#5: How AI is Helping (and Hurting) Marketers
Welcome back to another episode of The Junction.
We are jumping right into it, talking about
one of my favorite things in the world.
Marketing. Yep.
One of Chase's favorite things in the world, AI.
Automation.
Well, I love marketing, too.
I know, but if it's kind
of like the perfect blend, right?
So you're definitely on the
bleeding edge of this stuff.
And I've been meaning to ask you, do you
have some sort of bot that's like surfacing the
latest, greatest AI content that's dropping right into your
Slack channel or discord or something?
How do you stay on top of this stuff?
It is moving so fast.
It actually died on June 30.
Reddit had a big API FUPA and I don't know,
we don't have any poll on this podcast, but if
we did, I would tell them to revert what they
did so they could bring Apollo back.
But it goes to a lot of the communities
where people are enthusiastic about what's going on.
And not that the community died, but the vehicle
through which that was happening, the wheels fell off.
But I say all that it is moving so quickly
that it is extremely difficult to keep up with.
If you're involved in GitHub, you can have
a community project where people are actively coding
with you on behalf of maybe your own
project and y'all are pushing the barriers forward.
Well, one of these projects, Autogpt, has had thousands of
pull requests where people are coding and helping out to
push the boundary and you can't even keep up.
The founder, if you will, or the guy that
started it is struggling to build out a team
of volunteers that can keep up with this stuff.
And what's really interesting is that it's moving so
fast that the news can't keep up with it. Right?
And the only way to keep up with it is to have your
hands in and on it and doing anything with it at all.
Because if you don't, you're not going to be
able to keep track of all the news.
In the beginning, you know, this when this stuff first
started coming out, I don't know, a few months ago
in the last year, I was very apprehensive.
There was an enthusiastic defensive,
you might say defensive.
I was apprehensive.
Okay, I'm okay with change. Right.
I feel like I am an adapter adopter of
new strategies, especially within my domain of marketing.
But the way that it was positioned to me in my
feed on LinkedIn or on Slack was almost scary, right?
Like I'm like I kind of went to the is my job okay?
Is my job know, are they basically going to
outsource marketing to a bot that's going to go
build this website and post on social?
So it was the initial fear of not understanding and
then to your point, actually using it and then becoming
open minded to, well, how are other marketers using it?
And there was very few talking about
it in a positive connotation too.
Initially it was, oh, the copywriting is know all you
marketers that are going to go write blogs, all of
your content's going to get buried in Google SERP.
And that's where my head was at.
And so I'm thinking, yeah,
you can't write authentic content.
There's no way that you can possibly
go do this with this tool.
And then once you started using it, especially once
I started to learn more about having your own
tool set, because that was the other thing was
anything I put in here, it's going to remember.
And I wanted to be careful.
You know me, I'm very compliant.
Yeah, but that's where my head went.
And then once I started actually getting my
hands on the keyboard and using transcripts and
asking it questions and even then I was
a little apprehensive like, where's this all going?
But then I started seeing the insights
and I was like, this is silly.
Yeah, if we're not using, there's no way
we can't not use these tools in marketing.
Well, you as a marketer should
know that bad news sells, right?
And good news doesn't.
And I'm going to plug John
Krasinski here for a hot take.
He came out with this show in the
middle of COVID called Some Good News.
And the whole thing was about very familiar.
Yes, good News, great dude, love all
of his shows, he does great work.
But the media loves to sell this idea
of like, hey, this hot new thing came
out and it's going to destroy your world.
Well, that's going to get a click.
Whereas, hey, auto GPT came out and it's
going to make your life so much easier.
Naturally gets less clicks.
And I don't necessarily despise
the media because of that.
I know they're trying to report and
do as best as they can.
I mean, nobody walks into the room and
says, hey, how do I report on this? Poorly.
But ultimately I can understand why a lot of people would
start to wonder about their job, even as maybe I'm not
the best developer in the room, but some of our developers,
I think you're the only developer in this room.
Okay, that might be true, but some of us are
wondering like, well, is this thing going to outcode me?
Right?
Can it code better than I can?
And it actually goes back to one of the episodes
that we had a few back where we were talking
about original content and maybe in the past, at least
if you're in the SEO world, they called it content
is king, marketing maybe a little bit content is king.
And I think what we're going to naturally
progress into is that original content is king.
And that's one of the things that large language models
are not going to be capable of for the time
being, right, is coming up with these brand new ideas
that gain traction that people get excited about.
And I won't say they'll never be able to do
it, they're not be able to do it yet because
they're naturally predicting stuff based off of knowledge that they
have pulled out of the existing internet.
If you looked at GPT 3.5 or 4.0 right.
A lot of that's based on
everything that's 2021 and prior. Right.
Anything beyond that, they really
don't have a good understanding.
So anyway, content is king is the old way
and original content is king is the new way.
Well, and I think when you start to consider
the use cases in a safe or compliant environment,
right, because initially even we did Ven technology, we
came out with a no use policy.
Not because we didn't want to embrace it, we just
wanted to use it responsibly on behalf of our clients
and our partners and also to protect our team.
Absolutely.
So we've since opened up the conversation internally
and we're having conversations like this with our
teams and obviously we've run things through legal.
You should do that if you are considering
deploying an AI strategy, seek legal counsel to
kind of give them some use cases.
Just kind of a side note there.
But I've started to look more at this tool
as that, I don't know, internal Google, right?
Like it's the things that you can
go ask it in your own database.
So couple of very, I think low hanging fruit
in terms of use cases for marketers today.
Expedite your can I keep going back to call
transcripts, but that's a really easy use case.
A lot of us are using Zoom.
I don't actually know if Microsoft teams has some
sort of transcription service or part of it.
I know when we even drop videos in
Slack, slack has an attempt at transcribing it.
But even then these tools are so cheap if you need
to go pump it into a tool to give you a
transcript and then you can go upload it into OpenAI.
Obviously it's a lot easier in a tool like OpenAI because
I ran into this with using Chat GPT where I was
having to upload it in chunks and that was frustrating.
Oh yeah.
I'm actually still running into a little bit of the
tokens limit on some of mine just given the call.
But if you can go just take historical
transcripts and start putting those into a tool
like one of these AI tools because there
are a lot out there outside of OpenAI.
I think Jasper is another one I've heard a lot
of marketers using, I haven't used it myself, but you
can collect insights so much more quickly than you would
in a more traditional calming through and listening to videos.
But other ways that marketers go collect
insights from their clients, for example, or
the market is conducting surveys, right?
Yeah, help it, it can go write it for you if
you still want to deploy a survey, but I would argue
that the data is already living in your CRM.
I always want to know what calls did prospects ask?
Like, I'm asking the sales team what calls?
Yeah, I don't know.
I'm onto the next one. Right? Yeah.
It's not because they don't remember.
It's just like they got to move on to the
next they got to move on to the next thing.
I've tried asking at the end of the week. Yeah.
What was the top question you got from a prospect?
If you do that once a week, every week
for 52 weeks, you got 52 prospect questions and
then you're going to start to have trends.
But what if I can just go compile that information?
Oh, naturally. Yeah.
So those are just like that's one thing is
just the research aspect of it, but then uncovering
new products or verticals that you'd be going into
just based on it, looking at the transcripts and
the things that your prospects and partners are saying.
Yeah, okay.
I got to switch it up though.
I'm probably the most optimistic
person in the building.
I love to think about the best case scenario.
I want you to put on your pessimistic hat,
right, and tell me, how is this bad?
How is this hurting marketers?
Here's a quick example in my mind, I'm thinking like, well,
I could ask this thing to write a blog, right?
Write a blog about how salesforce is great for
my company and it spits something out and I
put it on HubSpot, I publish it and boom,
hey, check me out, I just wrote a blog.
Tell me how that is bad for a company.
Well, it's a slippery slope because especially if you don't have
AI assisted or you have no editorial oversight, so you certainly
never want to go put a blog out from Chat, GPT
or any other tool that you didn't scrub or put your
own brand, tone, voice, all those things.
But you also have to consider if
you can do it that fast.
And let's say you said, wow, that one
blog took me ten minutes or an hour.
Even if it took you 2 hours, for us, it could
take 8 hours to actually go and do the research and
write a blog and talk to the subject matter experts.
Let's say you go generate ten more
and you just start publishing them out.
Google's going to see that, wow, they
usually post one blog a week.
Every Monday, they just posted ten. Right.
That's the worst thing you could do.
That's a simple example.
Like you're going to get downranked and
when people go search on those terms,
you're not going to get surfaced. Right.
That's the first thing.
I just think because you can do it fast
doesn't mean you should then deploy it fast.
So if you can go generate content quickly, then
you just bought yourself more time to package it
and prepare it for deploying out to your audience.
So that's just like my first word of caution.
I think that's a really
dangerous, really slippery slope. Yeah.
The blog is an easy one.
And then in terms of volume, just because you can do
it fast doesn't mean that you should now be posting.
Well, we can go generate ten blogs a week.
That doesn't mean you should go
publish ten blogs a week. Right.
I think any marketing department, especially if you a
have content focus, you have content pillars and things
that you obviously want to talk about.
And then you have the we want to entertain, we
want to inspire, we want to educate type thing.
And if you just start willy nilly throwing content
out there, even if it's all within a very
specific set of keywords, where is the gut check?
Like, you're bringing in leads now based on
what you think is the keyword that's going
to bring them into the site.
And now they're asking questions of your sales team
that they read something in some article, some blog.
How's your sales rep going to even hot dang. Yeah.
How are they going to know how to respond to that?
Right.
I think in general, just as much as
marketing's job is to educate the prospect, we're
always doing the same thing internally. Right.
And it's not necessarily education, but more awareness of
here's how we're positioning our company out in the
market and our service and our products.
And so you're making sure
that you're in alignment internally.
Would you say that I'm digging myself a hole?
If, as a business owner, I'm going and writing 20
blogs on something I think I'm and maybe I am
a true expert in, but I'm letting really, without a
ton of oversight, letting GBT write the blogs for me.
Like, am I digging myself a hole?
Or is that something that I could dig out of?
Or by the time I realize it, it's going to be too late.
Are you starting fresh?
Yeah, absolutely.
I'm trying to split up a new business.
I'm trying to figure out how to do marketing.
I'm doing everything.
I'm wearing all the hats.
I'm going to have GBT write me
20 blogs on my topic of interest.
Am I already screwed or is that something
that I can pull myself out of?
I think again, there is no such thing in
our space, or at least as long as we
will use the tool that we don't have some
sort of editorial like it's human assisted AI.
There has to be a human element.
And if the blog, once the person you've drawn
in your prospect or you've drawn in, you've created
a community, an audience around particular pieces of content.
And if once they get on the
phone with you, there's a disconnect.
It doesn't sound like you or you don't
know the things that you're putting out in
your blogs or your content, then you've lost.
Like you are digging yourself a hole
and you can't get that sale back.
So I do think it's a slippery slope.
I think it's a great tool for, hey, what
are things top of mind to your prospect and
trying to understand your persona and the challenges that
they're facing and then what are the keywords?
And then going and writing kind of some outlines based
on that, maybe even if it's you're the business owner,
and you pull up the prompt and you record yourself
talking about it for ten minutes, put that in a
chat GPT and then have it write you a blog. Boom.
That's something we need to do.
That little thing that you said we should do where
we carve out and circle, highlight, whatever and put that
on on social where the guy goes and actually does
it while he's recording, because that's like golden.
I mean, that's the biggest thing
any marketer will tell you.
As long as I've been in the space and
I heard someone talking about it last week who's
been doing this a lot longer than me.
If you can just get it's always like, if
I can just get five minutes of your time,
let's walk and talk right in the hallway.
Because you know that that salesforce consultant or that
product developer or that CEO is onto their next
meeting or into their next task and they don't
want to talk to marketing, right?
Or maybe they are legitimately too
busy or they're legitimately too busy.
And so it was always like, I have this visual of
me almost running down the hallway, like writing down, like, just
give me three bullet points and I could build a blog
or a piece of content off of that.
Now, if you can just get it's like,
oh, let's get them on a webinar.
Yeah, let's get them on a podcast.
And now you have 30 minutes or an hour of
them actually talking their vernacular, their tone of voice.
And so we're kind of like moving into that.
If you are an expert in your field
and you're starting a business, just start pull
out your phone and voice memo.
We tell Scott to do that all the time.
And now take that transcript, put it in there, and
tell it to summarize it into a blog format, right?
I mean, let's be real, though.
If you're only giving your marketing team five
minutes on the hallway while you're walking to
your next meeting, maybe you're doing it wrong.
No, I like to think that I've developed that
rapport over time, but there is that, right?
Like marketing teams, sometimes it can be
your own fault or to your own
detriment that you build out these personas.
Maybe at one point it was collaboratively with
sales or with the product teams, if you
kind of sit in a cross functional role.
But at one point in time, you built out a
persona and you know their challenges, like the back of
your hand and their objections and you're going and getting
those questions and you're writing content based off that.
And if you're not bouncing it off of those teams
or you're not mining, like regularly spending time sitting in
on their meetings to understand new product releases or how
sales is selling this new product or service, then yeah,
you can start doing stuff in a vacuum.
Well, I will be the first to say that.
I think it's like there's this saying, like,
when there's a slump in the economy, marketing
is the first to go kind of thing.
For me, as a personal standpoint, marketing will
be the last to go because it is
much harder to sell something without effective marketing.
So if you're one of those guys, like walking down the
hall and you've only got five minutes for your marketing team,
you should give me a ring and I'll tell you why.
Maybe you should flip that around.
Why don't we switch it up?
Talk about what's going on in the news. I love it.
Headlines. Okay.
Chat GPT isn't coming for your
marketing business, but is it?
Are they?
This is probably speaking to agencies, right?
Or someone who does marketing for a living.
Maybe this goes back to, oh, no,
that's going to start writing websites and
blog content and copywriting and designing images.
No one's going to need me or my skills.
Yeah, I think it goes into that what's
it called, where it's like you don't believe
that you are like this person.
There's a word for it, you have to fake
it till you make it kind of thing. Right.
If you don't believe truly in yourself that you
can deliver better work than Chat GPT can, then
naturally you're going to fall into this fallacy.
I have a perfect example for you. It's very timely.
So last summer, about this time a year here
at Venn Technology, we had this incident where someone,
after hours, not an employee, came to our headquarters
to get some sun, laid out a beach towel.
They were a little cold and
they wanted to get hot, right?
They needed some they were trying to get a
little nice little bronze going, get a little crispy.
So this gentleman lays out his towel.
We've got cameras everywhere, right?
Security cameras.
Bradley sees it on the ring cameras and
says, wow, someone is sunbathing outside of Venn.
This spiraled out of mean.
We sent the guy a the.
I'm like, hey, give me that video footage.
All of a sudden now we're spitballing ideas
about, well, this guy's building a tan outside
of an office that builds integrations.
That's our next marketing campaign. And boom.
Do you think Chachi BT could have told me that that
Ven screen would be the next big campaign of 2022?
I could get there in a really roundabout way,
but not from kind of a one shot.
Oh, yeah.
Now I have all this context
that's our next marketing campaign.
I don't know if how much more authentic can you get?
I mean, that truly is at least that works for Venn.
Some of our best ideas are things that happen
here in the office or around us, right?
We've been able to kind of
flex our personality, so to speak.
We're not that just kind of stuffy b to B, right?
Sort of firm.
It goes back into the relationship, like
what's going on in your life.
You don't have to have somebody to come tanning in
your parking lot to build out a marketing campaign.
It could be like, I don't know, the random news
stories where the truck falls on the ground, right, and
cash starts flowing out everywhere and you could just build
out original content based off of this.
It's where you run into trouble.
And with this headline, right, chat,
GPT isn't coming for your business.
And if you feel like it is,
maybe you aren't doing it right.
Maybe you need to rethink your marketing perspective,
your sales perspective, and how you're going about
and working and doing those things. Agreed.
You have another headline for us?
Oh, yeah.
It's in all bold.
It says, Why relying on AI.
I hate to laugh.
Why relying on AI in
advertising is hurting creative innovation.
And I think this particular article is
referring to the stable diffusions, the ones
that are generating images based on prompts.
And from a more creative artsy,
like, let's generate some graphics.
Why that can be problematic.
And to be frank, this is probably one of the
use cases where maybe there is a little bit more
concern from these folks because the artwork that is being
generated is actually kind of really good.
I haven't played with that at all, have you?
Oh, I've played a lot with it
and I've seen what people are doing.
And to be frank, this is an area where
and maybe I'm just like the tech guy and
all I do is read words all day, right?
Like, I'm not an art guy.
I can't tell the Mona Lisa from some fake one.
I have no idea.
And so maybe it's just my inability here
to tell from one thing from another.
But from what seen the use cases, and
Adobe has shown this to be really interesting.
You can take a small picture and I'm paraphrasing here.
You drag out the deal and it will fill in
the content of what it thinks the rest of the
picture would have looked like if it existed.
And it's actually pretty good.
And so my concern isn't that we're going to have,
from a marketing perspective, graphic design kind of die out
and GPT is really or the AI is really going
to take over here, but the bigger concern is going
to be like, well, who generated this?
You get more into the ethical kind of
things like, well, where did it train on?
Did it train on?
Copyrighted works and is that a problem
or should we be worried about that?
Are we going to get into a lawsuit because
I trained on the Mona Lisa and all these
copyrighted things or is Disney going to come after
me because I trained off a Mickey Mouse?
I think there is legitimate concern there and
you have to spend time thinking about it.
And if you are a business and you are
a marketing agency, you should spend more time here
thinking about what you should do because ultimately all
these models are trained off of data.
And from like a blog perspective, you can't really
copyright a blog, but you can copyright an image.
And that's where things get really murky because nobody knows
the data that the model was used to train on.
And then it starts generating things that
kind of look like Mickey Mouse.
But I don't know, maybe if you got into
a courtroom, maybe you'd win, maybe you wouldn't.
But if they figured out that you trained on copyrighted
material, that's going to be a big legal battle.
So is it hurting the advertising industry?
I honestly don't know. I don't know.
Mel, what do you think all these things are true?
I agree with you.
It's something that if you're not already
understand, I haven't played with any of
the more artistic imagery generating type stuff.
I really haven't.
I've been more preoccupied and fascinated at here's,
digging into the CRM and just trying to
understand how we can generate written content.
I will say in an earlier episode you talked about
how the authentic content will rise to the top.
There was a great example the other day I
saw I think I was scrolling through Instagram.
There was a business that hired a nine year old
kid to decorate their restaurant with his doodles and he
was actually getting in trouble at school for doodling.
I mean, you talk about and now
this thing is like blown up.
It's all over the headlines.
Oh yeah, they didn't have an AI bot do that.
Nine year old kid came in and
plastered the walls and his doodles.
It's actually pretty neat.
So things like that.
We've thrown around some ideas here internally.
I'm not going to go too far, but we've thrown around
some ideas that would require, let's just say some illustration and
could we leverage tools to go do it faster? Probably.
But there's also an element of just pure authenticity
that you're never going to see that anywhere else
because you've hired an artist or you're working with
someone internally to go generate that.
So I don't know, I'm kind of on the
fence when it comes to I think it's great
from a we all want to be on brand.
Is that the right color usage and font?
And let's follow the brand guidelines.
Can we feed it that?
Can it fact check us?
Are there tools out there that can we know
we can preload up templates and things like that.
But I don't know.
I don't really know where my take is on kind of the
visual arts of it all because I come from and myself and
Randall, who's on my team, is in the room too.
We both, at one point in time or another,
in our backgrounds, have been painters and illustrators.
And you kind of have this there's something to that.
No one's going to take away my paintbrush. Right.
If I want to go paint at home, I can. Right.
If I want to start selling it,
I can give me that paintbrush.
I don't know anyone's going to buy what I'm painting.
But I do echo your sentiment around understanding.
If you're training on something, are there copyright infringements
and how is it going to be used?
Well, one thing we haven't even talked
about is you think about, from a
business perspective, material that is generated. Right.
And it's got the picture of a face. Right.
Somebody that when I was at Texas A
M working at May's Business School, we actually
had professional or executive MBA students come in
and we would take pictures of them naturally.
Aggies are really good looking, of course, not
pictured on this recording is my eye roll.
Oh, no, I can't see it. That's great school.
It's too small.
But they would take pictures of these folks right.
And then they'd show up in the
marketing material and look great, of course. Right.
But now you can generate photorealistic images
of people, and those models, if you
will, are now no longer needed.
And who was a part of that training set.
Right.
How did the model know what
the picture should look like?
Well, it was trained off of somebody.
Well, that person can copyright their own image. Yeah.
Anyway, there's a lot of murkiness here, and it's
probably the one area of AI that I am
just not the best person to talk to because
I can't speak to it one way or another.
Some of the content that I've seen
generated by AI is really good.
You're like, wow, that looks really awesome.
The only place that I can definitively say that AI will
maybe never get there is when you look at paint.
Like actual paintbrushes. Sure.
You can't necessarily replicate that, and somebody would
probably say, well, actually, I can do that.
But for the most part, right now, AI is
contained within the Internet or some kind of digital
experience for the folks that are painting.
With paint you probably don't have
to worry about for a while. Yeah.
Well, in general, I do think there's
a lot of upside for marketers.
And as is kind of the conversation from
episode to episode, we realize there's risks.
We get stuck in the optimism category
a lot because we're excited about it.
We're exploring what it can do.
But I do think that we need to continue to
understand how we can glean insights faster as marketers.
And get those out to the
market quicker, answer questions faster.
If you roll out an annual survey once a year, by
the time you get those insights out to your too late.
Way too late. Right.
And same goes internally even.
We did a webinar recently about finance and AI.
I absolutely leveraged OpenAI for the recording to
help me build the outline of the blog,
pull out key quotes, key moments.
It would have taken me so much longer and
I would have probably had a few more weeks
to actually get it out to the market.
And we had a lot of excitement around it and momentum.
So figuring out how to use these tools to capitalize
on that and then also to be able to pivot
because your message could be maybe you have to, like
we talked about in the Drip campaign, what's not working.
You know, if you spend six months doing that
and you're not asking those those questions naturally, and
then you have a tool that can at least
help quickly give you some feedback. Yeah, I don't know.
It's pretty powerful.
Well, we're interested in how you are
utilizing the AI models, the language models.
If you are an individual that is utilizing
stable diffusion, tell us how we were wrong.
Tell us how you are using it and why or why not.
This isn't a big deal for you guys.
Feel free to email us at thejunction@ventechnology.com and
we look forward to hearing from you guys.
And if you want to send us
an audio clip, do that as well.
We'll totally play it on a future episode.
Send us your images.
Or send us your images.
Maybe we'll tie like, a blog into this whole thing.
Let's ask Chat GPT if we should do that.
That's a great idea.
We just promoted our chat GPT AI
into our kind of AI CEO.
We could have a co host anyway.
You're my co host. I'm your co host. Right.
But we can there's not a third co host.
Chat GPT.
Yeah, like, they could join us.
No, we could change our segments to
where we start asking and questions.
See, look at us go.
We're iterating on the fly.
I like it.
Keep it automated, y'all.