Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Samantha (00:05):
Hi, welcome to another
episode of 1000 Ways to Market
Your Business, a podcast broughtto you by Pushing the Envelope.
I am Samantha Scott, and today Iam delighted to be joined by two
of my awesome colleagues, Triciaand Paige.
And we're gonna talk about AIand copywriting, dos, don'ts,
tips, tricks, all the things.
So ladies, introduce yourselves.
Paige (00:24):
My name is Paige.
I am PTE's Digital OperationsCoordinator.
And within my role I do a lot ofAI research, making sure that I
can fill the gaps for any painpoints that our team has that AI
could replace or streamline.
I stay up to date on the everevolving landscape of forums and
new platforms and softwares, andjust inform our team so that we
(00:46):
can use it to streamline andmake our lives easier.
Tricia (00:50):
And I'm Tricia.
I'm our Engagement Strategist,so I handle a lot of community
relations and then a lot oforganic social.
So I use AI for brainstormingand copywriting when it comes to
content for our clients ortrying to figure out events and
how to tie them in with ourclients for different goals.
So me and Paige collaborate alot between the data and the AI,
to kind of use it to benefitourselves and our clients.
Samantha (01:13):
So let's dive in.
This is a juicy, really deeptopic.
I think we could go on for daysabout it, but let's start by
discussing what's the biggestmisconception as it relates to
AI and using it for business.
I would say, things I've heardquite often, is it's coming for
our jobs, or it doesn't actuallyproduce good content, which I
think has to do with how you'reprompting it.
(01:35):
But those are just some of thethings off the top of my head
that I've heard.
I'd love to hear from yourperspective, you know, Paige,
you live and breathe AI andoptimization and how to use it,
but Tricia, you're actuallyusing it for content creation
specifically.
So tell me your thoughts.
What's the biggestmisconception?
Paige (01:52):
So I think first and
foremost, I do agree people feel
threatened that AI's gonna takeover their job.
But that being said, one ofthose biggest misconceptions is
that AI can completely replacethe human touch and replace
certain processes withinmarketing or business function.
It's a great tool, but behindit, and to your point about how
you're using AI, it needsstrategy, it needs to know
(02:14):
what's working and using data toinform it.
In a lot of ways it can helpwith speed and efficiency and
improving processes, but it'snever going to fully take a good
majority of positions becauseit's just gonna make those
positions do better at theirjob.
Samantha (02:29):
In fact, it's even
created a position.
A lot of companies now arefinally catching up to us and
having an AI ambassador like wehave Paige.
That's her job and a largefunction of it is to actually
analyze it and optimize the useand support the team.
I think it's a really importantcomponent because yes, humans
are still involved andnecessary, but also because we
have to make sure that we'reusing it the right way and that
it's being used consistentlyacross a team.
(02:50):
If you don't have someonespearheading that, you've got
people just going rogue, usingit different ways and not
efficiently.
And that's when you get into thesituation of poor output because
it's poor input.
But Tricia, I'd love to get yourpoint too.
Tricia (03:03):
Yeah, and kinda like
Paige said, it's a lot of, you
gotta have that human touch andI think that's where it kind of
plays a role in my job.
If I were to only use ChatGPT tocreate captions or social media
content, you're gonna look at itand say, oh, ChatGPT wrote that.
And we do a lot of tweaking forbrand voice and things like
that, but there's still justsome element that you have to
revise based on your client'sneeds and their goals way better
(03:25):
than Chat does, no matter howmuch you inform it.
We've even ran into instanceswhere it will make up promos.
If we would just put that outthere, we would have clients
knocking on our door for thatmatter, saying, why are we
offering this?
You know?
So even just in the sense offact checking and looking for
that in content.
But it definitely helps kind ofspeed that process along.
(03:46):
It makes the process so muchmore efficient when you're kind
of writing a month of content ata time, that can take a lot of
time when you just use your ownbrain power versus using Chat to
kind of help brainstorm and getthose ideas really flowing.
It really helps speed thatalong.
Samantha (04:00):
Yeah.
That brings us to the nextpoint, which is idea generation.
How do we best leverage this forthis?
Because if you just go to Chatand say, hey, write me a month's
worth of Facebook content for arestaurant.
That's so vague.
I mean, think about asking ateam member or an intern or an
employee to do that.
They may not come up with stuffand you have to know, well,
okay, well is this a Mexicanrestaurant?
Is it a fine dining restaurant?
(04:20):
Is it takeout?
Is it a food truck?
I mean, there's so manyvariables, so making sure you're
giving it really good data is socritically important.
And one way that we are doingthat is we educate Chat or any
other tool we're using.
We reference ChatGPT often, butthere are many other tools we
use as well, Perplexity forresearch, things like that.
But we teach it, and that's themost important first step that I
(04:42):
think so many people are missingwhen they're using AI to help
with content generation orreally for anything else, even
idea creation.
Because if it has no contextabout the company that you're
trying to create ideas orcontent for, it's not gonna be
able to do its best job.
So the way we take it is weeducate it about each individual
client.
So if you're in-house, tell itabout your business.
(05:03):
Explain, this is who we are.
This is what we do.
This is our target audience.
This is our unique valueproposition or our
differentiator.
This is our brand voice, how wewanna come across, this is our
personality.
Give it all the backgroundinformation and then tell it
what your strategy is.
So what's your marketingstrategy?
Which you should have, annually,but then also if you have a
campaign you're working on.
Give it all of the resources andinformation so it can be
(05:25):
successful.
So you kind of have to help it,help you, and then see what it
comes up with and then furtherrefine it and say, yes, I like
this, or don't like this, orchange this wording.
But then we take it a stepfurther and use the data that
Paige creates every month in herreports.
And then we educate it and say,okay, Chat, we had these posts
go up, these performed reallywell.
These did not.
Create more of these and less ofthose.
(05:46):
So you have to continue toeducate.
I think that's another thingthat people miss out, is if you
educated it in the beginning,but then you just kind of set it
and forget it, it's not gonnawork as well.
But if you continue to tell it,yes, this did well or this did
not, it's gonna create bettercontent over time and also
better ideas because then it'llknow, oh, okay, the audience
really liked this, so I cancreate more of that.
Paige (06:07):
Yeah.
And to your point withcontinuously educating it, I
think it's also important, Ikind of like to think of it as
like talking to it like anotherteam member as like a human and
telling it exactly what youliked in its output, what you
didn't like, how can it beimproved and giving it clear
direction.
The vague like, give me acaption for this isn't gonna be
successful and you're gonnacontinue to get bad outputs,
(06:28):
that's gonna not save you time,'cause you're gonna have to make
more edits.
But continue to tell it what'sworking, what's not, be
specific, and also pairing thatwith what's currently happening
in those forums and stuff that Imentioned.
How are people learning new waysto use their chat bots and like
how are they speaking to themand what are some different,
like even formulas for this iswhat I'm looking for and you can
(06:48):
tell it like you are a digitalmarketing expert or you
specialize in email marketing.
And there's a lot of newresearch that emerges about how
these formulas of how to speakto it and give it really good
feedback for better outputs.
Samantha (07:01):
Yeah, absolutely.
Engineering prompts, for thosethat are technically interested
in this.
Absolutely, and there's a ton ofthose available.
I follow a number of people onLinkedIn that share tips,
infographics, sheets like that,that help with prompts and
customizing that.
Make sure you're using thememory function as well and just
storing anything that you can,educating it because, to your
point, you have to tell itexactly what you want.
(07:22):
Once you do that, then it can dothe best work for you, but if
you don't take that step, it'snot gonna perform as well.
And I think that's where peopleget tripped up or that's where
they're getting that kind of toyour point, Tricia, oh, that's
clearly written by Chat.
So I think it's really importantto keep in mind, you have to
educate it, but also really bestrict about ensuring your brand
and your strategy is clear, thatyou continue to remind it of
(07:43):
that, and that you're checkingit as the human against how you
wanna position your company,your business, or your clients.
Tricia (07:50):
And I would add onto
that, that I talk to Paige all
the time about this'cause shekind of spearheads our AI,
trying to find ways to be niceto Chat.
I've changed my mindset of kindof treating it like a colleague
and giving it that feedbackbecause it's a lot of trial and
error with what are the rightprompts that are getting the
content that really resonatesthat I really like with minimal
(08:10):
tweaking versus, sometimes Iprompt it when I'm in a rush or
something and it's like, wow,that totally missed the mark.
When some people I think get intheir heads that, well, if I
have to spend the time to giveit so much background I can just
do it myself and save that time.
When really what you get in thelong run after just kind of
front loading it with thatinformation helps so much.
And then just tweaking thoseprompts.
(08:31):
Like Paige said, you're adigital marketing expert, or
you're working in communityrelations, public relations, or
whatever, just to kind of helpChat get into that mindset.
But I've noticed too, the moreyou give it feedback on that was
good.
And for me it was kind of hardto say thank you or please, or
great job to a computer.
It really does make a differencebecause it's telling Chat that,
oh hey, they really like thiscontent and it knows to do more
(08:53):
of that.
Samantha (08:54):
Give Chat a pat on the
back every once in a while, you
know?
Paige (08:58):
Yeah.
Tricia (08:59):
It sounds silly, but you
have to do it.
Paige (09:01):
It can be frustrating
when you're getting outputs that
aren't what you're looking for,and that's when you wanna
respond and be like, no, tryagain.
But if you're cleaning up thosememories and you're using those
formulas for how you speak to'em, giving content feedback and
specific feedback about what'sworking, what's not, giving the
data upfront, of course,adhering to the AI policies that
(09:22):
you should have in place.
You don't want to puteverything.
Some information is sensitiveand it should be only
information that is available tothe public.
You can look on an Instagram howmany likes something has or see
some insights.
But overall, if you combine allthose things, it will be less
frustrating to be nice to Chatbecause it'll have all the tools
it needs to give you higherquality outputs.
Samantha (09:43):
Yeah, I like to think
of it as, you guys have both
mentioned this, it's acolleague.
It's a partner in this.
You wouldn't hire somebody newon your team and have them sit
down, give them no background,and then say, go do this.
Right.
We wouldn't do that.
There's an onboarding process,so it's the same thing with
Chat, and to your point, Tricia,invest the time once and it will
pay dividends in the long runand then continue to be a
partner with it and give it theinformation and resources it
(10:06):
needs.
Now, we've talked a lot abouteducating it, how to go about
doing that and working with it.
But I also wanna touch on kindof the big elephant in the room,
which are risks and limitations,right?
So you mentioned the AI policy,Paige.
I think that's super, superimportant.
We have one on our website.
There's a lot of thingshappening with the government
and whether they're going torequire this or not.
I'm not legal counsel.
You can look into that on yourown.
(10:26):
But I would say definitelyeducate yourself on that, one.
But then also, two, be veryaware of the limitations.
AI is not perfect.
Fact check.
So make sure that if you'reasking it for information, tell
me about this or what about Xthat you're double checking
that, that it's accurate.
Perplexity is a fantasticresource for research.
I love it.
And it's also awesome because itwill cite where it's getting the
(10:49):
information, but then it alsogives you secondary prompts so
you can dig deeper and reallyget more information.
Make sure you're doing yourhomework with that and you're
aware of it.
And then also make sure, to yourpoint, Tricia, that if you're
asking it to help with ideation,that it doesn't just come up
with willy-nilly campaigns andyou just run with it.
You really have to be diligentto make sure this applies to the
brand, that it makes sense, thatit aligns with the strategy.
Are there any other AIlimitations, risks, or things
(11:11):
that I missed that we shouldtalk about, Paige?
Paige (11:14):
I think overall because
this landscape is evolving and
there's new platforms popping upall the time, you wanna be
transparent with your clientsand it's a really good idea to
just have a policy in place.
This can change over time, ofcourse, once you make one, you
can make adjustments to it.
The platforms that you're using,keep a list of them, make sure
you know if there's any updatesto that platform specifically.
(11:34):
Keep it as a toolbox, but it'snot something that's stagnant.
It continuously changes.
Remaining transparent and havingthat in place, keeping that
list, updating it frequently,would be best practice so you
don't get caught in a situationwhere you're using
misinformation or you're usingcontent that people can tell
isn't authentic and it's writtenby a bot.
Your engagement's gonna drop offin the long run.
Samantha (11:55):
Do the right thing,
always, is our motto.
Do the right thing whethersomebody's looking or not.
And I think that's really,really important is just kind of
self-police yourself in thismatter and take the safety road
I would suggest.
There are ways to do that whereyou can turn off the kind of way
that Chat can send informationout into the...
what is it?
The LLM.
Making sure that if you havework that you're doing and you
(12:17):
wanna keep it private, that youhave that turned off.
So it's not putting that outthere for the world to see.
But first and foremost, youshould just never put anything
into these programs that wouldbe proprietary or private in any
way.
I know a lot of people will evenmake up fictitious names for the
businesses or organizations thatthey're working with to make
sure that there's nothing thatcould possibly get out there
that ties back to it.
(12:37):
And just make sure that you'realways keeping it nice and neat
and that you're doing the rightthing, the right way as it
relates to that.
Tricia, any thoughts from yourperspective about things to
avoid or risks or just stuff tokeep in mind as it relates to
using AI for content?
Tricia (12:52):
Yeah, I think the
biggest thing is, we talk a lot
in content about authenticity.
That's a big shift right now incontent.
Everyone wants to see anauthentic brand.
They don't necessarily wanna seethe picture perfect, the kind of
front your business wants to puton, they wanna see the day to
day.
They wanna see the real you in asense.
And a lot of businesses can dothat in a variety of ways, but I
think that's kind of one of thethings some brands kind of
(13:13):
teeter on the idea of AI of if Iuse it I'm not being authentic
because I'm not the one creatingthat or whatever it might be.
But I think that's where thathuman touch comes in of using AI
to kind of streamline theseprocesses, kind of speed things
along and give you that goodbase.
But it can't replicate thathuman emotion.
So there's always tweaks to bemade that refine that brand
voice to really tie into theemotion that your audience
(13:35):
wants.
Or it may pull data that, if youdidn't give it enough
information, it may be from thewrong industry.
So to your point of factchecking.
There's just a lot of refinementthat can be done with the human
touch, that it doesn'tnecessarily mean that you're
lacking authenticity.
You're really creating it andmaking sure that it's there for
your business, but kind ofspeeding things along using it
(13:56):
as a tool.
It’s not going to replaceanything, it cannot replace
things, and if so, everybusiness that puts content out
there in the restaurantindustry, it's all gonna sound
the same.
So no one's gonna stand out.
But when you use it for theideas or just to kind of give
you that base of content, thenyou add in those emotions, you
add in those touches that reallymake your brand what it is.
And I think that's where thatauthenticity comes in, in a
(14:19):
little bit faster way, you know?
Samantha (14:21):
Yeah, same with if
you're using Gemini, or
Grammarly, or any of thosetools, educate it and help it
understand the brand voice, soit just gets you one step closer
to that authentic voice.
And Paige, you mentioned ifyou're doing work for clients,
but I would say even if you'rean in-house team, this applies.
If you've got a small marketingdepartment in-house, make sure
that you're all using a teamChat so it's gaining the
information from every personthat's involved and that you're
(14:43):
all working toward the samevoice, the same outcome.
Yeah, make sure you're doing allthose things so that you can get
the very best output possible.
Tricia (14:51):
Yeah, and I would say
that's one thing that we're even
working on refining kind of inour processes.
It seems like it takes thatextra time to give it more
information as you go along.
But if I'm already reviewingcontent and I know it's been
approved by a client, we wantmore of those captions and that
verbiage.
If I'm already looking at it, ittakes a couple more minutes to
throw it in Chat and say, hey,this is approved.
We really like this length orthis tone, or whatever it might
(15:14):
be.
And Paige kind of does the samething when she pulls data.
We're already evaluating thisdata, what's performing, what's
not, so taking those coupleextra minutes to just put it in
Chat and kind of inform it justas we're getting these things in
real time.
That really helps refine thecontent that you're getting.
Taking that time to do thelittle bit of input really
refines the output.
And the more you do that, thecloser you get to where you
(15:36):
wanna be to just add in thosequick human touches.
Samantha (15:38):
So let's talk about
when to use AI and when to avoid
it, because we talked about alot of ways that we can use it
in terms of idea creation orcontent generation, that kind of
stuff.
But I think there's some ways asit relates to that, that we
should avoid it and probablysome best practices or things
that we see where it's really,really helpful.
So, Tricia, where should weabsolutely use it?
(16:00):
And then Paige, I'd love to hearfrom you where we should not use
it.
Tricia (16:05):
Absolutely.
I think some great places tostart with using AI for
copywriting, kinda like we'vebeen talking a lot, is social
media captions or ad copy.
That really helps keep thingsfresh.
If you're writing a month ofcontent, that can be a lot.
And AI, the more you train itand the more details you give
it, it can even really help kindof space out your content to
where you’re not talking aboutthe same thing three posts in a
(16:25):
row and then you forget about itthe rest of the month.
So even small tweaks like that,it really helps kind of refine
that and kind of get those ideasand kind of the flow of content
a really good baseline for that.
Samantha (16:35):
Yeah, you can even
direct it and say, for example,
if you are a business that has amarketing strategy and a
specific campaign, you can say,I want X posts about the primary
strategy.
I want X about this campaign.
I want this many that are justgeneral value added, et cetera.
So you can really guide it toyour point of varying that
content, making sure it's goodquality content, not just
volume.
Tricia (16:55):
Yeah, absolutely.
And especially because there aretimes that, again, if I'm in a
rush, or I kind of just, I don'treally pause and take a moment
to see what all I need to put init.
I give it a prompt, I kind oftell it a couple things, and
then I realize, oh, I didn'tmention this campaign.
It's very easy to say, hey, Iforgot, can you add in three of
these posts or two of theseposts?
It'll shuffle everything aroundto really make it flow.
(17:17):
All you have to do is give itthat information and it's a
really good starting place forsocial media content and ad
copy.
And another way that I reallylike to use it or we as PTE is
for subject lines for emails.
You want those to be reallyengaging, and you can even tell
it give me three options tofocus more on this, or lead with
this.
Trying to give it those keywords.
That's kind of thatbrainstorming part of AI.
(17:40):
A lot of times it'll give mefive or six subject line
options, and I see two that Ikind of like, but I don't love
them by themselves.
It gives me that kind of ideaof, okay, can I mash'em up this
way?
Or, ooh, I really like that theyused that word.
Even if you're not necessarilycopying and pasting straight
from a subject line, it canreally give you a lot of great
options that can even then leadto good performance with open
(18:00):
rates and things like that.
Especially with email subjectlines you want them to be super
engaging, super enticing,because that's the first thing
users are gonna see, so you wantthem to be able to click it and
then open that email.
And then another thing that Ireally like to use AI for is to
repurpose content.
So like I said, we kind of tellAI what's really working for our
clients and what's not.
And if we have a post or a videothat worked really well, you can
(18:23):
then tell it to take that andkind of make social media
captions from it.
If we have a blog, for example,that is performing really well
or it's a really hot topic thathas a lot of really good
verbiage or maybe even datapoints in it, you can give it
that information and say, hey,can you give me an email
campaign for this, or ad copy,or whatever.
You've already done the bulk ofthe work to find that data and
(18:43):
get that information, that itcan streamline that process
super quickly to kind of breakout points for social media
captions, et cetera, et cetera,versus you having to go through
your blog and ok well I reallylike that point, and I really
like that point.
It can just really like analyzethat really well and kind of
streamline repurposing content.
Samantha (19:03):
That's a great point
and it reminded me of something
else you can do is not justtelling it I am a digital
marketing expert, or I am amarketing person, but actually
tell it who you're trying towrite for.
So I'll do that sometimes whenI'm trying to write a LinkedIn
post, I'll say, here's thevideo, or here's the particular
topic, and I want this to appealto X, Y, Z.
It could be a business owner ora marketing department leader,
or fill in the blank and I wantit to be engaging or I want it
(19:24):
to be informative.
So working both sides of that,how you want it to be written in
your voice, but also who you'retrying to attract and engage.
Great points, Tricia.
All right, so tell us what notto do, Paige.
Give us the rules, theguidelines.
Paige (19:38):
I would say the first and
foremost thing, we've kind of
touched on emotional connection,the emotional connection that AI
lacks.
So anything that requires astrong emotion behind it.
You can use it for likeideation, or as like a template,
but I think overall I would keepthings that are very emotional,
such as brand storytelling andlike the narrative and the
(19:58):
identity of your brand.
When you're making thatbaseline, that should be
something that comes from whenyou're creating your overall
marketing strategy and yourbranding from the start.
If you start at the baselinewith AI, everything that you
make from your brand story andfrom your brand voice and from
your marketing strategy is goingto be an extension of
automation.
So you want it to be authenticand genuine from the start.
(20:22):
And then the other thing I wouldsay is just any high stakes
content like PR statements orthought leadership content.
I would steer clear from AI inthose scenarios.
Samantha (20:30):
Yeah, absolutely.
Good points.
Well, this has been afascinating conversation about
one teeny tiny way that we canuse AI, specifically as it
relates to content creation.
So we talked a lot aboutauthenticity, the human element,
how to educate it at thebeginning, how to continue to
educate it so it can create thebest content, different ways you
can use AI in your contentwriting.
(20:51):
I hope this has been helpful forthose of you who have listened
in.
Thank you very much.
Thank you, Tricia, thank you,Paige, for sharing your
expertise.
We appreciate it and I hope youall tune in for another episode
of 1000 Ways to Market YourBusiness.