• 21:58
  like with a table and being like,
• 22:00
  copy and paste this prompt into this.
• 22:02
  And then we were trying to like do
• 22:04
  Claude projects and all that stuff,
• 22:05
  but it just was like so difficult.
• 22:08
  Now it's like way easier
• 22:10
  to have this whole like, HQ setup.
• 22:13
  Got it. Okay, cool.
• 22:15
  So what I'm gathering from what you're
• 22:18
  what you're outlining here,
• 22:19
  is that you're having AI interview
• 22:22
  you into building this context document,
• 22:25
  and you're effectively
• 22:27
  uploading your taste,
• 22:30
  if you will, into the AI.
• 22:32
  And I think that word taste
• 22:34
  is obviously in the zeitgeist right now.
• 22:36
  It's you go to New York, San Francisco.
• 22:38
  That's all I can talk about is test,
• 22:39
  test, test.
• 22:40
  Alex taste.
• 22:41
  So like, how are you?
• 22:43
  How are you sort of thinking about like,
• 22:45
  that human taste as you are
• 22:48
  briefing content.
• 22:49
  My other question is,
• 22:51
  are these context documents?
• 22:53
  It sounds like you may have
• 22:54
  multiple, right?
• 22:55
  It sounds like maybe there's one
• 22:56
  that is very focused on
• 22:58
  you as the tastemaker.
• 23:00
  And then maybe there's another one
• 23:02
  that's focused on the brand
• 23:03
  specifically, like
• 23:04
  how are you
• 23:05
  sort of thinking
• 23:05
  about these different verticals
• 23:07
  as you're building
• 23:08
  your, your briefs, your assets, etc.?
• 23:11
  Yeah. So
• 23:14
  it's always funny
• 23:15
  when people talk about taste
• 23:16
  because I'm like,
• 23:17
  most likely
• 23:17
  the people that talk about taste
• 23:18
  don't have the taste because.
• 23:23
  That's all these Silicon
• 23:24
  Valley bros being like, yo,
• 23:25
  it doesn't have taste.
• 23:27
  Yeah, exactly.
• 23:28
  So,
• 23:31
  yeah.
• 23:32
  So anyway, let that let
• 23:35
  that sit for what it needs to be.
• 23:38
  But jokes aside, I think, you know,
• 23:40
  when I had just mentioned about, like,
• 23:42
  a person
• 23:43
  being able to
• 23:45
  see something in the world
• 23:47
  and then find
• 23:48
  how it can come back
• 23:49
  and be a creative idea,
• 23:52
  that's actually
• 23:53
  where the taste really exists.
• 23:55
  And you can only develop taste by like,
• 23:58
  experience, unfortunately.
• 24:00
  Like it's
• 24:00
  not something you can just like,
• 24:02
  I just have it.
• 24:04
  It's like,
• 24:05
  no, like you kind of need to like watch
• 24:08
  a lot of content and be like,
• 24:09
  what's good, what's bad and just know it.
• 24:11
  Right.
• 24:11
  And like,
• 24:12
  you can refine that palate,
• 24:14
  if you will,
• 24:15
  of what's going to work
• 24:17
  and being able to be creative.
• 24:19
  Like I actually think that's like,
• 24:22
  you know,
• 24:22
  creative director tends to be this
• 24:24
  like very ominous, vague term.
• 24:26
  And it can be from like
• 24:28
  a title perspective.
• 24:29
  But I think that's like,
• 24:31
  you know,
• 24:31
  you think about the people
• 24:33
  who are putting together
• 24:35
  like beautiful Pinterest boards.
• 24:36
  You're thinking about people
• 24:37
  who are grabbing inspiration
• 24:39
  all the time.
• 24:39
  Like,
• 24:40
  those are people that are really,
• 24:42
  really good
• 24:42
  at being able to generate ideas.
• 24:46
  So I think that, like,
• 24:48
  your ideation should really come from you
• 24:52
  as a human.
• 24:52
  And I think that's where it's like,
• 24:53
  I want to compress
• 24:55
  so much of other things
• 24:56
  so that they have more time
• 24:57
  to just sit and exist
• 24:59
  and try to find interesting things.
• 25:01
  Like if I'm thinking
• 25:03
  about a lot of my client conversations
• 25:05
  that I've had in the last
• 25:06
  two days, it's Wednesday.
• 25:07
  So Monday and Tuesday
• 25:09
  I had a conversation
• 25:11
  with a telehealth company
• 25:13
  that we work with
• 25:14
  because I had another friend
• 25:15
  tell me about something
• 25:17
  with their health,
• 25:18
  and it made me think, oh,
• 25:20
  that's interesting.
• 25:21
  Like, that's an interesting trigger.
• 25:22
  Let me talk to them about this.
• 25:23
  And then like, hey,
• 25:24
  that might be something, right?
• 25:26
  That's just me
• 25:27
  existing, being friends
• 25:29
  with another person.
• 25:30
  I'm not asking them
• 25:31
  so that I can get some like internal info
• 25:34
  so that I can make them ad,
• 25:35
  but it's just like
• 25:36
  I existed in the world.
• 25:37
  But because I'm somebody who's like,
• 25:40
  that's interesting.
• 25:40
  Like, I think humans are interesting
• 25:42
  and I love how what's motivating them.
• 25:44
  And like that motivated
• 25:46
  that guy
• 25:47
  to go get blood work
• 25:48
  and go get testing and stuff.
• 25:50
  That is actually something
• 25:51
  that we need to figure out, like,
• 25:53
  what is that motivation?
• 25:54
  Maybe there's multiple reasons, right?
• 25:55
  And I'm starting to ask that question.
• 25:57
  If I was behind a computer
• 25:59
  and not out to dinner with my friend,
• 26:01
  I would have missed that.
• 26:02
  So like and that that idea
• 26:06
  probably is going to unlock
• 26:08
  like a lot of money from my client.
• 26:09
  And so you think about like,
• 26:11
  oh, well, is that work?
• 26:12
  Is that it's like,
• 26:14
  I guess, you know, like,
• 26:16
  was I sitting behind
• 26:17
  making the best brief ever? No.
• 26:19
  But the best briefs come from like,
• 26:21
  great ideas and connecting it
• 26:22
  and understanding people.
• 26:23
  And you can't understand people
• 26:25
  if you're not with people like that.
• 26:27
  I think is just like the heart,
• 26:28
  like the funniest thing.
• 26:29
  It's like if you're grinding all day,
• 26:31
  it's like,
• 26:31
  no, you should probably
• 26:32
  watch some Love Island
• 26:33
  because most people
• 26:34
  are watching Love Island.
• 26:35
  And if you
• 26:36
  watch Love Island, it's like crazy, like
• 26:40
  you're watching people in real time
• 26:42
  be motivated by actions
• 26:43
  and completely complex ways.
• 26:45
  It's like it's insane, right?
• 26:47
  So I say all that because like,
• 26:50
  that is where you can really get
• 26:53
  great ideas
• 26:54
  and thinking about things is
• 26:55
  like existing.
• 26:56
  So that's how you can build
• 26:57
  your taste, your palate
• 26:58
  and like all that stuff.
• 26:59
  Now going back
• 27:00
  to like the context documents though,
• 27:02
  which is like
• 27:04
  we have different
• 27:05
  you have the brand context document.
• 27:07
  That is true. Right?
• 27:09
  So you want to say I,
• 27:11
  I know that
• 27:12
  like whatever,
• 27:12
  I want to talk about these things
• 27:14
  for this brand,
• 27:15
  maybe the theme, maybe whatever.
• 27:17
  And then I want to put it in this format,
• 27:19
  72 questions,
• 27:20
  because I saw this from Vogue.
• 27:22
  And then
• 27:25
  then you've got the different
• 27:27
  types of kind of writing documents.
• 27:31
  We have one for statics,
• 27:32
  we have one for UGC briefs,
• 27:34
  and then we have one for studio videos.
• 27:37
  And they all have different elements
• 27:38
  because those we shoot
• 27:40
  those kind of differently.
• 27:41
  The hardest thing
• 27:43
  in the most important thing,
• 27:44
  I think
• 27:45
  to note
• 27:46
  whenever it comes
• 27:46
  to how you use AI with creativity
• 27:51
  and specifically with creative work,
• 27:53
  is that you are giving it rules, right?
• 27:56
  The context document says
• 27:58
  this is how we want to write
• 27:59
  static briefs, right?
• 28:00
  A great creative
• 28:01
  strategist
• 28:02
  is someone who knows
• 28:04
  when to break those rules.
• 28:05
  And so 72 questions probably would break
• 28:09
  those rules
• 28:10
  of what a studio vision
• 28:12
  video typically is for us.
• 28:14
  And so it's probably not ideal
• 28:17
  to use that context document
• 28:19
  because we're really doing something
• 28:21
  completely different. Right.
• 28:22
  It's not as much direct response.
• 28:24
  We're going to be,
• 28:26
  you know, asking questions or whatever.
• 28:27
  So like knowing
• 28:29
  when to break rules is important.
• 28:31
  And I think it's
• 28:31
  the hardest thing about AI is because
• 28:34
  llms are rule.
• 28:36
  It's rule based, right?
• 28:37
  It's predicting the rules
• 28:38
  and what you think the word should be.
• 28:41
  And so
• 28:42
  it's not very good at like breaking out
• 28:45
  of those rules.
• 28:46
  And I think that's
• 28:47
  where the human element is, is like
• 28:49
  if you have
• 28:50
  as much human parts of your ideation,
• 28:53
  you know, like,
• 28:55
  like how to get something
• 28:56
  really interesting
• 28:57
  that taste, if you will,
• 28:58
  and then you should have
• 29:01
  individual types of things,
• 29:03
  like I like to have
• 29:04
  my static adds a certain way.
• 29:06
  We have certain rules
• 29:07
  of how we want to write things,
• 29:09
  design things.
• 29:10
  Same with UGC, same with studio.
• 29:13
  But I think over time,
• 29:14
  like we might flesh that out even more.
• 29:16
  We might have like a, you know,
• 29:19
  maybe we build out
• 29:20
  like a specific style of video.
• 29:22
  It has its own context document.
• 29:24
  That way you're like,
• 29:25
  hey, write it this way or whatever,
• 29:28
  but for right now, that's what we have.
• 29:30
  But I think always know
• 29:31
  when to break those rules is,
• 29:32
  I think, important.
• 29:33
  So I think if you have that human element
• 29:35
  in the ideation phase
• 29:36
  that actually will prompt you to be like,
• 29:38
  no, we're breaking the rules.
• 29:40
  Like we need to do something
• 29:41
  wild or very different.
• 29:42
  And I think the best content is when
• 29:45
  people know when to break those rules.
• 29:47
  And they do do that. Right.
• 29:48
  So the best examples that we love
• 29:50
  in advertising tend to break the rules.
• 29:52
  And yet we try to create
• 29:54
  very rules based content.
• 29:56
  So I think just being careful of of that
• 29:58
  you need both.
• 29:59
  But if you're like, wow,
• 30:01
  I really wanted a very different idea.
• 30:03
  It's like
• 30:04
  it's probably not going to come
• 30:05
  from your context document output.
• 30:08
  So got it. That makes sense.
• 30:10
  And I
• 30:11
  don't want to belabor this point to much.
• 30:12
  I don't want to move on
• 30:13
  because I know you're building
• 30:14
  something that I would love to see,
• 30:17
  but it feels like
• 30:18
  maybe there's like a streamlined approach
• 30:20
  if you're doing something sort of,
• 30:22
  you know,
• 30:22
  let's just say you're coloring
• 30:23
  within the lines,
• 30:24
  you know,
• 30:24
  you've done this a million times.
• 30:25
  You have this context document.
• 30:28
  The beginning part of your process is ID
• 30:30
  come up with the idea
• 30:30
  on my own,
• 30:31
  cross-reference the context document
• 30:33
  to make sure
• 30:34
  the brief looks okay,
• 30:35
  write the brief and refine it,
• 30:38
  and then come back at the back end
• 30:39
  and have the kind of like document
• 30:41
  review it.
• 30:41
  And sure,
• 30:42
  like all my stats are correct,
• 30:43
  maybe beef it up a bit
• 30:44
  and then put it into production.
• 30:46
  If you're thinking sort of like,
• 30:48
  you know, coloring outside the lines,
• 30:49
  thinking outside the box,
• 30:50
  you know, insert your metaphor here.
• 30:52
  You are saying, okay,
• 30:54
  maybe I actually don't
• 30:55
  want to use any context.
• 30:56
  This is going to be
• 30:57
  entirely human process
• 30:58
  where AI actually doesn't even
• 31:00
  necessarily make sense to be involved.
• 31:01
  Yeah.
• 31:02
  And then you could
• 31:03
  after you build it in a more human,
• 31:05
  you could run it through
• 31:07
  that context
• 31:08
  document of the brand and say, like,
• 31:10
  are we missing anything or whatever?
• 31:12
  Like, like,
• 31:13
  is there other hesitations
• 31:14
  we need to hit on?
• 31:15
  So like, you don't always have to use it
• 31:17
  AI to help you create or like right.
• 31:21
  It can help you edit as well.
• 31:22
  So I think those are
• 31:23
  those are different ways to use it.
• 31:26
  But I think, you know,
• 31:27
  you definitely want to have,
• 31:29
  you know, space and time to create.
• 31:32
  Yeah, that makes complete sense.
• 31:33
  And honestly having that as a fact
• 31:35
  checker, that's pretty brilliant.
• 31:36
  I feel like we should
• 31:37
  probably be doing that more,
• 31:39
  just like
• 31:39
  checking stats
• 31:40
  and just like
• 31:41
  maybe this stat is better than that set.
• 31:43
  Okay, cool.
• 31:44
  So you've been sort of talking
• 31:46
  about maybe like built a
• 31:49
  AI tool sort of similar to motion
• 31:51
  and maybe
• 31:52
  like it's a little bit more personalized
• 31:54
  to your, to your studio.
• 31:55
  Maybe you can share that
• 31:56
  and maybe talk through it.
• 31:58
  So I think that we're in an interesting
• 32:01
  time within,
• 32:02
  I think, SaaS and I think that,
• 32:06
  you know, SaaS is
• 32:07
  they are a business, right?
• 32:09
  They have their own business.
• 32:11
  I find that
• 32:13
  within kind of software itself,
• 32:16
  they have to kind of
• 32:18
  build for the middle.
• 32:20
  So
• 32:22
  there's always times where you're like,
• 32:24
  if only I had this one feature.
• 32:25
  Right?
• 32:27
  And then they ask people
• 32:28
  for features and stuff,
• 32:29
  and they start building those features
• 32:31
  so that you feel like complete.
• 32:32
  But it's like, man,
• 32:34
  sometimes you get into a program
• 32:35
  and you're like, this is so great.
• 32:37
  I love this program.
• 32:38
  This is awesome.
• 32:39
  Maybe it doesn't have everything,
• 32:40
  but we make it work, right?
• 32:41
  It's like out of the box. Good.
• 32:42
  And then like two years later,
• 32:44
  you're like, yo, what is this?
• 32:46
  It gets.
• 32:47
  The expansion of product is just insane.
• 32:50
  You're like, I don't need all of this.
• 32:53
  I didn't ask for this.
• 32:54
  And now you're charging me
• 32:55
  for AI credits.
• 32:57
  And I personally think your stuff sucks
• 33:00
  and you're making me pay
• 33:02
  like, ten x that, like, no way.
• 33:04
  Right now,
• 33:06
  hopefully nobody sponsors
• 33:08
  this this podcast,
• 33:10
  but I've not been as public about this.
• 33:12
  But I will right here.
• 33:13
  So you'll get the hottest of takes.
• 33:14
  I think that if you look
• 33:16
  at the major players of, of analytics,
• 33:20
  their model of charging is egregious,
• 33:24
  if I'm being honest.
• 33:25
  Because if I am a and see who's
• 33:28
  working with a brand
• 33:29
  and spending millions of dollars a month
• 33:32
  not all on my ads,
• 33:33
  but just millions of dollars
• 33:34
  a month on meta,
• 33:35
  which we have multiple brands doing that.
• 33:39
  You are charging me
• 33:40
  thousands of dollars to get some analytic
• 33:43
  like an analytics tool
• 33:44
  when I could just do a Facebook
• 33:45
  ads report. It's like insane, right?
• 33:48
  So like,
• 33:48
  I at a certain point was like
• 33:50
  I was looking at like,
• 33:51
  if I'm looking at motion, I'm like,
• 33:53
  I get a data analytics team
• 33:54
  with how much you guys
• 33:55
  are going to try and charge me.
• 33:56
  That doesn't make sense to do, right?
• 33:58
  Like I might as well just hire
• 34:00
  like a brilliant, like,
• 34:02
  you know,
• 34:02
  whatever mathematician at this point,
• 34:04
  which I, you know.
• 34:05
  So that being said, I was like,
• 34:08
  there is so many things that like
• 34:11
  we could just like simplify
• 34:13
  if we could just have like these 2
• 34:15
  or 3 things like, it's like,
• 34:17
  this is all we really need.
• 34:18
  So I spent the time trying
• 34:22
  to build something of equivalent to it.
• 34:25
  Now, is it the most perfect thing now?
• 34:30
  Does it get the job done?
• 34:32
  Yes.
• 34:33
  Does it save us thousands of dollars
• 34:35
  in software? Absolutely.
• 34:37
  And I think what's cool is that, like,
• 34:40
  there have been features
• 34:42
  that I thought would be helpful
• 34:44
  and then they've become not helpful.
• 34:46
  So we just delete it. Right.
• 34:48
  And I think that's what's really cool.
• 34:51
  And I think potentially one day
• 34:53
  SaaS will be able to have like
• 34:55
  you get the out of the box,
• 34:57
  you know, 80%.
• 34:58
  And then you can code by the code,
• 35:01
  the 20% that you need.
• 35:02
  But as of right now,
• 35:03
  that's not how it works.
• 35:04
  So you kind of have to do it yourself.
• 35:06
  So I'm going to share my screen
• 35:09
  so you can kind of see it.
• 35:10
  But basically some of this will have
• 35:14
  to be blurred just because of
• 35:17
  confidentiality.
• 35:18
  But I will
• 35:20
  share my screen
• 35:21
  so you can kind of see it.
• 35:22
  But this is called Verdana.
• 35:25
  Cool.
• 35:25
  So basically
• 35:29
  what you'll see is
• 35:30
  like you have this overview.
• 35:32
  You're able to see like
• 35:34
  kind of high level last 14 days.
• 35:37
  Like you can pick your time. Right.
• 35:39
  So like it's just the Facebook.
• 35:41
  You're just plugging in the Facebook
• 35:43
  API into this
• 35:45
  and you're just grabbing that data
• 35:47
  in and it's syncing. Right.
• 35:49
  So I can kind of
• 35:50
  see how things are improving versus
• 35:52
  the prior period. Very simple.
• 35:54
  There is a creative.
• 35:57
  So this is like your actual
• 35:58
  like table view.
• 36:00
  So you're able to kind of
• 36:02
  this is our custom report that we do like
• 36:04
  these are this is what we look at right.
• 36:06
  Like nothing too crazy here.
• 36:09
  We actually have we we
• 36:11
  this is a newer client
• 36:12
  so we haven't tagged them yet.
• 36:13
  But we do have tags of like by
• 36:17
  is this a founder ad.
• 36:18
  Is this an actor. Is it no talent?
• 36:21
  The type of hook,
• 36:22
  the style and things like that.
• 36:25
  So there's a lot of
• 36:25
  like tagging that we do,
• 36:27
  but you're able to like
• 36:28
  see the ad and then you can click in
• 36:32
  and then it will load
• 36:33
  because you'll
• 36:34
  actually play the video itself,
• 36:36
  which is pretty cool.
• 36:38
  And sometimes it doesn't fully load
• 36:40
  because of the API,
• 36:41
  but it always has like a
• 36:44
  a preview link.
• 36:46
  But you can see like
• 36:47
  we've got some really cool stuff
• 36:48
  here, right?
• 36:49
  Like you can see the retention curve,
• 36:52
  which is pretty cool.
• 36:54
  Basically everything
• 36:55
  you would need to know about this.
• 36:58
  So like there's some other finer
• 37:01
  details in here that's pretty cool.
• 37:03
  But like I think overall
• 37:05
  like what I love about
• 37:06
  this is the fact that like
• 37:08
  what matters most to
• 37:12
  to a strategist is like,
• 37:14
  can I just quickly
• 37:16
  see what's
• 37:17
  working in the account and what is the
• 37:21
  like, co-create the whole rates,
• 37:22
  like what's actually working.
• 37:24
  And it doesn't need to be that complex,
• 37:27
  right.
• 37:28
  Like I think that like,
• 37:31
  you know,
• 37:31
  people are like
• 37:32
  and then you're going to connect
• 37:33
  all the AI and the LMS
• 37:34
  and like all these things
• 37:35
  and these software companies
• 37:36
  are doing like all of this,
• 37:37
  like analytics and stuff.
• 37:39
  And it's like,
• 37:40
  honestly,
• 37:41
  if you have a good strategist
• 37:42
  and like
• 37:43
  they can just view these things
• 37:45
  like it actually is, it simplifies it.
• 37:48
  And I think oftentimes
• 37:49
  we're trying to
• 37:50
  almost overcomplicate it and like do this
• 37:54
  like hand wavy stuff when it's like
• 37:57
  does this add work yes or no.
• 37:59
  And if it's working
• 38:01
  like what are some elements as to why.
• 38:04
  So yeah,
• 38:05
  I think that people just over complicate
• 38:07
  some things that I think,
• 38:10
  you know, just make it
• 38:11
  way harder to use their application.
• 38:13
  So we are really just focused on like,
• 38:15
  what do we need
• 38:17
  so that we can get back
• 38:18
  to being creative as much as possible?
• 38:22
  Yeah okay.
• 38:23
  That makes sense.
• 38:24
  So tool like this
• 38:25
  it seems like commensurate with motion
• 38:27
  atria etc..
• 38:28
  And you know, I guess one question I have
• 38:30
  and this isn't necessarily related,
• 38:32
  and obviously
• 38:33
  you probably built this
• 38:33
  with cloud code or Codex.
• 38:37
  So this is more of a question
• 38:39
  for creative strategy in general.
• 38:42
  These tools are very helpful
• 38:43
  with iterative learning.
• 38:45
  Going in looking at
• 38:46
  oh, this is at a great thumbs up, right?
• 38:48
  Like maybe we should use this hook
• 38:49
  for a different persona or whatever.
• 38:51
  When you sort of think
• 38:53
  like, you know,
• 38:54
  not to insert myself at all, but like,
• 38:56
  you know,
• 38:56
  I generally like to think of
• 38:58
  iterative content and sort of net
• 39:00
  new content like net new testing.
• 39:01
  So when you're strategists and you said
• 39:04
  if a good strategist comes
• 39:05
  in, I will say caveat
• 39:07
  finding a good strategy
• 39:07
  is really fucking hard. Yeah.
• 39:11
  Yeah.
• 39:11
  It's it's way easier to code stuff.
• 39:14
  It's a fine great talent right now.
• 39:16
  So very very true.
• 39:18
  So this is great iterative learning.
• 39:20
  How do you take an analytics tool
• 39:22
  that you built with AI
• 39:24
  and turn that into testing
• 39:25
  entirely new tests,
• 39:27
  entirely new personas,
• 39:28
  angles, things
• 39:29
  that aren't
• 39:30
  necessarily going to
• 39:30
  be evident in the data?
• 39:34
  Yeah.
• 39:35
  I like made a post to you about this,
• 39:37
  like especially with reviews,
• 39:38
  because a lot of people
• 39:39
  will say like, oh, look at reviews.
• 39:41
  And it's like, well,
• 39:42
  if you're advertising a specific angle,
• 39:44
  you're most likely going
• 39:45
  to attract a certain people.
• 39:46
  Like it's just going to kind of
• 39:48
  like reinforce that for you.
• 39:52
  I like this is where I like
• 39:54
  I genuinely like hate to say this,
• 39:56
  but that that's
• 39:57
  actually where the taste comes in.
• 39:58
  Like that is actually where
• 40:00
  like the person comes in
• 40:02
  because you have to be always open
• 40:06
  to understanding.
• 40:08
  How can you open
• 40:09
  and expand the aperture of the funnel
• 40:12
  right then?
• 40:12
  That comes from new styles, new personas,
• 40:15
  like people you want to talk to.
• 40:17
  So like that's the strategy part.
• 40:20
  Like like yeah. Like that's it. Right.
• 40:22
  I think that when you have that new idea
• 40:27
  where AI can help you
• 40:28
  is maybe start to start to expand it.
• 40:30
  I'll give you an example, which is that
• 40:33
  we're working with this one brand.
• 40:34
  It's a home goods brand,
• 40:35
  but the it's a weighted blanket.
• 40:39
  And so it's like okay great.
• 40:40
  Like way to blanket there
• 40:42
  science behind that.
• 40:43
  There's a health component to it. Great.
• 40:45
  Now like we started to expand.
• 40:49
  Well how do we talk about this.
• 40:52
  Not just saying like hey
• 40:53
  everybody needs great sleep okay.
• 40:55
  That's true.
• 40:56
  But now we're saying,
• 40:59
  okay,
• 41:00
  well what about people
• 41:01
  who are neurodivergent
• 41:03
  and they feel overstimulated
• 41:05
  and the weight actually
• 41:06
  helps them with that?
• 41:07
  Now we're doing the research
• 41:09
  to like figure that out
• 41:11
  if that's true or not.
• 41:12
  And then we're doing that
• 41:13
  and then we're thinking about like, well,
• 41:14
  how can we break this down
• 41:16
  into other health components?
• 41:18
  Somebody on dialysis,
• 41:20
  somebody on blah, blah, blah, blah.
• 41:21
  Right.
• 41:21
  And like it just now,
• 41:22
  you're now you're generating like,
• 41:24
  oh, you could do this, you could do this.
• 41:25
  And now maybe you're using
• 41:26
  AI to like be like,
• 41:28
  are there other studies out there that,
• 41:30
  you know, say weight is helpful, right.
• 41:32
  Like,
• 41:32
  but that initial prompt of asking
• 41:35
  AI that came from a human
• 41:38
  being thoughtful, right.
• 41:39
  And being like,
• 41:40
  how can we do something different, right?
• 41:43
  And so I think that like that is where
• 41:48
  the human aspect is.
• 41:50
  So you can expand it.
• 41:51
  Now where the tool can come in
• 41:53
  is is like,
• 41:53
  let's say you have an idea, you now go
• 41:56
  and make a few statics,
• 41:58
  see how that works.
• 41:59
  Are you getting,
• 42:01
  you know, good click through.
• 42:02
  Is that expanding that like
• 42:04
  are you seeing good early signs
• 42:06
  so that you feel like I can de-risk going
• 42:08
  and making a big video about whatever,
• 42:12
  you know, an idea that you had, like
• 42:14
  it's riskier
• 42:15
  when you're expanding off stuff
• 42:16
  that you don't know. Right.
• 42:17
  So like that's
• 42:18
  where an analytics tool can come in
• 42:20
  to say like, hey, I had this early test.
• 42:22
  This is what I was hoping for.
• 42:24
  This is my hypothesis.
• 42:25
  And now we're getting some early signs.
• 42:27
  Let's double down on it
• 42:28
  and let's continue on. Right.
• 42:30
  And so
• 42:30
  I think that's where those two things can
• 42:32
  kind of
• 42:32
  play in is like data can help refine or,
• 42:36
  you know, say yes or no to an idea.
• 42:39
  But like that idea,
• 42:40
  you want to get something
• 42:41
  that new or different.
• 42:43
  That's where
• 42:44
  like you need
• 42:45
  to have
• 42:46
  somebody who they
• 42:47
  themselves are generative, right?
• 42:49
  Like,
• 42:50
  you can't hope that the
• 42:52
  AI is going to be generative,
• 42:53
  if that makes sense. Yeah.
• 42:54
  I mean,
• 42:55
  I think you touched on a point
• 42:56
  that people would just like, don't
• 42:57
  talk about
• 42:58
  enough is like,
• 42:59
  you know,
• 42:59
  AI is built on
• 43:00
  historical human knowledge.
• 43:02
  Like,
• 43:02
  if we're not sort of thinking forward,
• 43:05
  you know,
• 43:05
  these things are going to be
• 43:06
  perpetually retroactive
• 43:07
  and we're never going
• 43:08
  to have any new ideas,
• 43:09
  and we're just going to be in this
• 43:10
  like negative feedback loop
• 43:11
  where every outage is,
• 43:13
  you know, the same. Right?
• 43:14
  And so like,
• 43:15
  I think I think people and listen,
• 43:17
  I was in an AI psychosis
• 43:19
  state for a very long time
• 43:21
  somewhat come out of it.
• 43:23
  Welcome, welcome.
• 43:25
  Yeah. It's good. It's good to be back.
• 43:27
  It's good to be back.
• 43:27
  But listen
• 43:28
  there's obviously
• 43:29
  we've talked on many podcasts
• 43:31
  like there's plenty of ROI here.
• 43:32
  I don't think AI is going to be this
• 43:34
  like fundamental foundational tool
• 43:38
  that everyone it's
• 43:39
  going to replace all these workers.
• 43:40
  I genuinely don't think that anymore.
• 43:41
  I think
• 43:42
  if I had a better margin account,
• 43:45
  I would be shorting space
• 43:47
  times over.
• 43:48
  But I guess what I'm
• 43:49
  trying to say here is that,
• 43:51
  like AI is very good
• 43:52
  at retroactively understanding
• 43:54
  what has worked in your account.
• 43:55
  Like Matt
• 43:55
  just showed,
• 43:56
  like he built his own motion effectively.
• 44:00
  And motion is a great tool
• 44:01
  that you can use out of the box as well,
• 44:03
  but it's not going to be the tool
• 44:05
  for you to come up with new ideas.
• 44:06
  And I think that is
• 44:08
  where a lot of folks get into trouble,
• 44:10
  because they are relying
• 44:11
  too heavily on a tool
• 44:12
  that is looking at history
• 44:14
  rather than sort of looking forward.
• 44:16
  Yeah, if I could jump in
• 44:17
  like I think like
• 44:18
  it's like
• 44:19
  you can vibe code a dashboard,
• 44:21
  but why do you need it? Right.
• 44:23
  And like, it's so fun to be like,
• 44:25
  I made this, I mean, this, but it's like,
• 44:26
  no, no, no, but like,
• 44:27
  how are you actually using this thing?
• 44:29
  Right.
• 44:29
  And that's why I was like,
• 44:30
  there were things that I thought like,
• 44:32
  oh, this would be helpful.
• 44:33
  And then I'm like, this is not helpful.
• 44:35
  Like, I'm not even using this like,
• 44:36
  get rid of this thing. Right.
• 44:37
  So I think that like,
• 44:39
  that's why
• 44:40
  I have issue
• 44:41
  with a lot of like sass
• 44:42
  because it's like
• 44:42
  they are trying to like feature
• 44:44
  Max like the heck out of it to justify,
• 44:47
  like we're bringing you
• 44:48
  value or whatever.
• 44:49
  But it's like most of the time
• 44:51
  you don't really need a lot of that.
• 44:53
  I think that, like,
• 44:54
  what's great about AI
• 44:55
  is that
• 44:55
  there are some really like core
• 44:58
  things that it can do well,
• 44:59
  like whether it's it can take an idea
• 45:03
  and then like iterate off of that idea
• 45:06
  and like do it more.
• 45:07
  So if you're like,
• 45:08
  I need ten or like this like great.
• 45:10
  And like that
• 45:12
  honestly is a lot of work
• 45:14
  that's like not even fun anyway.
• 45:16
  Like for the strategist. Right.
• 45:17
  And so it's like, that's great.
• 45:20
  Let AI duplicate a lot of stuff
• 45:24
  so that you again,
• 45:25
  have more time to be like,
• 45:26
  what's the new thing
• 45:27
  that I can like generate?
• 45:29
  And I think that's really cool.
• 45:30
  One other thing to think about
• 45:33
  in terms of the
• 45:35
  like coming up with new ideas
• 45:36
  and how you do that also in this tool
• 45:38
  and this can be completely shared.
• 45:41
  We have this kind of like
• 45:44
  we can save ads basically.
• 45:47
  So these are some ads
• 45:48
  that like we've saved.
• 45:50
  And what's cool about it is like
• 45:53
  we
• 45:55
  have like
• 45:56
  it generate the script of the video.
• 46:00
  I started getting this bread delivered.
• 46:01
  You know you can download that video.
• 46:04
  So basically what is foreplay.
• 46:06
  What is trea.
• 46:08
  And then we have like
• 46:11
  like what the verbal hook is.
• 46:13
  The on screen hook is
• 46:14
  and then like what the visual hook is.
• 46:16
  And you can always change this
• 46:17
  but like very cool
• 46:19
  because like you can also
• 46:20
  just like use this as,
• 46:22
  like a way of like, man,
• 46:23
  I really like the framework of this video
• 46:26
  and I want to use it for different,
• 46:28
  like whatever.
• 46:30
  But like sometimes
• 46:31
  you see good frameworks and you're like,
• 46:33
  yeah, I want to make that.
• 46:34
  And it like makes it very easy.
• 46:37
  There's a bit of like an analysis here.
• 46:39
  I don't know some of the stuff.
• 46:40
  I'm like, yeah, I might as well just like
• 46:42
  take it out.
• 46:42
  I don't know if we use it completely,
• 46:43
  but I think it's really interesting.
• 46:46
  And then,
• 46:46
  you know,
• 46:47
  we have a board you can like
• 46:48
  make the board for a client.
• 46:50
  So you have that.
• 46:52
  So these are like interesting
• 46:54
  like elements that I think are like good.
• 46:57
  And like
• 46:57
  we have this thing where you can like
• 46:58
  star the verbal hooks
• 47:00
  or like any of these hooks.
• 47:01
  And so we have like a hooks library
• 47:03
  if you will.
• 47:04
  This is like on the newer side.
• 47:05
  So like still building it out.
• 47:06
  But like if I was a strategist
• 47:08
  and we had like 100 hooks, it'd be like,
• 47:11
  well this is a great hook.
• 47:13
  Like I want to use this,
• 47:14
  but instead of it being like,
• 47:16
  if you're a cocktail by the pool
• 47:17
  type of person, you need this.
• 47:18
  It'd be like, if I'm a, you know, ten
• 47:21
  step skincare person, like,
• 47:24
  you need this or whatever, right?
• 47:25
  Like you can start to like,
• 47:27
  think about how that, like curation
• 47:29
  can help you
• 47:30
  with coming up with ideas
• 47:31
  and kind of, again, blending things.
• 47:34
  And so we think about that a lot as well.
• 47:36
  And again,
• 47:37
  super simple tool
• 47:39
  to kind of like build out and again,
• 47:42
  like it's like there's purpose behind it.
• 47:44
  The goal is how does this help
• 47:46
  the strategist
• 47:47
  come up with ideas
• 47:48
  and do their job better.
• 47:50
  And I think that
• 47:50
  that's like that's important.
• 47:52
  The the thing I will say is that,
• 47:54
  like with anything
• 47:55
  that you're building, like
• 47:56
  making sure you get team by in
• 47:58
  and getting them educated is important.
• 48:00
  I find that a lot of leaders are like,
• 48:02
  this is so awesome.
• 48:03
  And then like,
• 48:04
  it's not being used and whatnot.
• 48:06
  And so I think, like
• 48:07
  just making sure
• 48:08
  that you understand
• 48:09
  that the adaptation of your tools
• 48:12
  and what you're doing
• 48:13
  is actually takes a lot longer.
• 48:14
  But we find that it's been very helpful,
• 48:17
  like traditionally speaking,
• 48:18
  because of just how like Facebook
• 48:21
  and all that stuff was working.
• 48:22
  You know, our editors
• 48:24
  maybe not
• 48:25
  didn't always have like a view of
• 48:26
  like what's happening in the accounts.
• 48:28
  And now they can just log in
• 48:30
  and see everything.
• 48:31
  And so there's been a lot of like value
• 48:33
  add from the tool itself.
• 48:35
  But like also know that
• 48:36
  like you've got to do
• 48:37
  trainings,
• 48:38
  you've got to walk through people
• 48:39
  and you've got to
• 48:39
  encourage them to use it
• 48:41
  because they may not use it if you don't.
• 48:43
  Yeah, I think it's a really good point.
• 48:44
  The buy in,
• 48:45
  something I'm even guilty of, it's very,
• 48:47
  very hard.
• 48:47
  It's very, very hard to get your team
• 48:49
  fully bought into this stuff.
• 48:50
  Okay.
• 48:51
  We do have to wrap,
• 48:52
  but if I were to sort of encapsulate
• 48:54
  everything here, it's just like,
• 48:55
  you know, iteration
• 48:56
  and building
• 48:58
  upon existing winners and concepts.
• 49:00
  Use AI, right?
• 49:01
  Like you have an incredible tool bank.
• 49:04
  You have all this data.
• 49:05
  Matthew showed
• 49:06
  a really, really powerful tool
• 49:07
  they've built internally
• 49:08
  that sort of dissects
• 49:10
  ads that they pull in for meta.
• 49:11
  They utilize that
• 49:12
  and as their foundation to build
• 49:15
  maybe a hypothesis
• 49:17
  that they came up with on their own,
• 49:18
  but then let the
• 49:20
  AI really,
• 49:20
  really help you with iteration,
• 49:21
  but free up your own time
• 49:23
  to do your own ideation
• 49:25
  and come up with sort of.
• 49:26
  Net new concepts. Is that Sarah?
• 49:28
  That's exactly it. Cool.
• 49:30
  All right.
• 49:30
  Any any sort of last questions.
• 49:32
  Are you reading anything interesting
• 49:34
  any any sort of parting words
• 49:36
  you like to share?
• 49:37
  I think similar to you.
• 49:40
  There's been times where I've gone
• 49:41
  maybe too deep with the AI,
• 49:43
  and I think that, like,
• 49:45
  there's there's a balance
• 49:46
  in the process of creativity.
• 49:48
  And so I think find your final works for
• 49:51
  you and that's different for each person.
• 49:54
  Spend the time to like build out
• 49:56
  the context documents to like do that.
• 49:58
  Know that you're going
• 49:59
  to have to refine it over time,
• 50:00
  but spend the time to do it
• 50:01
  because you will get
• 50:02
  you will have gains from that,
• 50:03
  but also know
• 50:05
  when you need to write down,
• 50:06
  you know,
• 50:07
  pen and paper,
• 50:08
  some ideas
• 50:08
  and like come up with headlines
• 50:10
  and like push yourself for that.
• 50:11
  So you know I think try to flex between
• 50:15
  varying different,
• 50:16
  you know, ways in which to create.
• 50:18
  I think a great creative
• 50:19
  person knows how to do that.
• 50:21
  And so I think like, you know,
• 50:23
  find what works for you
• 50:24
  to get you the best ideas.
• 50:25
  At the end of the day,
• 50:27
  if you're in advertising,
• 50:28
  the goal is to make ads that win that.
• 50:31
  That's it. Right?
• 50:32
  And you may not need 100
• 50:33
  ads, you may just need a few 1020.
• 50:36
  And most people are spending
• 50:38
  not enough money
• 50:39
  to need hundreds of generated ads anyway.
• 50:41
  So you probably should
• 50:43
  focus on
• 50:44
  less is more,
• 50:45
  and most likely that's going to be you.
• 50:47
  So I think just thinking about that
• 50:49
  and that can maybe ease your process
• 50:51
  and kind of build it in a thoughtful way
• 50:53
  for how those two things come.
• 50:55
  AI and your own creativity
• 50:57
  kind of come into harmony
• 50:59
  with one another.
• 51:00
  You need both of them at this point,
• 51:02
  but I think they only work
• 51:05
  if you have that harmony
• 51:07
  and you can blend it together.
• 51:08
  And that's going to be different
• 51:09
  for each person.
• 51:10
  So, you know,
• 51:11
  you got to find what works for you.
• 51:12
  But I think spend the time
• 51:14
  to kind of find
• 51:14
  that harmony is going to be
• 51:16
  where you can be the most effective.
• 51:18
  Nice spoken from the man himself.
• 51:22
  Appreciate you coming on
• 51:23
  and we'll talk soon.
• 51:25
  Thank you.