April 20, 2023 • 48 mins
On this episode of Industrial Strength Marketing, host James Soto sits down with Parry Malm, CEO and co-founder of Phrasee, to talk about his journey as a marketer and entrepreneur. Malm discusses how he fell into entrepreneurship and the challenges of marketing in today's technology-focused landscape. Soto and Malm delve into the world of AI, its impact on marketing, and the ethics of it all. This episode also explores the intersection of marketing, technology, and business strategy.
Articles and resources reference in this episode:
If you haven’t already, please review Industrial Strength Marketing to let us know your thoughts on the show.
About the Show
As a top manufacturing podcast and YouTube channel, we focus on what matters most to industrial marketers, sellers, and growth-minded manufacturing executives.
Featuring inspiring conversations with manufacturing and B2B marketing leaders on winning moves and lessons learned, this show exists to deliver insights that help you grow your business.
Are you looking to share your expertise with industrial marketers, influencers, and leaders? Click here to be a guest on the show.
About the Show:
The Industrial Strength Marketing Show is a top manufacturing podcast that explores the personalities, cutting-edge strategies, tools, and technologies transforming the industrial and manufacturing sectors. Each episode, hosted by James Soto, covers marketing, sales, business development, and integrating martech and AI into industrial B2B strategies. Tune in to gain actionable insights to help you stay ahead in the industry.
Let’s connect:
- Visit the podcast homepage
- Tweet us at @Industrialsays, @jamessotoism
- Follow James Soto on LinkedIn
- Email us at podcast@Industrialstrengthmarketing.com
- Subcribe to the YouTube channel
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Perry Malm (00:00):
I think what's happened to marketing in the
last 20 years is really sad.
Where marketers used to be, youknow, Don Draper types, sip your
whiskey in the corner office andcome up with crazy big ideas and
go hard with them. And now 90%of marketers time is spent in a
whole bunch of differentsoftware platforms and Excel
(00:24):
spreadsheets.
James Soto (00:34):
Welcome to the industrial strength Marketing
Show, we challenges leadersacross the supply chain to make
their people marketing andtechnology, their businesses a
strength of their business. Hi,I'm James Soto. I'm the founder
of industrial the family andstrategic marketing brands
trusted by leading organizationsaround the world. I'm sitting
down with innovators withindustrial marketers, and
(00:56):
technology leaders to talk abouttheir careers, their insights,
marketing, and technology thatmakes an impact on their
businesses. I'm here today withParry Malm, CEO of frasi. I've
been following crazy for a longtime, I've been speaking about
it at events around the countryin the manufacturing technology
space. And we're gonna reallyget into some great things
(01:17):
today. So Perry, welcome to theshow.
Perry Malm (01:21):
Yeah, thanks.
James Soto (01:22):
We're really glad to have you here. And I just
thought we just get right intoit, like, tell, tell our
audience a little bit about yourjourney as a marketer, as an
entrepreneur, you know, and, youknow, just as an expert, and
your little slice of the AIworld?
Perry Malm (01:38):
Well, I mean, like most people, I fell into it. I
think people who aspire to beentrepreneurs are constantly
chasing their, their tails. Andthe people who I think tend to
have the most success are peoplelike me who never intend on
doing it. But discover aproblem, experience a problem,
(01:59):
and then figure out ways tosolve that, that problem. The
fact that I'm I know a bunchabout AI is cool and stuff.
Doesn't get me a lot of friendsat parties and whatnot. But
ultimately, you know, the mostimportant thing is that we solve
customers problems, the factthat we use AI is cool. But the
most important thing is theproblems get solved.
James Soto (02:21):
Yeah. And, you know, that is that is the fundamental
reasons of starting a business.
So when you looked at startingthe organization, and as the
right place at the right time,what was the quintessential
problem in the market you'resolving for?
Perry Malm (02:35):
Yeah, so this goes back a number of years, when I
was a branch site marketer, Iwas living in Amsterdam at the
time, and we would send outmillions and millions of
messages to millions, millionsof people. And I used to always
look at it and be like, what isthe right message to send out
via I mean, back then this wasin 2007, or something. So it's
(02:57):
mostly on email and brochuresand direct mail and stuff. And
he's always go like, we'respending millions of euros.
Surely, we should spend a bit oftime to figure out what language
is good. So I started doing abunch of tested and tried to
build a bunch of models and onething led to another and before
long, I had the light bulbmoment, going, if I'm
(03:19):
experiencing this problem, thenprobably others are too. And it
turns out, I was right.
James Soto (03:25):
Yeah. And tell us a little bit about Phrasee. Like
one of the things we've heardfrom our audience, is they want
to hear unapologetically whatyou do. How does it help them?
Hey, James, you know, you know,I know, they're gonna say, have
Perry, tell us more about whatFreezy does? And, you know, what
are these amazing things thatyou're unlocking with your
technology?
Perry Malm (03:45):
All right, well, welcome to the shameless plugs
zone.
James Soto (03:50):
No, one apologetic,
Perry Malm (03:54):
The concept with Phrasee is very simple. We use
technology to not just createcontent, but to create good
content. Now, you may wonderwhat good means. And it means a
number of things. First of all,content that's generated needs
to be on brand, and it needs tobe trustworthy. It can't be
(04:15):
factually inaccurate. It alsocan't invert and people with the
need to continuously approvethings because they don't trust
the AI itself. And ultimately,that must perform strongly. So
that's ultimately what phrase hedoes. We we generate and
optimize and understand thelanguage that makes people
(04:38):
click. So the greatest andeasiest way to think about it is
that if you get emails, pushmessages, Facebook ads,
whatever, from companies likeGroupon or Walgreens are like
loads and loads of companies,then you've experienced raising
technology without knowing andyou've probably bought something
that's a direct result of thelanguage we not only generated
(04:59):
but optimized, and dependingupon what you bought, either
say, you're welcome, or I'msorry, I can't be in charge of
your own shopping habits.
James Soto (05:08):
Yeah, I have used an example from phrasing that was
like driving like a 30%improvement rate. And I think it
was subject lines for Facebookads, it was something about,
like, I used one of the examplesI think he showed, which was,
you know, one, you know, onelevel of copy that was around,
like, oh, take this vacation,and, you know, summer savings.
(05:31):
And then it's like sun sea sand,fun, you know, like, get in on
the action, bunch of emojis. Andwhat was amazing about it, and
this one felt magical to me,which was that the AI was
outperforming, you know,something a human row. And also
as was unlocking, like the useof emojis that like, didn't have
(05:54):
bias they didn't, there wasactually some thinking and
advantages to how it how itactually performed the job. And,
you know, we've I've done this,I've spoken with that example,
at a major technology event, andpeople actually kept picking the
AI is the better copy. So. Sothat's like, one of the things
that inspired me to reach out.
Because, you know, I think whenI kind of look at things, and
(06:19):
what inspired me, in thisepisode was really a quote from
Arthur C. Clarke. And it about,you know, it really kind of
dives a little bit into thisworld. And it was in his book
profiles of the future that hewrote any sufficiently advanced
technology is indistinguishablefrom magic. And I think that is,
(06:40):
I think, a lot of what we'reseeing out here. So when, in the
context of that, like, what doesthat make you think of when you
think about what we can now dowith the technology? You know,
in terms of AI for marketing forcontent, stuff that people can
see, they don't have to questionit, you know, they can actually
use it. You know, what, when youthink about like it being
(07:01):
distinguishable from magic,like, what does that, you know,
what do you think of in thecontext of your world and AI
marketing?
Perry Malm (07:09):
Yeah, I think people who view it as magic probably
don't understand how it actuallyworks. So it might be worth
spending a moment to explainwhat a lot of this stuff does.
Effectively, what AI is, is it'svery fast moving statistical
models. So there's sort of twomain schools of it. One is
(07:30):
prediction, or decisioning. AI.
And the second is generate ofAI. So the former prediction and
decisioning is effectively youknow, when you go into Excel,
and you got a scatterplot, andyou go like insert trendline.
That's effectively what the vastmajority of machine learning
systems are doing, except withmany more variables, many, many
more data points at a muchhigher speed. From a generative
(07:53):
standpoint, obviously, everybodyis, has been exposed to chat
GPT, which is like prettyimpressive technology. Use Cases
notwithstanding. But what iteffectively is, is it's a huge
statistical model that canpredict the next best word at
scale. So it can be given aprompt, and based upon that
(08:14):
prompt, it can then go, well,here's how I should structured
that answer using this huge sortof, like, multibillion parameter
vector space of, of languagetokens. So while it appears
magical to the layperson, whatit effectively is, is a huge
amount of really fast movingstatistics, which presents many
opportunities, but also manyrisks.
James Soto (08:38):
Yeah, yeah, for sure. And I think we're now
getting in the world where, youknow, this has been opened up to
pretty much everybody. And, youknow, we're at the point where
we're looking at like phraseengineering, because you can
even take that to the next leveland leveraging the technology,
and get even better and moresuccinct answers. And so, you
know, as you really look at thecore technology, and this is
(09:02):
kind of where our conversationstarted. Going. LinkedIn was,
you know, with that magic, thereare real issues, when we look at
that predictive models, thepatterns, the patterns, a
pattern apply, you know, applythe insights and generative AI.
You know, it comes to me likethose key impacts, we see
impacts on a number of areas,and I think there's like three
(09:24):
different areas. So, you know,we have, you know,
personalization, you know, andprivacy, we have the bias and
algorithms as an issue, whichreally can impact, you know, the
human element to potentiallynegatively impact outcomes. And
then there's the impact. I thinka lot of folks are feeling
(09:45):
today, whether you're a brandside marketer, or you're even in
an agency so to speak, or you'rean industrial business, it's
like we're seeing inmanufacturing, you know, big
displacement due to automationand AI is also going to have an
impact. So there's that thirdthing which Does that impact on
employment? And so withoutgoing, you know, mainstream,
(10:06):
it's just pretty amazing. So,you know, what, what do you see,
you know, like this, you know,chat GPT the gender of AI, or
even, you know, what's crazydoing that is so truly
remarkable, you know, that, youknow, you're gonna have to, you
know, really kind of come outwith, you know, real answers to
those questions. So, you know,so since it's been thrust on us
(10:29):
all, you know, what I'm, youknow, what are your concerns
regarding the ethical use of AI?
So, you know, just set, let'ssay, just started with
personalization and privacy, youknow, how does that apply to it,
you know, crazy is doing?
Perry Malm (10:45):
Yeah, well, we're, we're fortunate in some ways,
we're, we're a little bitdetached from many of the
challenges that that hyperpersonalization and big data
warehouses companies face, butjust sort of diving into the
whole school of generative AIand some of the the ethical
conundrums which one must face,particularly with large language
(11:07):
models, they're trained uponeffectively the entire Internet.
So what they did is they scrapedthe entire internet, more or
less, so all of Wikipedia andall of your old Tumblr, blogs,
and all this kind of stuff. Andthen they basically turned it
into a to a series of individualtokens, which are like words and
(11:27):
stuff, and then produced thishuge statistical model. So like,
from a technical standpoint,it's absolutely fascinating. But
the challenge is, it's trainedupon the entire Internet. And I
don't know if you've ever beenon the internet, but it's filled
with idiots, like 99% of thecontent on the internet has been
(11:51):
generated by an idiot. Peoplewith bias, people with wormhole
world views, all this sort ofcrazy stuff, all this sort of
embedded bias that we have, inUS only at a huge scale. So what
that means is that this model isinexorably biased. If you ask it
(12:13):
to produce content, targetingwomen, or targeting men, or if
you ask you to write jobdescriptions for specific job
roles, or if you ask it to, thisis a trick I pulled to generate
the pseudocode algorithm topredict somebody's credit score
(12:35):
based upon their gender andethnicity, then you can really
start to uncover the biases.
Some of them are overt, you knowabout race and gender and stuff
like that. But some of them areimplicit biases invisible to the
human eye. And I think that'sreally quite dangerous. You
know, the way that I phrase chatGPT is it's sort of like an
economist. They're alwaysconfidence, and sometimes,
(13:03):
right. So now we have thisauthoritative, authoritative AI
system that can producecompelling answers that only
sometimes are right. I don'tknow, that strikes me as being
quite existentially challengingfor truth for the internet. And
moreover, for marketers, whereif you're in trusting systems to
(13:26):
generate content for you, how doyou know if it's right? How do
you know if it's biased or not?
And ultimately, how do you knowif it's good?
James Soto (13:35):
Yeah, it's almost like freedom, you know, the we
created here in the States, wecreated a country that's based
on, you know, with that freedomis the assumption that we'll use
it responsibly. So, so if we'relooking at an internet of
idiots, so to speak, is it youknow, that we've really gone out
there, and, you know, theassumption is, when we're
scraping with, you know, gettingall that data with AI, that that
(13:58):
content was created, you know,with good with good thinking and
responsibility. And I think wecan argue there's a lot of bad
stuff out there, but there's alot of good too. But the
question is, what's, you know,how do we how do we scrutinize
that, you know, and I think inthis world, you know, that,
that, you know, we do have thisopportunity to personalize our
(14:20):
advertisements, you know,campaigns. So, it's not just
what we write, but it's actuallyhow we're doing that targeting
and personalization. And so theAI can tailor marketing messages
to be more relevant. And thennow we're talking about
influence and, and those things,you know, how to be effective by
looking at behavior,preferences, demographics, but,
(14:41):
you know, that relies to yourpoint on collecting a ton of
data and analyzing a vast amountof data. So in that context, you
know, we look at those privacyconcerns. So it's really
essential that we, we we reallybe transparent about Howard
collecting and using customerdata. And obviously we have to
(15:04):
give, you know, the userempowerment to opt out. But in
that context like how do yousee, let's solve for that in the
sense that, you know, how shouldbusinesses that are engaged in
AI share how they collectcustomer data? How do you see
that?
Perry Malm (15:21):
Yeah, I mean, it depends on what data is being
collected. And how sensitive thedata is. A good example is that
if somebody is collectingpersonal health data, then
presumably you need to be veryupfront about it. But if
somebody is collecting, like,when you bought dog food last, I
mean, you could certainly makearguments about how his personal
(15:46):
data, but also like, who reallycares, right? I think just being
honest and upfront with peopleis the key. I mean, it's when
companies, you know, pull outCambridge Analytica and collect
data and use it for spuriousmotives, which I think is is
(16:07):
morally wrong, but also leads tosocial ills. But like, if
you're, if your like, cornerstore, knows how many Red Bulls
you bought in the last week, whocares?
James Soto (16:21):
It's an inventory question. What do we need for
Joe? Or for Perry? You know, ifyou look, you know, and to the
extent that, you know, you haveyour business here, but if you
were to paint a picture of whatlike privacy looks like now, you
know, versus like, what, like,what paint a picture of it like
now versus let's say, five yearsfrom now, you know, what do you
(16:45):
see privacy looking like, youknow, given the rate of
evolution and innovation, andwhat's really going to be put in
market?
Perry Malm (16:52):
Yeah, I think it's gonna vary jurisdiction to
jurisdiction, and it'sultimately going to be
regulatory frameworks, which areput in. So I know that the the
EU already has some ratherstrong privacy laws with GDPR,
and all this kind of stuff. AndI think those are going to
strengthen. And we're probablygoing to see some level of
(17:16):
regulatory frameworks put in bythe EU to cover off generalized
AI systems like chat, GPT, andwhatnot. In the US, I think the
path will be quite different.
There will be regulations putin, but it's going to be driven
through case law, rather thanlegislators. So there's already
two interesting lawsuits, whichare either happening or have
(17:40):
been threatened in this space. Iknow that there's a class action
suit, which is either in theprocess of being filed or has
been filed versus GitHub copilot, claiming that there's
been copyright infringement of,of their source data that's
collected through open sourcecode and proprietary code, which
(18:01):
is held on GitHub, I don't knowwhat's going to happen with
that. The second one is, I wantto say stability, AI has been
threatened with a letter ofintent from Getty Images, who
claimed that their imagegeneration AI, illegally scrapes
Gettys imagery. So they'reclaiming copyright infringement.
(18:23):
Also, the EU tend to be more topdown, they have you know, it's a
it's a code of laws type thing,where they'll come in, and put
in a framework, which willprobably be clunky. But it'll
keep a bunch of bureaucrats inwork in Brussels, so good for
them in the US, is going to keepa bunch of lawyers, kids, in
private schools. So that'ssomething which we should feel
(18:46):
thankful about. And then in theUK, I don't know, we'll just do
something, it'll probably bewrong, and nobody cares.
James Soto (18:54):
Yeah, so it's the Wild Wild West, as we say, here.
And there's a lot happening.
And, you know, when we were whenwe were kind of going back and
forth on LinkedIn a little bit,you know, the kind of second
point here is that there's realethical concerns around the, you
know, the use of AI andmarketing. You know, the
potential for bias inmanufacturing was a huge, huge
(19:15):
skilled labor gap. So we do alot of recruitment marketing for
manufacturers and folksthroughout the supply chain. So,
you know, there's a lot ofpotential for bias in the
algorithms that power AI, andthe marketing of their, of it
with it. And these AI systemsare only as unbiased as the data
and the people who, you know,who've influenced the training.
(19:35):
And so if that data is, youknow, biased or skewed in some
way, you know, like, yeah, thethinking is, it'll be reflected.
So, like, suppose like, AI, likea system is trained on data,
like disproportionatelyrepresenting certain demographic
groups. You know, like we, youknow, we were obvious They will
(19:58):
say, you know, they'll make someinappropriate or wrong
assumptions. And so, ultimately,you know, we'll be unfairly, you
know, targeting people ordiscriminating, you know, either
intentionally or notintentionally. So, ultimately,
to that point, when you look atthe, you know, ethical use of
AI, looking at the biases, youknow, how are you looking at
(20:23):
mitigating that risk? Andbecause you're your front facing
with content and copy? So how dowe, you know, with marketers
having to really look atdiversity, equity, inclusion and
belonging? How are you ensuring,like, how do you see us ensuring
that we're being diverse andrepresentative, you know, in
actually what you are building?
Perry Malm (20:44):
Yeah, so first of all, what many people think is
that the source data is theproblem, and it sort of is, but
without that huge volume ofsource data, the large language
models wouldn't be possible inthe first place. So you're kind
of stuck with a huge amount ofdata that probably is biased and
over represented with middleclass plus white males, right?
(21:08):
But the thing is, you need thathuge volume, you need the
billions of parameters to getthe level of, of complexity that
large language model has. Sothen the question has to be, how
can you post process the output,and set rules and frameworks to
(21:28):
make sure that when it generatessomething that is either
hallucinating facts or isovertly bias that you can
overcome it. And that's reallywhere we've focused over the
years, I mean, we came up beforelargely with models existed, we
built our own languagegeneration system, which, which
is still in use, we'veintegrated a bunch of the large
(21:49):
language model stuff. But whatseparates us is that we've
already built all of these toolsthat can guide it at source, and
then also manicure it after thefact. Many systems do not have
this, and I would imagine, atopen AI, they're very actively
(22:11):
injecting directionality. Sowhen it first came out in
December, you could have agenerate all sorts of
reprehensible stuff, all theobvious stuff, it could, it
could, you could tell it to saythat, that Hitler was a super
nice guy or something completelywild like that. And they put in,
you know, a whole bunch of awhole bunch of these rules. The
(22:34):
challenge with that is thatsomebody is always going to
figure out a new way to be anasshole. So how do you keep up
with it? Fortunately, forFreezy, we're in a much work
contained environment, becausewe're working with with
marketers, the challenge becomesmore about not the ability to
generate content, that's justthe catch the challenges, then
(22:56):
the yards after the catch to getto the end zone. And that's
really where our focus is.
James Soto (23:01):
At. And so that's an iterative improvement of, of
what you're saying, as you'regetting insights, you know, in
real time, is that, is that howyou see? Exactly,
Perry Malm (23:11):
exactly, it's like, no, the ability to generate
language is effectively acommodity. Um, you got shut GPT,
which is effectively like amillion monkeys on a million
typewriters, and maybeeventually, they will write the
complete works of CharlesDickens. But you need to be able
(23:31):
to tell what's good for bed,before you put it out in the
wild. So that's really where ourfocus is, is to tell good from
bad in terms of brand, in termsof accuracy, in terms of
trustworthiness, and ultimately,in terms of performance?
James Soto (23:47):
I think that's a great point. It's, it's, if if
this world and the ability toput out content, you know, not
is no longer a differentiator,it's a capability, it's an
expectation, we can use thismachine assisted technology to
create copies of great content.
And that's, you know, availableto all, you know, how do you
win? How do you go up andbeyond, and to the extent that
(24:09):
you have your niche here, andyou're focused on marketing, you
know, because it has a differentstandard, everything from the
brand, to the audience, andpeople who are really trying to,
you know, consider thoseconstituents. And essentially,
you guys are connecting the dotsto make sure that you can drive
real value engagement,entertainment, you know,
education, to drive, you know,to inspire people to take
(24:33):
action. And then as thosesystems work, you can
continuously improve it toimprove outcomes. And that's the
beauty of AI. It's hard forhumans to replicate that, you
know, in terms of, you know,iteratively, you know,
continuously improving content.
And so, you know, so that, that,you know, there's a there's a
(24:53):
big issue here and kind of toget into the third issue here,
and curious to see how it's inpack With us, the third area
we've seen, you know, you know,outside of the bias side of it
is the impact on employment. Soas we look at, we look at
employers, right now inmanufacturing. There's about
(25:16):
800,000 open jobs in the USmanufacturing, manufacturing has
been going through a lot ofchange with automation, we've
seen displacement of jobs,because you can, you know, where
robot can do it better, youshould be having a robot where
you need humans to operate thoserobots and have created
opportunities. But I think thewriting's on the wall, the
(25:38):
combination of, of automation,you know, of AI, and you know,
the ability for that AI to getsmarter and smarter and smarter,
is really going to havedisplacement. And so when we
look at this in the context ofus as marketers, and how do you
really see that impact onemployment?
Perry Malm (26:02):
Yeah, I mean, certainly, if we go back to the
cotton mills in Lancashire, inthe Victorian era,
James Soto (26:12):
The Industrial Revolution, that's where it
started the spinning jenny's.
And they made those so eightpeople could do the work of one,
that's literally what we westart the Industrial Revolution
story on the car.
Perry Malm (26:22):
Yes, absolutely. And you had the Luddites, who came
into the factories, and theytried to bash the machines
going, it's going to disrupt ourway of life. And like, sure, you
know, a few of the artisans whohad you know, lose their front
room, had to retrain, find adifferent living, whatever. But
it created so many more jobsafterwards. And that's how these
(26:43):
things work. If you think aboutthe, like, what 1995 or
something when the internetreally started proliferating. A
lot of people in newspaperssaid, Oh, this is going to is
going to kill our jobs. And itdidn't like there's more people
writing news content now thanthere was back then. So like,
(27:06):
when people worry about specificjobs being lost, usually it's
because they have a vestedinterest in perpetuating the
status quo. What automation isgoing to do what AI is going to
do is it's just going to meanthat will lead to have millions
of people in the workforce,retrained and focused on doing
(27:30):
something else. I don't believethe scare mongering of people
saying that we're going to haveyou know, 20% structural
unemployment in the future. Itjust doesn't make sense. Like,
history doesn't indicate that'sgoing to happen. There will be
some job categories that AI willprobably threatened. And it's
(27:52):
going to be probably low leveladministrative work. And
hopefully, some high levelconsultancies
James Soto (28:03):
Is that a dig?
Perry Malm (28:06):
The thing is, what these large language models can
do, is they can aggregate let'ssay best practice, right? It
control across millions ofwebsites, and people who have
written about topic X pick anybusiness strategy or any
business conundrum. And it cangive you the vanilla ice cream
(28:28):
tried and tested answer. Right?
It's probably not the mosteffective answer, but it's not
going to be the worst answer.
It'll be the vanilla ice creamanswer. And that's what a lot of
consultants, do. They give thevanilla ice cream answer because
they've read the textbooks. Butno, you don't need them. So I
(28:49):
think it's going to be quiteinteresting in 1520 years time
to see what the role ofconsultancies are, there will
still be a role for them, butit's not going to be giving
pithy advice on businesschallenges. Yeah,
James Soto (29:06):
you know, so you hit two things here that I think are
really important. So one, youknow, you know, there's a
consultancy, and I'll talk aboutframeworks a little bit, but
then there's this, you know,first thing he said, on the
Industrial Revolution, so, so weshould look at this with
optimism, because when you lookat that displacement that had in
the early Industrial Revolution,there was humanity were like,
(29:27):
they needed X amount of men todo a job to harvest in an
agricultural society, you know,and you would work there and you
would work seasonally, a lot ofseasonal work, the world was a
lot different and what theIndustrial Revolution unlocked
and yes, it may have displacedfolks in you know, in the
textiles industry. But what itdid is a credit the factory in
(29:49):
the first time, you know, youwould literally be able to go to
a place and work year round. Youknow, what factories were
impacted by the weather and theconditions and, and so you can
literally go and it created, inessence, the middle class and a
place where you can go tohopefully get an apprenticeship
and work and, and have a careerand really create a
generational, you know, work. SoI do think that displacement
(30:13):
created the ability for folks towork like never before and know
here the Homestead Act in theUnited States took people and
they traveled, they traveledwest, right. And that failed.
But what happens is they cameback to the urbanization
movement, and they came back tothe city centers where the
factories were. And so that is avery, very true tale, a very
(30:35):
interesting story. So for everydisruption, you know, 100 years
ago, 100, a little over 100years ago, the US economy was
97%. Agricultural. And nowthat's down to three. What
displaced it the IndustrialRevolution, then what's this?
And then now we look at, youknow, that was a primary part of
(30:55):
our economy, and now it'shovering around 11%. What
displaced it, you know,information and technology, so
you've got to make your way ofliving life and doing business
obsolete before generationaltechnology and market forces,
you know, do do for sure. So Ido think there's this really
great opportunity moving forwardto really, you know, look at
(31:16):
these, you know, operaopportunities of the of the
future, because there's, thereis going to be change, for sure.
And for people to like, hold onto those constituencies and, you
know, have those worries, youknow, we have to, you know,
understand where we're going tobe and something I heard, I
(31:37):
think Gary Vee said it was that,you know, 10 years ago, I was
saying like, you know, if you'rea yellow page ad salesman, you
better be looking at what Googleand Yahoo are doing. You know,
when they're first startingYahoo, formed in 94, Google in
98, and I worked for anindustrial directory, which is
the biggest print industrialdirectory, and they wanted me to
(31:58):
help them really kind of evolvetheir team and their go to
market model, especially theirsales organization, to be
digital leaning. So we're allgoing through these disruptions.
And so, you know, as we, youknow, look at these same tools
and this disruption, how do yousee that playing out, you know,
for marketers, you know, becausethere's just so many things, you
(32:20):
know, yes, there's aredundancies? Yes, there's
things that can be automated,but like, where's the writing on
the wall? What areas?
Perry Malm (32:29):
Yeah, you know, I came up as a marketer. And I
think what's happened tomarketing in the last 20 years,
is really sad. Where marketersused to be, you know, Don Draper
types, sip your whiskey in thecorner office and come up with
crazy big ideas and go hard withthem. And now 90% of marketers
(32:54):
time is spent in a whole bunchof different software platforms
and Excel spreadsheets. Andthey're expected to, to just
execute, you know, they're justoperators. That's all they are.
There's very little thinkingdone in marketing. You know,
there was a four P's right, likeproduct, price place promotion,
(33:18):
I think it is marketing. No,it's promotion, and product,
price and place are sort ofdivvied off elsewhere. And I
think it's a real shame, becausegood marketers can contribute a
huge amount of value to theentire business into the
products, which are being builtinto, you know, how you actually
(33:42):
distribute it, all this kind ofstuff. And yet know the majority
of marketers time is like, likedeciding, oh, should we up our
bid on this Google Ad by 20cents or something? It's just,
it's taken a very interestingjob, and has made it very
(34:04):
boring. And I think that's areal shame. So hopefully, what
some of these new technologieswill do is get rid of a bunch of
the boring stuff, automate, oweda whole bunch of the boring
stuff, and make marketing.
Interesting again, I love
James Soto (34:20):
it. So you're talking about the manufacturing
experience, right? We have toautomate out, remove repetitive
work and go to high creative,high cognitive skills. You know,
like, that's what we want to doas humans, like, everyone. Some
people say they're creatives.
And those people are notcreatives that straightener,
there's data and studies thatsay, like, No, we're trained out
of that every kid's creative,they play with their kids, and
(34:42):
you can recapture thatcreativity. And I do think, you
know, that's what's happening. Ibelieve the writing's on the
wall, that if there is a processor something you repeat, those
things need to be you know,really considered for candidates
for these AI you know, Toempower technologies related to
marketing sales data, you know,the whole mix there. And that is
(35:04):
going to be framed to folks, butit's true. How much of our time
are we spending doing thingsthat, you know, only humans can
do. And that's the imagining,you know, our head of studios
says, I need time to dream,James, where's our dream time,
and whether your draw you, youknow, you're Draper or you're,
(35:25):
you know, you're just the, youknow that you're the marketing,
art director somewhere, youreally want to be able to do
those things. Because thesetechnologies, if we use them,
right, can really free up andcapture our imagination. You
know, I'd hate to see a worldwhere AI is being creative for
us completely. And we're devoidand detached from the process.
(35:49):
And then lastly, you mentionedthe displacement of
consultancies, I find that veryinteresting, because I think
that when you look at a lot ofconsultancies, like McKinsey and
others and Deloitte, they haveframeworks and students study,
they're frameworks they use, andthey're well vetted, they're
very robust. And, you know, I dosee that same issue happening
(36:12):
with consultancies, where youcan use these, you know, AI to
really disseminate like reallygood frameworks that you can
make your own. And that you canwalk through whether you're
going to go to market strategy,you're doing a transformation
effort, you're looking atBusiness Model Generation, or
you know, positioning, there'sframeworks for these things, and
(36:32):
that I believe, through thesetechnologies and the ability to
create these things. You canalso co pilot that with AI and
Bureau research, to reallycreate great ways that you can
get humans because frameworksyou have to put a human through
so a business can make decisionsabout it to really impact how
that will affect audiences ormarketers can do the same. We
(36:53):
all have frameworks, and I thinkthe the age of the consultancy,
we all want to be paid based onour thinking, not just making
things or doing things or doingsomething by the hour. It's
really about where do we bringthat strategy, that creativity,
that dreaming to the to theforefront? And I think AI could
be opening up a whole new set ofpossibilities. And, you know, to
(37:14):
your point, Perry, I just wantto key in on, I think you just
have some really great points isthat if we're going to resist
this, you know, from thestandpoint of our our biases and
holding on to the past, I thinkwe're missing the point, we're
going through yet another set ofevolutionary forces that we
(37:34):
really have to contend with. Soyeah, totally Yeah, to,
Perry Malm (37:42):
like, people can be Luddites, and they can pretend
it's not happening and they canfight it. That's fine. But like,
what happens the Luddites?
James Soto (37:55):
Know what to believe and know what a Luddite is. I
actually write about it, becausethat's one of the things I
talked about in teaching folksabout the Industrial Revolution.
And then like how that form the,you know, the unions and the
weekend was great it because ofthe fact Yeah, that's a
byproduct of the IndustrialRevolution. There wasn't even a
weekend such a thing of theweekend. So, you know, I think
(38:17):
there's not going to be a lot ofLuddites that, you know, you
know, would want to, you know,make a bigger workweek or, but
yeah, there's, there's a lot,there's a lot for us to upon us
now. And COVID had a huge impacton that.
Perry Malm (38:31):
Like, like, there's even a world. Yeah, I'm not sure
if this is gonna happen or not.
I learned a long time ago tolike, not make bold predictions,
because the world isunpredictable. And that's one of
the wonderful things about thehuman experience, right. But
there is a world where we becomevery homogenized in the world
(38:53):
experience becomes very boring.
That's one of my fears. So like,no, like back in 1998, my
brother and I backpacked aroundEurope for three months. And we
were really surprised anddelighted by so much we saw like
we saw all sorts of new thingscrazy things like her different
(39:16):
languages even use differentcurrencies back then it was
before the Euro was adopted, theTreaty of Maastricht. But no, if
you go on holiday, first of all,the experience is very
homogenized, because every cityhas a McDonald's and a Starbucks
and an Irish Pub. But you alsolike research it before you go
(39:39):
away, so you already know whatyou're going to see before you
go. And I think one of the riskswith AI, is it it makes the
entire human experience likethat, where we, we we optimize
everything in automate, automateeverything in our lives. which
(40:00):
means that the world becomes onebig blob. And I think that we as
people, as humans, and asmembers of society, need to
ensure that there's still thingsthat surprise and delight us. If
we ever get to the point wherewe're never surprised or
delighted. You may as well justjump off a bridge.
James Soto (40:25):
Yeah. And to the extent that something is so
predictive, prescriptive andpersonalized every time. And
that becomes our world, right?
We expect that and we open upNetflix writes, predictive,
prescriptive and personalized.
Where are those things that younever would have ever thought to
do? Or run into? Wherespontaneity? Where's that
(40:46):
serendipity in our lives? And tothe extent that this becomes an
overlay? It could be soulcrushing? I completely agree
with that. And I think as welook at the ethical use of AI,
you know, personalization andprivacy, those are big issues.
But I think the net outcome ishow does it affect the human
condition? You know, I think aswe look at, you know, you know,
(41:09):
even the workforce, as you know,as you look at, you know, tell
us, like, how big is your team?
By the way, it crazy how many,how many humans do you got,
Perry Malm (41:18):
you have roughly about 100 people? So 100
James Soto (41:22):
people, you know, do you have this conversation on
how this, all these things aregoing to impact them? And, you
know, is it the responsibilityof your organization to really
kind of consider that, andmitigate potentially some of
these impacts on them?
Perry Malm (41:39):
Yes, and no. So, ultimately, you know, a
company's objective andfiduciary duty is to its
shareholders, which is a verydepressing way to think about
business, but that's ultimatelythe truth. So like, with that
said, we do have a duty of careto our employees, also, we need
(42:03):
to make sure that what we'redoing, Will, we'll advance them
in their careers and expose themto things that they that they
can carry forward in 10 yearstime and when, you know, maybe
Freezy will be will be like, onthe NASDAQ or maybe maybe will
just cease to exist, you neverknow what's gonna happen in 10
(42:23):
years. But like, do we do wetalk to them deeply about like,
how AI is gonna affect theirjobs and whatnot? No, because we
we live and breathe it. Andpeople don't join crazy with the
expectation that AI is not goingto affect them. They join crazy
because they know that AI willaffect them. So I'm probably not
(42:46):
well placed to, to answer thatquestion. Because we live and
breathe this stuff every day.
James Soto (42:52):
It let me let me ask you this. So we have a lot of
industrial marketers, we have alot of executives of, you know,
industrial companies that, youknow, in our sphere and our
audience, these are marketers,sellers, people on the digital
side of the organization, tryingto bring together marketing
sales, customers experience, youknow, what would you say to
marketers, who are really tryingto grasp with what's the role,
(43:15):
you know, machine assistedwhatever, you know, in their
roles, like, how do they? Whatwould you have to say to them in
terms of how they need to reallyreconcile AI? What, if anything,
should they be doing right now?
Perry Malm (43:30):
I would say, figure out the things that you like
doing the least. And then use AIto do those things. You know, I
mean, a great example is likewriting loads and loads of
marketing content over and overand over is quite arduous.
Especially the leg short stuff,like subject lines, right? Who
(43:54):
wants to sit there and writesubject lines every day, or
Google ads or feed like Facebookads, when you're doing this
stuff at scale? Who wants to dothat? It sucks. Like it's not,
this is not what Don Draperdoes, you know, he doesn't sit
there on Madison Avenue writinga subject line. Nobody wants to
do that. So automate it, that'sone phrase he does. But if you
(44:18):
like doing some stuff, then keepdoing it. I think, you know,
there's so many different toolsand so many different
opportunities to automate andoptimize every part of the value
chain in business, andparticularly in marketing. So I
think marketers should be alittle bit selfish. Do the stuff
(44:40):
you want to do. And then Viceoffered to do the stuff you
don't want to do? I mean, what'sthe downside?
James Soto (44:47):
Yeah, yeah. Minding that balance that you were
talking about before, whereyou're not just over automating
it and your job becomes soullessand let it free you. But if it's
not freeing you, you know, thenYou know, what is it? You know?
What is it doing for you ifyou're caught up, you know, in
the minutia of the technology, Ithink the chief martech a list
(45:09):
of all the marketing relatedsoftware technologies, I've seen
it grow from 500 to 1000. Andyou know, just last year reach
10,000, marketing, automation.
And now obviously, growing AIrelated technologies, it's so
massive, it's, it's almostimpossible to keep up with all
the distinct categories oftechnologies that are out there.
(45:31):
So as we wrap up, you know, getgive us that yet again, plug
for, you know, for whatinformation listeners can use to
follow learn more about Freezy.
You know, what are you guys upto? What is the cool events?
But, you know, I know everyone'sgonna be dying to learn more
about you. And frankly, how dowe how do we connect with
Perry Malm (45:54):
you? That's super easy. frasi.co. And I'm the only
Perry mallam in the world.
Weirdly, I looked on LinkedInthe other day, and there's a
Perry Milne spelled p e r ry, inNorth Dakota. So I don't know
how I feel about that I feellike 80% Less unique than I did
before. But Perry with a male,I'm the only one. And as far as
(46:20):
freezing goes, look like, if theproblem you've got is you need
to produce highly effectivecontent at scale, then Freezy is
a good solution for you. And weshould talk. If that's not your
problem, we can still talk Ijust want to try to sell you
anything.
James Soto (46:42):
Yeah. Well, that's great. And, you know, thank you
so much for making, you know,time to be with us here today.
You know, I think that, youknow, what you've built is an
amazing company. The technologyreally works. I love the fact
that, you know, seeing andseeing using the technology,
(47:06):
that we're not in the best guestgame, we really can use AI to
really get to better outcomes.
And as we review the amazingoutputs, because you know, we're
ultimately making decisionsaround those outputs, we're
getting to a better product, andnot just like the marketers,
like not just the business ortheir brand likes, but the
audience's like, and it reallygives them value that compels
(47:26):
them to do something that reallygets them, you know, hopefully a
real well thought out and greatvaluable proposition that they
can consider, and whether it's amessage or a product. That's
what's really great about thetechnology. So so so Perry with
an egg. Thanks for being on theshow. Great, listen, thank you.
All right, awesome. So on thatnote, I just want to thank you
(47:50):
guys in the audience so much formaking time in your day to
listen to the industrialstrength Marketing Show. I just
hope you heard even one thingthat inspires you to make
marketing, technology and peopleeven more strength of your
business. So for more insightsfrom industrial marketers, or if
you just like to reach out to usvisit industrial strength
(48:13):
marketing.com You'll see us onthe podcast. You'll see us all
over the internet and online onYouTube. And we'd love to hear
from you thumbs up subs. Let'shave a great time. We'll catch
you next week.
I think what's happened to marketing in the
last 20 years is really sad.
Where marketers used to be, youknow, Don Draper types, sip your
whiskey in the corner office andcome up with crazy big ideas and
go hard with them. And now 90%of marketers time is spent in a
whole bunch of differentsoftware platforms and Excel
(00:24):
spreadsheets.
James Soto (00:34):
Welcome to the industrial strength Marketing
Show, we challenges leadersacross the supply chain to make
their people marketing andtechnology, their businesses a
strength of their business. Hi,I'm James Soto. I'm the founder
of industrial the family andstrategic marketing brands
trusted by leading organizationsaround the world. I'm sitting
down with innovators withindustrial marketers, and
(00:56):
technology leaders to talk abouttheir careers, their insights,
marketing, and technology thatmakes an impact on their
businesses. I'm here today withParry Malm, CEO of frasi. I've
been following crazy for a longtime, I've been speaking about
it at events around the countryin the manufacturing technology
space. And we're gonna reallyget into some great things
(01:17):
today. So Perry, welcome to theshow.
Perry Malm (01:21):
Yeah, thanks.
James Soto (01:22):
We're really glad to have you here. And I just
thought we just get right intoit, like, tell, tell our
audience a little bit about yourjourney as a marketer, as an
entrepreneur, you know, and, youknow, just as an expert, and
your little slice of the AIworld?
Perry Malm (01:38):
Well, I mean, like most people, I fell into it. I
think people who aspire to beentrepreneurs are constantly
chasing their, their tails. Andthe people who I think tend to
have the most success are peoplelike me who never intend on
doing it. But discover aproblem, experience a problem,
(01:59):
and then figure out ways tosolve that, that problem. The
fact that I'm I know a bunchabout AI is cool and stuff.
Doesn't get me a lot of friendsat parties and whatnot. But
ultimately, you know, the mostimportant thing is that we solve
customers problems, the factthat we use AI is cool. But the
most important thing is theproblems get solved.
James Soto (02:21):
Yeah. And, you know, that is that is the fundamental
reasons of starting a business.
So when you looked at startingthe organization, and as the
right place at the right time,what was the quintessential
problem in the market you'resolving for?
Perry Malm (02:35):
Yeah, so this goes back a number of years, when I
was a branch site marketer, Iwas living in Amsterdam at the
time, and we would send outmillions and millions of
messages to millions, millionsof people. And I used to always
look at it and be like, what isthe right message to send out
via I mean, back then this wasin 2007, or something. So it's
(02:57):
mostly on email and brochuresand direct mail and stuff. And
he's always go like, we'respending millions of euros.
Surely, we should spend a bit oftime to figure out what language
is good. So I started doing abunch of tested and tried to
build a bunch of models and onething led to another and before
long, I had the light bulbmoment, going, if I'm
(03:19):
experiencing this problem, thenprobably others are too. And it
turns out, I was right.
James Soto (03:25):
Yeah. And tell us a little bit about Phrasee. Like
one of the things we've heardfrom our audience, is they want
to hear unapologetically whatyou do. How does it help them?
Hey, James, you know, you know,I know, they're gonna say, have
Perry, tell us more about whatFreezy does? And, you know, what
are these amazing things thatyou're unlocking with your
technology?
Perry Malm (03:45):
All right, well, welcome to the shameless plugs
zone.
James Soto (03:50):
No, one apologetic,
Perry Malm (03:54):
The concept with Phrasee is very simple. We use
technology to not just createcontent, but to create good
content. Now, you may wonderwhat good means. And it means a
number of things. First of all,content that's generated needs
to be on brand, and it needs tobe trustworthy. It can't be
(04:15):
factually inaccurate. It alsocan't invert and people with the
need to continuously approvethings because they don't trust
the AI itself. And ultimately,that must perform strongly. So
that's ultimately what phrase hedoes. We we generate and
optimize and understand thelanguage that makes people
(04:38):
click. So the greatest andeasiest way to think about it is
that if you get emails, pushmessages, Facebook ads,
whatever, from companies likeGroupon or Walgreens are like
loads and loads of companies,then you've experienced raising
technology without knowing andyou've probably bought something
that's a direct result of thelanguage we not only generated
(04:59):
but optimized, and dependingupon what you bought, either
say, you're welcome, or I'msorry, I can't be in charge of
your own shopping habits.
James Soto (05:08):
Yeah, I have used an example from phrasing that was
like driving like a 30%improvement rate. And I think it
was subject lines for Facebookads, it was something about,
like, I used one of the examplesI think he showed, which was,
you know, one, you know, onelevel of copy that was around,
like, oh, take this vacation,and, you know, summer savings.
(05:31):
And then it's like sun sea sand,fun, you know, like, get in on
the action, bunch of emojis. Andwhat was amazing about it, and
this one felt magical to me,which was that the AI was
outperforming, you know,something a human row. And also
as was unlocking, like the useof emojis that like, didn't have
(05:54):
bias they didn't, there wasactually some thinking and
advantages to how it how itactually performed the job. And,
you know, we've I've done this,I've spoken with that example,
at a major technology event, andpeople actually kept picking the
AI is the better copy. So. Sothat's like, one of the things
that inspired me to reach out.
Because, you know, I think whenI kind of look at things, and
(06:19):
what inspired me, in thisepisode was really a quote from
Arthur C. Clarke. And it about,you know, it really kind of
dives a little bit into thisworld. And it was in his book
profiles of the future that hewrote any sufficiently advanced
technology is indistinguishablefrom magic. And I think that is,
(06:40):
I think, a lot of what we'reseeing out here. So when, in the
context of that, like, what doesthat make you think of when you
think about what we can now dowith the technology? You know,
in terms of AI for marketing forcontent, stuff that people can
see, they don't have to questionit, you know, they can actually
use it. You know, what, when youthink about like it being
(07:01):
distinguishable from magic,like, what does that, you know,
what do you think of in thecontext of your world and AI
marketing?
Perry Malm (07:09):
Yeah, I think people who view it as magic probably
don't understand how it actuallyworks. So it might be worth
spending a moment to explainwhat a lot of this stuff does.
Effectively, what AI is, is it'svery fast moving statistical
models. So there's sort of twomain schools of it. One is
(07:30):
prediction, or decisioning. AI.
And the second is generate ofAI. So the former prediction and
decisioning is effectively youknow, when you go into Excel,
and you got a scatterplot, andyou go like insert trendline.
That's effectively what the vastmajority of machine learning
systems are doing, except withmany more variables, many, many
more data points at a muchhigher speed. From a generative
(07:53):
standpoint, obviously, everybodyis, has been exposed to chat
GPT, which is like prettyimpressive technology. Use Cases
notwithstanding. But what iteffectively is, is it's a huge
statistical model that canpredict the next best word at
scale. So it can be given aprompt, and based upon that
(08:14):
prompt, it can then go, well,here's how I should structured
that answer using this huge sortof, like, multibillion parameter
vector space of, of languagetokens. So while it appears
magical to the layperson, whatit effectively is, is a huge
amount of really fast movingstatistics, which presents many
opportunities, but also manyrisks.
James Soto (08:38):
Yeah, yeah, for sure. And I think we're now
getting in the world where, youknow, this has been opened up to
pretty much everybody. And, youknow, we're at the point where
we're looking at like phraseengineering, because you can
even take that to the next leveland leveraging the technology,
and get even better and moresuccinct answers. And so, you
know, as you really look at thecore technology, and this is
(09:02):
kind of where our conversationstarted. Going. LinkedIn was,
you know, with that magic, thereare real issues, when we look at
that predictive models, thepatterns, the patterns, a
pattern apply, you know, applythe insights and generative AI.
You know, it comes to me likethose key impacts, we see
impacts on a number of areas,and I think there's like three
(09:24):
different areas. So, you know,we have, you know,
personalization, you know, andprivacy, we have the bias and
algorithms as an issue, whichreally can impact, you know, the
human element to potentiallynegatively impact outcomes. And
then there's the impact. I thinka lot of folks are feeling
(09:45):
today, whether you're a brandside marketer, or you're even in
an agency so to speak, or you'rean industrial business, it's
like we're seeing inmanufacturing, you know, big
displacement due to automationand AI is also going to have an
impact. So there's that thirdthing which Does that impact on
employment? And so withoutgoing, you know, mainstream,
(10:06):
it's just pretty amazing. So,you know, what, what do you see,
you know, like this, you know,chat GPT the gender of AI, or
even, you know, what's crazydoing that is so truly
remarkable, you know, that, youknow, you're gonna have to, you
know, really kind of come outwith, you know, real answers to
those questions. So, you know,so since it's been thrust on us
(10:29):
all, you know, what I'm, youknow, what are your concerns
regarding the ethical use of AI?
So, you know, just set, let'ssay, just started with
personalization and privacy, youknow, how does that apply to it,
you know, crazy is doing?
Perry Malm (10:45):
Yeah, well, we're, we're fortunate in some ways,
we're, we're a little bitdetached from many of the
challenges that that hyperpersonalization and big data
warehouses companies face, butjust sort of diving into the
whole school of generative AIand some of the the ethical
conundrums which one must face,particularly with large language
(11:07):
models, they're trained uponeffectively the entire Internet.
So what they did is they scrapedthe entire internet, more or
less, so all of Wikipedia andall of your old Tumblr, blogs,
and all this kind of stuff. Andthen they basically turned it
into a to a series of individualtokens, which are like words and
(11:27):
stuff, and then produced thishuge statistical model. So like,
from a technical standpoint,it's absolutely fascinating. But
the challenge is, it's trainedupon the entire Internet. And I
don't know if you've ever beenon the internet, but it's filled
with idiots, like 99% of thecontent on the internet has been
(11:51):
generated by an idiot. Peoplewith bias, people with wormhole
world views, all this sort ofcrazy stuff, all this sort of
embedded bias that we have, inUS only at a huge scale. So what
that means is that this model isinexorably biased. If you ask it
(12:13):
to produce content, targetingwomen, or targeting men, or if
you ask you to write jobdescriptions for specific job
roles, or if you ask it to, thisis a trick I pulled to generate
the pseudocode algorithm topredict somebody's credit score
(12:35):
based upon their gender andethnicity, then you can really
start to uncover the biases.
Some of them are overt, you knowabout race and gender and stuff
like that. But some of them areimplicit biases invisible to the
human eye. And I think that'sreally quite dangerous. You
know, the way that I phrase chatGPT is it's sort of like an
economist. They're alwaysconfidence, and sometimes,
(13:03):
right. So now we have thisauthoritative, authoritative AI
system that can producecompelling answers that only
sometimes are right. I don'tknow, that strikes me as being
quite existentially challengingfor truth for the internet. And
moreover, for marketers, whereif you're in trusting systems to
(13:26):
generate content for you, how doyou know if it's right? How do
you know if it's biased or not?
And ultimately, how do you knowif it's good?
James Soto (13:35):
Yeah, it's almost like freedom, you know, the we
created here in the States, wecreated a country that's based
on, you know, with that freedomis the assumption that we'll use
it responsibly. So, so if we'relooking at an internet of
idiots, so to speak, is it youknow, that we've really gone out
there, and, you know, theassumption is, when we're
scraping with, you know, gettingall that data with AI, that that
(13:58):
content was created, you know,with good with good thinking and
responsibility. And I think wecan argue there's a lot of bad
stuff out there, but there's alot of good too. But the
question is, what's, you know,how do we how do we scrutinize
that, you know, and I think inthis world, you know, that,
that, you know, we do have thisopportunity to personalize our
(14:20):
advertisements, you know,campaigns. So, it's not just
what we write, but it's actuallyhow we're doing that targeting
and personalization. And so theAI can tailor marketing messages
to be more relevant. And thennow we're talking about
influence and, and those things,you know, how to be effective by
looking at behavior,preferences, demographics, but,
(14:41):
you know, that relies to yourpoint on collecting a ton of
data and analyzing a vast amountof data. So in that context, you
know, we look at those privacyconcerns. So it's really
essential that we, we we reallybe transparent about Howard
collecting and using customerdata. And obviously we have to
(15:04):
give, you know, the userempowerment to opt out. But in
that context like how do yousee, let's solve for that in the
sense that, you know, how shouldbusinesses that are engaged in
AI share how they collectcustomer data? How do you see
that?
Perry Malm (15:21):
Yeah, I mean, it depends on what data is being
collected. And how sensitive thedata is. A good example is that
if somebody is collectingpersonal health data, then
presumably you need to be veryupfront about it. But if
somebody is collecting, like,when you bought dog food last, I
mean, you could certainly makearguments about how his personal
(15:46):
data, but also like, who reallycares, right? I think just being
honest and upfront with peopleis the key. I mean, it's when
companies, you know, pull outCambridge Analytica and collect
data and use it for spuriousmotives, which I think is is
(16:07):
morally wrong, but also leads tosocial ills. But like, if
you're, if your like, cornerstore, knows how many Red Bulls
you bought in the last week, whocares?
James Soto (16:21):
It's an inventory question. What do we need for
Joe? Or for Perry? You know, ifyou look, you know, and to the
extent that, you know, you haveyour business here, but if you
were to paint a picture of whatlike privacy looks like now, you
know, versus like, what, like,what paint a picture of it like
now versus let's say, five yearsfrom now, you know, what do you
(16:45):
see privacy looking like, youknow, given the rate of
evolution and innovation, andwhat's really going to be put in
market?
Perry Malm (16:52):
Yeah, I think it's gonna vary jurisdiction to
jurisdiction, and it'sultimately going to be
regulatory frameworks, which areput in. So I know that the the
EU already has some ratherstrong privacy laws with GDPR,
and all this kind of stuff. AndI think those are going to
strengthen. And we're probablygoing to see some level of
(17:16):
regulatory frameworks put in bythe EU to cover off generalized
AI systems like chat, GPT, andwhatnot. In the US, I think the
path will be quite different.
There will be regulations putin, but it's going to be driven
through case law, rather thanlegislators. So there's already
two interesting lawsuits, whichare either happening or have
(17:40):
been threatened in this space. Iknow that there's a class action
suit, which is either in theprocess of being filed or has
been filed versus GitHub copilot, claiming that there's
been copyright infringement of,of their source data that's
collected through open sourcecode and proprietary code, which
(18:01):
is held on GitHub, I don't knowwhat's going to happen with
that. The second one is, I wantto say stability, AI has been
threatened with a letter ofintent from Getty Images, who
claimed that their imagegeneration AI, illegally scrapes
Gettys imagery. So they'reclaiming copyright infringement.
(18:23):
Also, the EU tend to be more topdown, they have you know, it's a
it's a code of laws type thing,where they'll come in, and put
in a framework, which willprobably be clunky. But it'll
keep a bunch of bureaucrats inwork in Brussels, so good for
them in the US, is going to keepa bunch of lawyers, kids, in
private schools. So that'ssomething which we should feel
(18:46):
thankful about. And then in theUK, I don't know, we'll just do
something, it'll probably bewrong, and nobody cares.
James Soto (18:54):
Yeah, so it's the Wild Wild West, as we say, here.
And there's a lot happening.
And, you know, when we were whenwe were kind of going back and
forth on LinkedIn a little bit,you know, the kind of second
point here is that there's realethical concerns around the, you
know, the use of AI andmarketing. You know, the
potential for bias inmanufacturing was a huge, huge
(19:15):
skilled labor gap. So we do alot of recruitment marketing for
manufacturers and folksthroughout the supply chain. So,
you know, there's a lot ofpotential for bias in the
algorithms that power AI, andthe marketing of their, of it
with it. And these AI systemsare only as unbiased as the data
and the people who, you know,who've influenced the training.
(19:35):
And so if that data is, youknow, biased or skewed in some
way, you know, like, yeah, thethinking is, it'll be reflected.
So, like, suppose like, AI, likea system is trained on data,
like disproportionatelyrepresenting certain demographic
groups. You know, like we, youknow, we were obvious They will
(19:58):
say, you know, they'll make someinappropriate or wrong
assumptions. And so, ultimately,you know, we'll be unfairly, you
know, targeting people ordiscriminating, you know, either
intentionally or notintentionally. So, ultimately,
to that point, when you look atthe, you know, ethical use of
AI, looking at the biases, youknow, how are you looking at
(20:23):
mitigating that risk? Andbecause you're your front facing
with content and copy? So how dowe, you know, with marketers
having to really look atdiversity, equity, inclusion and
belonging? How are you ensuring,like, how do you see us ensuring
that we're being diverse andrepresentative, you know, in
actually what you are building?
Perry Malm (20:44):
Yeah, so first of all, what many people think is
that the source data is theproblem, and it sort of is, but
without that huge volume ofsource data, the large language
models wouldn't be possible inthe first place. So you're kind
of stuck with a huge amount ofdata that probably is biased and
over represented with middleclass plus white males, right?
(21:08):
But the thing is, you need thathuge volume, you need the
billions of parameters to getthe level of, of complexity that
large language model has. Sothen the question has to be, how
can you post process the output,and set rules and frameworks to
(21:28):
make sure that when it generatessomething that is either
hallucinating facts or isovertly bias that you can
overcome it. And that's reallywhere we've focused over the
years, I mean, we came up beforelargely with models existed, we
built our own languagegeneration system, which, which
is still in use, we'veintegrated a bunch of the large
(21:49):
language model stuff. But whatseparates us is that we've
already built all of these toolsthat can guide it at source, and
then also manicure it after thefact. Many systems do not have
this, and I would imagine, atopen AI, they're very actively
(22:11):
injecting directionality. Sowhen it first came out in
December, you could have agenerate all sorts of
reprehensible stuff, all theobvious stuff, it could, it
could, you could tell it to saythat, that Hitler was a super
nice guy or something completelywild like that. And they put in,
you know, a whole bunch of awhole bunch of these rules. The
(22:34):
challenge with that is thatsomebody is always going to
figure out a new way to be anasshole. So how do you keep up
with it? Fortunately, forFreezy, we're in a much work
contained environment, becausewe're working with with
marketers, the challenge becomesmore about not the ability to
generate content, that's justthe catch the challenges, then
(22:56):
the yards after the catch to getto the end zone. And that's
really where our focus is.
James Soto (23:01):
At. And so that's an iterative improvement of, of
what you're saying, as you'regetting insights, you know, in
real time, is that, is that howyou see? Exactly,
Perry Malm (23:11):
exactly, it's like, no, the ability to generate
language is effectively acommodity. Um, you got shut GPT,
which is effectively like amillion monkeys on a million
typewriters, and maybeeventually, they will write the
complete works of CharlesDickens. But you need to be able
(23:31):
to tell what's good for bed,before you put it out in the
wild. So that's really where ourfocus is, is to tell good from
bad in terms of brand, in termsof accuracy, in terms of
trustworthiness, and ultimately,in terms of performance?
James Soto (23:47):
I think that's a great point. It's, it's, if if
this world and the ability toput out content, you know, not
is no longer a differentiator,it's a capability, it's an
expectation, we can use thismachine assisted technology to
create copies of great content.
And that's, you know, availableto all, you know, how do you
win? How do you go up andbeyond, and to the extent that
(24:09):
you have your niche here, andyou're focused on marketing, you
know, because it has a differentstandard, everything from the
brand, to the audience, andpeople who are really trying to,
you know, consider thoseconstituents. And essentially,
you guys are connecting the dotsto make sure that you can drive
real value engagement,entertainment, you know,
education, to drive, you know,to inspire people to take
(24:33):
action. And then as thosesystems work, you can
continuously improve it toimprove outcomes. And that's the
beauty of AI. It's hard forhumans to replicate that, you
know, in terms of, you know,iteratively, you know,
continuously improving content.
And so, you know, so that, that,you know, there's a there's a
(24:53):
big issue here and kind of toget into the third issue here,
and curious to see how it's inpack With us, the third area
we've seen, you know, you know,outside of the bias side of it
is the impact on employment. Soas we look at, we look at
employers, right now inmanufacturing. There's about
(25:16):
800,000 open jobs in the USmanufacturing, manufacturing has
been going through a lot ofchange with automation, we've
seen displacement of jobs,because you can, you know, where
robot can do it better, youshould be having a robot where
you need humans to operate thoserobots and have created
opportunities. But I think thewriting's on the wall, the
(25:38):
combination of, of automation,you know, of AI, and you know,
the ability for that AI to getsmarter and smarter and smarter,
is really going to havedisplacement. And so when we
look at this in the context ofus as marketers, and how do you
really see that impact onemployment?
Perry Malm (26:02):
Yeah, I mean, certainly, if we go back to the
cotton mills in Lancashire, inthe Victorian era,
James Soto (26:12):
The Industrial Revolution, that's where it
started the spinning jenny's.
And they made those so eightpeople could do the work of one,
that's literally what we westart the Industrial Revolution
story on the car.
Perry Malm (26:22):
Yes, absolutely. And you had the Luddites, who came
into the factories, and theytried to bash the machines
going, it's going to disrupt ourway of life. And like, sure, you
know, a few of the artisans whohad you know, lose their front
room, had to retrain, find adifferent living, whatever. But
it created so many more jobsafterwards. And that's how these
(26:43):
things work. If you think aboutthe, like, what 1995 or
something when the internetreally started proliferating. A
lot of people in newspaperssaid, Oh, this is going to is
going to kill our jobs. And itdidn't like there's more people
writing news content now thanthere was back then. So like,
(27:06):
when people worry about specificjobs being lost, usually it's
because they have a vestedinterest in perpetuating the
status quo. What automation isgoing to do what AI is going to
do is it's just going to meanthat will lead to have millions
of people in the workforce,retrained and focused on doing
(27:30):
something else. I don't believethe scare mongering of people
saying that we're going to haveyou know, 20% structural
unemployment in the future. Itjust doesn't make sense. Like,
history doesn't indicate that'sgoing to happen. There will be
some job categories that AI willprobably threatened. And it's
(27:52):
going to be probably low leveladministrative work. And
hopefully, some high levelconsultancies
James Soto (28:03):
Is that a dig?
Perry Malm (28:06):
The thing is, what these large language models can
do, is they can aggregate let'ssay best practice, right? It
control across millions ofwebsites, and people who have
written about topic X pick anybusiness strategy or any
business conundrum. And it cangive you the vanilla ice cream
(28:28):
tried and tested answer. Right?
It's probably not the mosteffective answer, but it's not
going to be the worst answer.
It'll be the vanilla ice creamanswer. And that's what a lot of
consultants, do. They give thevanilla ice cream answer because
they've read the textbooks. Butno, you don't need them. So I
(28:49):
think it's going to be quiteinteresting in 1520 years time
to see what the role ofconsultancies are, there will
still be a role for them, butit's not going to be giving
pithy advice on businesschallenges. Yeah,
James Soto (29:06):
you know, so you hit two things here that I think are
really important. So one, youknow, you know, there's a
consultancy, and I'll talk aboutframeworks a little bit, but
then there's this, you know,first thing he said, on the
Industrial Revolution, so, so weshould look at this with
optimism, because when you lookat that displacement that had in
the early Industrial Revolution,there was humanity were like,
(29:27):
they needed X amount of men todo a job to harvest in an
agricultural society, you know,and you would work there and you
would work seasonally, a lot ofseasonal work, the world was a
lot different and what theIndustrial Revolution unlocked
and yes, it may have displacedfolks in you know, in the
textiles industry. But what itdid is a credit the factory in
(29:49):
the first time, you know, youwould literally be able to go to
a place and work year round. Youknow, what factories were
impacted by the weather and theconditions and, and so you can
literally go and it created, inessence, the middle class and a
place where you can go tohopefully get an apprenticeship
and work and, and have a careerand really create a
generational, you know, work. SoI do think that displacement
(30:13):
created the ability for folks towork like never before and know
here the Homestead Act in theUnited States took people and
they traveled, they traveledwest, right. And that failed.
But what happens is they cameback to the urbanization
movement, and they came back tothe city centers where the
factories were. And so that is avery, very true tale, a very
(30:35):
interesting story. So for everydisruption, you know, 100 years
ago, 100, a little over 100years ago, the US economy was
97%. Agricultural. And nowthat's down to three. What
displaced it the IndustrialRevolution, then what's this?
And then now we look at, youknow, that was a primary part of
(30:55):
our economy, and now it'shovering around 11%. What
displaced it, you know,information and technology, so
you've got to make your way ofliving life and doing business
obsolete before generationaltechnology and market forces,
you know, do do for sure. So Ido think there's this really
great opportunity moving forwardto really, you know, look at
(31:16):
these, you know, operaopportunities of the of the
future, because there's, thereis going to be change, for sure.
And for people to like, hold onto those constituencies and, you
know, have those worries, youknow, we have to, you know,
understand where we're going tobe and something I heard, I
(31:37):
think Gary Vee said it was that,you know, 10 years ago, I was
saying like, you know, if you'rea yellow page ad salesman, you
better be looking at what Googleand Yahoo are doing. You know,
when they're first startingYahoo, formed in 94, Google in
98, and I worked for anindustrial directory, which is
the biggest print industrialdirectory, and they wanted me to
(31:58):
help them really kind of evolvetheir team and their go to
market model, especially theirsales organization, to be
digital leaning. So we're allgoing through these disruptions.
And so, you know, as we, youknow, look at these same tools
and this disruption, how do yousee that playing out, you know,
for marketers, you know, becausethere's just so many things, you
(32:20):
know, yes, there's aredundancies? Yes, there's
things that can be automated,but like, where's the writing on
the wall? What areas?
Perry Malm (32:29):
Yeah, you know, I came up as a marketer. And I
think what's happened tomarketing in the last 20 years,
is really sad. Where marketersused to be, you know, Don Draper
types, sip your whiskey in thecorner office and come up with
crazy big ideas and go hard withthem. And now 90% of marketers
(32:54):
time is spent in a whole bunchof different software platforms
and Excel spreadsheets. Andthey're expected to, to just
execute, you know, they're justoperators. That's all they are.
There's very little thinkingdone in marketing. You know,
there was a four P's right, likeproduct, price place promotion,
(33:18):
I think it is marketing. No,it's promotion, and product,
price and place are sort ofdivvied off elsewhere. And I
think it's a real shame, becausegood marketers can contribute a
huge amount of value to theentire business into the
products, which are being builtinto, you know, how you actually
(33:42):
distribute it, all this kind ofstuff. And yet know the majority
of marketers time is like, likedeciding, oh, should we up our
bid on this Google Ad by 20cents or something? It's just,
it's taken a very interestingjob, and has made it very
(34:04):
boring. And I think that's areal shame. So hopefully, what
some of these new technologieswill do is get rid of a bunch of
the boring stuff, automate, oweda whole bunch of the boring
stuff, and make marketing.
Interesting again, I love
James Soto (34:20):
it. So you're talking about the manufacturing
experience, right? We have toautomate out, remove repetitive
work and go to high creative,high cognitive skills. You know,
like, that's what we want to doas humans, like, everyone. Some
people say they're creatives.
And those people are notcreatives that straightener,
there's data and studies thatsay, like, No, we're trained out
of that every kid's creative,they play with their kids, and
(34:42):
you can recapture thatcreativity. And I do think, you
know, that's what's happening. Ibelieve the writing's on the
wall, that if there is a processor something you repeat, those
things need to be you know,really considered for candidates
for these AI you know, Toempower technologies related to
marketing sales data, you know,the whole mix there. And that is
(35:04):
going to be framed to folks, butit's true. How much of our time
are we spending doing thingsthat, you know, only humans can
do. And that's the imagining,you know, our head of studios
says, I need time to dream,James, where's our dream time,
and whether your draw you, youknow, you're Draper or you're,
(35:25):
you know, you're just the, youknow that you're the marketing,
art director somewhere, youreally want to be able to do
those things. Because thesetechnologies, if we use them,
right, can really free up andcapture our imagination. You
know, I'd hate to see a worldwhere AI is being creative for
us completely. And we're devoidand detached from the process.
(35:49):
And then lastly, you mentionedthe displacement of
consultancies, I find that veryinteresting, because I think
that when you look at a lot ofconsultancies, like McKinsey and
others and Deloitte, they haveframeworks and students study,
they're frameworks they use, andthey're well vetted, they're
very robust. And, you know, I dosee that same issue happening
(36:12):
with consultancies, where youcan use these, you know, AI to
really disseminate like reallygood frameworks that you can
make your own. And that you canwalk through whether you're
going to go to market strategy,you're doing a transformation
effort, you're looking atBusiness Model Generation, or
you know, positioning, there'sframeworks for these things, and
(36:32):
that I believe, through thesetechnologies and the ability to
create these things. You canalso co pilot that with AI and
Bureau research, to reallycreate great ways that you can
get humans because frameworksyou have to put a human through
so a business can make decisionsabout it to really impact how
that will affect audiences ormarketers can do the same. We
(36:53):
all have frameworks, and I thinkthe the age of the consultancy,
we all want to be paid based onour thinking, not just making
things or doing things or doingsomething by the hour. It's
really about where do we bringthat strategy, that creativity,
that dreaming to the to theforefront? And I think AI could
be opening up a whole new set ofpossibilities. And, you know, to
(37:14):
your point, Perry, I just wantto key in on, I think you just
have some really great points isthat if we're going to resist
this, you know, from thestandpoint of our our biases and
holding on to the past, I thinkwe're missing the point, we're
going through yet another set ofevolutionary forces that we
(37:34):
really have to contend with. Soyeah, totally Yeah, to,
Perry Malm (37:42):
like, people can be Luddites, and they can pretend
it's not happening and they canfight it. That's fine. But like,
what happens the Luddites?
James Soto (37:55):
Know what to believe and know what a Luddite is. I
actually write about it, becausethat's one of the things I
talked about in teaching folksabout the Industrial Revolution.
And then like how that form the,you know, the unions and the
weekend was great it because ofthe fact Yeah, that's a
byproduct of the IndustrialRevolution. There wasn't even a
weekend such a thing of theweekend. So, you know, I think
(38:17):
there's not going to be a lot ofLuddites that, you know, you
know, would want to, you know,make a bigger workweek or, but
yeah, there's, there's a lot,there's a lot for us to upon us
now. And COVID had a huge impacton that.
Perry Malm (38:31):
Like, like, there's even a world. Yeah, I'm not sure
if this is gonna happen or not.
I learned a long time ago tolike, not make bold predictions,
because the world isunpredictable. And that's one of
the wonderful things about thehuman experience, right. But
there is a world where we becomevery homogenized in the world
(38:53):
experience becomes very boring.
That's one of my fears. So like,no, like back in 1998, my
brother and I backpacked aroundEurope for three months. And we
were really surprised anddelighted by so much we saw like
we saw all sorts of new thingscrazy things like her different
(39:16):
languages even use differentcurrencies back then it was
before the Euro was adopted, theTreaty of Maastricht. But no, if
you go on holiday, first of all,the experience is very
homogenized, because every cityhas a McDonald's and a Starbucks
and an Irish Pub. But you alsolike research it before you go
(39:39):
away, so you already know whatyou're going to see before you
go. And I think one of the riskswith AI, is it it makes the
entire human experience likethat, where we, we we optimize
everything in automate, automateeverything in our lives. which
(40:00):
means that the world becomes onebig blob. And I think that we as
people, as humans, and asmembers of society, need to
ensure that there's still thingsthat surprise and delight us. If
we ever get to the point wherewe're never surprised or
delighted. You may as well justjump off a bridge.
James Soto (40:25):
Yeah. And to the extent that something is so
predictive, prescriptive andpersonalized every time. And
that becomes our world, right?
We expect that and we open upNetflix writes, predictive,
prescriptive and personalized.
Where are those things that younever would have ever thought to
do? Or run into? Wherespontaneity? Where's that
(40:46):
serendipity in our lives? And tothe extent that this becomes an
overlay? It could be soulcrushing? I completely agree
with that. And I think as welook at the ethical use of AI,
you know, personalization andprivacy, those are big issues.
But I think the net outcome ishow does it affect the human
condition? You know, I think aswe look at, you know, you know,
(41:09):
even the workforce, as you know,as you look at, you know, tell
us, like, how big is your team?
By the way, it crazy how many,how many humans do you got,
Perry Malm (41:18):
you have roughly about 100 people? So 100
James Soto (41:22):
people, you know, do you have this conversation on
how this, all these things aregoing to impact them? And, you
know, is it the responsibilityof your organization to really
kind of consider that, andmitigate potentially some of
these impacts on them?
Perry Malm (41:39):
Yes, and no. So, ultimately, you know, a
company's objective andfiduciary duty is to its
shareholders, which is a verydepressing way to think about
business, but that's ultimatelythe truth. So like, with that
said, we do have a duty of careto our employees, also, we need
(42:03):
to make sure that what we'redoing, Will, we'll advance them
in their careers and expose themto things that they that they
can carry forward in 10 yearstime and when, you know, maybe
Freezy will be will be like, onthe NASDAQ or maybe maybe will
just cease to exist, you neverknow what's gonna happen in 10
(42:23):
years. But like, do we do wetalk to them deeply about like,
how AI is gonna affect theirjobs and whatnot? No, because we
we live and breathe it. Andpeople don't join crazy with the
expectation that AI is not goingto affect them. They join crazy
because they know that AI willaffect them. So I'm probably not
(42:46):
well placed to, to answer thatquestion. Because we live and
breathe this stuff every day.
James Soto (42:52):
It let me let me ask you this. So we have a lot of
industrial marketers, we have alot of executives of, you know,
industrial companies that, youknow, in our sphere and our
audience, these are marketers,sellers, people on the digital
side of the organization, tryingto bring together marketing
sales, customers experience, youknow, what would you say to
marketers, who are really tryingto grasp with what's the role,
(43:15):
you know, machine assistedwhatever, you know, in their
roles, like, how do they? Whatwould you have to say to them in
terms of how they need to reallyreconcile AI? What, if anything,
should they be doing right now?
Perry Malm (43:30):
I would say, figure out the things that you like
doing the least. And then use AIto do those things. You know, I
mean, a great example is likewriting loads and loads of
marketing content over and overand over is quite arduous.
Especially the leg short stuff,like subject lines, right? Who
(43:54):
wants to sit there and writesubject lines every day, or
Google ads or feed like Facebookads, when you're doing this
stuff at scale? Who wants to dothat? It sucks. Like it's not,
this is not what Don Draperdoes, you know, he doesn't sit
there on Madison Avenue writinga subject line. Nobody wants to
do that. So automate it, that'sone phrase he does. But if you
(44:18):
like doing some stuff, then keepdoing it. I think, you know,
there's so many different toolsand so many different
opportunities to automate andoptimize every part of the value
chain in business, andparticularly in marketing. So I
think marketers should be alittle bit selfish. Do the stuff
(44:40):
you want to do. And then Viceoffered to do the stuff you
don't want to do? I mean, what'sthe downside?
James Soto (44:47):
Yeah, yeah. Minding that balance that you were
talking about before, whereyou're not just over automating
it and your job becomes soullessand let it free you. But if it's
not freeing you, you know, thenYou know, what is it? You know?
What is it doing for you ifyou're caught up, you know, in
the minutia of the technology, Ithink the chief martech a list
(45:09):
of all the marketing relatedsoftware technologies, I've seen
it grow from 500 to 1000. Andyou know, just last year reach
10,000, marketing, automation.
And now obviously, growing AIrelated technologies, it's so
massive, it's, it's almostimpossible to keep up with all
the distinct categories oftechnologies that are out there.
(45:31):
So as we wrap up, you know, getgive us that yet again, plug
for, you know, for whatinformation listeners can use to
follow learn more about Freezy.
You know, what are you guys upto? What is the cool events?
But, you know, I know everyone'sgonna be dying to learn more
about you. And frankly, how dowe how do we connect with
Perry Malm (45:54):
you? That's super easy. frasi.co. And I'm the only
Perry mallam in the world.
Weirdly, I looked on LinkedInthe other day, and there's a
Perry Milne spelled p e r ry, inNorth Dakota. So I don't know
how I feel about that I feellike 80% Less unique than I did
before. But Perry with a male,I'm the only one. And as far as
(46:20):
freezing goes, look like, if theproblem you've got is you need
to produce highly effectivecontent at scale, then Freezy is
a good solution for you. And weshould talk. If that's not your
problem, we can still talk Ijust want to try to sell you
anything.
James Soto (46:42):
Yeah. Well, that's great. And, you know, thank you
so much for making, you know,time to be with us here today.
You know, I think that, youknow, what you've built is an
amazing company. The technologyreally works. I love the fact
that, you know, seeing andseeing using the technology,
(47:06):
that we're not in the best guestgame, we really can use AI to
really get to better outcomes.
And as we review the amazingoutputs, because you know, we're
ultimately making decisionsaround those outputs, we're
getting to a better product, andnot just like the marketers,
like not just the business ortheir brand likes, but the
audience's like, and it reallygives them value that compels
(47:26):
them to do something that reallygets them, you know, hopefully a
real well thought out and greatvaluable proposition that they
can consider, and whether it's amessage or a product. That's
what's really great about thetechnology. So so so Perry with
an egg. Thanks for being on theshow. Great, listen, thank you.
All right, awesome. So on thatnote, I just want to thank you
(47:50):
guys in the audience so much formaking time in your day to
listen to the industrialstrength Marketing Show. I just
hope you heard even one thingthat inspires you to make
marketing, technology and peopleeven more strength of your
business. So for more insightsfrom industrial marketers, or if
you just like to reach out to usvisit industrial strength
(48:13):
marketing.com You'll see us onthe podcast. You'll see us all
over the internet and online onYouTube. And we'd love to hear
from you thumbs up subs. Let'shave a great time. We'll catch
you next week.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy And Charlamagne Tha God!
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.