August 3, 2025 • 30 mins
In this episode of The Digital Marketing Podcast, Daniel Rowles and Ciaran Rogers explore one of the most transformative shifts in digital marketing: the arrival of ChatGPT Agent and the rise of agent-based AI. No longer just about generating content, AI is now taking action on our behalf, logging into websites, filling out forms, analysing data, creating content, and even posting to platforms like LinkedIn. It’s here now, and it’s changing how marketing teams work from the ground up. Daniel shares his firsthand experiments with ChatGPT Agent, from connecting it to Canva to create and publish carousel posts, to testing its ability to analyse his entire LinkedIn history and recommend strategies for future content. While the possibilities are jaw-dropping, the hosts also explore the risks, security concerns, talent pipeline challenges, and the long-term implications for marketing roles. In this episode, you’ll learn: What agent-based AI really is and why it’s different from traditional AI chat tools How ChatGPT Agent combines deep research, web browsing, and autonomous task execution Real-world test cases, from generating content to posting directly on LinkedIn Why prompt quality and structure are more important than ever The security and compliance risks of connecting agents to business systems How agent-driven automation could replace many junior marketing tasks Why agility and learning culture are now critical for businesses to survive disruption The ethical and strategic questions marketers need to face around job design and human oversight The discussion goes beyond technology to look at the human side of digital marketing. Daniel and Ciaran debate the future of junior roles, how companies can avoid gutting their talent pipelines, and why brands that lean into humanity, creativity, and trust may stand out in an AI-driven landscape. Key Takeaways: ChatGPT Agent is a genuine step change, going from AI that suggests, to AI that does. Marketing teams can gain huge efficiency, but must double-check outputs and guard against overreliance. Security is now a critical concern, as agents gain access to business-critical systems. Junior marketing roles are at risk, companies must balance efficiency with long-term talent growth. Building an agile learning culture is the best defence against AI disruption. 📥 Access the show notes, tools, and links at: 📩 Join the free newsletter for access to monthly live sessions:
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Ciaran (00:00):
Welcome to the Digital Marketing Podcast, brought to
you by target internet.com.
My name is Kieran Rogers.
Daniel Rowles (00:06):
And I'm Daniel Rolls.
Ciaran (00:08):
And today, Daniel, we are discussing AI agents.
Daniel Rowles (00:20):
Okay, so let's just define what an AI agent
is, talk about the evolution andthen talk about why we are here.
This very specific period in time.
Ciaran (00:30):
Why so profound Yeah.
Daniel Rowles (00:32):
Well, it's gonna get a bit profound in this one actually.
So, definition of a agen AI oran agent-based AI is it uses
sophisticated reasoning anditerative planning to autonomously
solve complex multi-step problems.
Okay?
So this started when chattwo BT and Google Gemini.
Release deep research modes.
I mean, it started way beforethat, but that's what's kind
(00:54):
of brought it to everyone.
So deep research mode.
If you're not familiar, you goin, you've in chat, for example.
You've got your little plus buttonas you have in Gemini, and you
could select deep research mode.
You'd give it a task, maybe to do someresearch and it would go off and ask
you a series of questions, alright,what target market are we talking about?
What's the time periodyou're interested in?
Et cetera, et cetera.
(01:15):
And then it would go off andcomplete a task for you autonomously.
Now the thing is, that's not quick.
It might take 10, 15, 20, 30 minutes to gooff and do that thing, but that was their
first kind of move into agent-based ai.
Then what you also had was chatt,BT were building this thing called
operator, and operator was theability of chatt BT to use a browser.
(01:39):
Not to search in a browser, but to, youknow, log into a website to go through to
fill forms and all those kind of things.
What's happened and we've just done anews episode on this is that chat two.
BT released their agent, so chat two.
BT agent and chat.
Two.
BT agent suddenly takes theresearch, the ability to use a
browser to manipulate documents.
(02:00):
Gives you this new thing thatcan go off and do stuff for you.
It is a game changer.
Now, I should say I got myself intotrouble with this, so I got hold of this
a bit earlier, so it's, by the time thiscomes out, it would've been out for about
10 days chat agent and I got, hold itreally early on and I started playing with
it and I started creating lots of testcases and I tested the heck out of it.
(02:24):
I then went, look, what about I wantyou to research content, create content,
log into my profile on LinkedIn andpost it how I you to go into all my
LinkedIn posts, and I want you toanalyze all of 'em and tell me what
content I should do in the future.
How about I use the Canva connector inChat two piti, if you're not familiar,
there are lots of connectors in chatthat can connect to external tools.
(02:47):
And I was able to connect itto my Canva account and I could
do this through the agent.
Either I could get it to log in to myCanva account, take a load of content it
created, and create a load of carouselposts for me, and then take those carousel
posts and put them into LinkedIn and.
Ciaran (03:03):
What could possibly go wrong?
Daniel Rowles (03:05):
Well, it didn't, well, the thing was nothing went wrong.
It did it.
It did a really good job.
But there was a risk.
Right.
Now, what I should say, with allof this AI agent stuff, there
are human intervention points.
So for example, when it goes to myCanva account, it doesn't own my
password, so I have to take over.
So I click the button, I take overthe browser for a moment, and I put
my password and my username in, and itlogs in and it carries on my behalf.
(03:28):
So let's just take a step back from that.
How this LinkedIn post got meinto trouble was that I said,
this is game changing technology.
I reckon I can replace a marketingexec with seven decent prompts.
Okay.
And my, what I went on to say was thatmeans you'll have huge cost savings.
You won't, you know, theseagents don't need holidays.
They don't want bonus pay.
(03:49):
Their dogs don't get sickand all this kinda stuff.
And then I did add, which is thepeople, the bit that no one seemed
to read was that you'll also haveno talent pipeline 'cause you won't
have any junior people anymore.
You're gonna have no diversityof think in your organization.
You won't have any youngpeople potentially, you won't
have a working culture and.
That's gonna have quite abig impact on businesses.
What do you think businesseswill do about this now?
(04:11):
Classic social media.
Nobody read the bottom bit tosee that I was being a little bit
facetious about it and was saying,
Ciaran (04:16):
You had
Daniel Rowles (04:16):
yeah,
Ciaran (04:17):
we don't need,
Daniel Rowles (04:18):
we don't need market execs anymore.
Yeah.
And my favorite one was that commentsaid, you clearly didn't really test this.
And it just like, you know, myprofessional integrity was called
out to quite a large extent.
But anyway I, what I normally dowhen I get angry about a social
media post is I refuse to respondto it until the next day, and I've
normally calmed down by the next day.
Ciaran (04:35):
Always
Daniel Rowles (04:36):
so, so I went through and
Ciaran (04:37):
edge.
Daniel Rowles (04:38):
very politely pointed out to everyone that, actually,
I wasn't saying it was a hundredpercent, but let's, so, okay.
There's a couple ofthings to pick out there.
First of all.
I can now get this to do things Icould never do before, like log into my
LinkedIn and analyze all of my posts.
I don't have to export it first.
So there's phenomenallyuseful things to do.
I can get it to log in andpost social media stuff for me.
(05:00):
Is that a good idea?
Probably not.
It's kind of risky.
But I can get it to log inand buy something for me.
So I don't think we're very far away fromthe first legal case of saying chat PT
spent all my money and I told it not to.
So, which is, yeah, there are lots ofsafety features in place, but my point.
This is game changing because we,this is the first iteration and we
(05:20):
know how quickly this stuff's moving.
Ciaran (05:22):
Yeah.
Daniel Rowles (05:22):
In six months, this will be radically better than
it is now because it's learning.
Okay.
And therefore you will be able to verysimply say, okay, what do I do every day?
What does my marketing exec do every day?
Let's turn that into a reallygood step-by-step, prompt.
Go to this website, do this thing,download this, go here, do that,
(05:43):
and then just take my role andbreak it down into that series of
tasks, and I can automate those.
Now, we've had robotic processautomation for a while, but it falls
over because of the fact that actually.
You know, when the website changes,it can no longer handle it.
Well, actually now the AIcan kind of overcome that.
The first time you use, by the way,chat BT agent and you sit there
(06:05):
watching it, browsing your websiteand telling you what it's doing.
Oh, well that didn't work.
I'm gonna try this.
Oh no, that didn't work.
I'm gonna do this.
It's quite a mind blowing experience.
I thought, I felt it was like,I'd really, you feel like you
experienced something very new.
It feels very science fictionally thefirst time you kind of do it as well.
Ciaran (06:20):
We, well, we were chatting on text, weren't we last week?
Have you tried it?
I'm like, no.
And you were like, well,have you got access to it?
I'm like, I dunno.
And I went in and looked I've becomea bit obsessed with this, but it's the
little icon that they shoved in there,kind of looks like a little robo print.
So I'm seeing it.
A wonky mushroom,
Daniel Rowles (06:38):
Righty muscle.
I'm sure that wasn't theirintent from a design perspective.
Ciaran (06:41):
I know it's, well, I think what they're trying to, like the icons, trying
to like illustrate the point that it canopen a web browser and click and explore.
Like that's what it
Daniel Rowles (06:50):
right.
Ciaran (06:50):
like to me.
But it does look like a bit like asquare foot, like paw print to me.
Like if
Daniel Rowles (06:55):
So to,
Ciaran (06:56):
pauses
Daniel Rowles (06:57):
to me it's just a screen with a pointer clicking on it, but Okay.
Ciaran (07:00):
it's a pore look, it's a paw.
You can sit, it's a pore.
Right?
Daniel Rowles (07:04):
No,
Ciaran (07:04):
if you agree with me right in.
Daniel Rowles (07:06):
it's a screen, a pointer and some like little dashes that,
that are showing movement, I think.
But anyway,
Ciaran (07:12):
it's
Daniel Rowles (07:12):
let's,
Ciaran (07:13):
I,
Daniel Rowles (07:14):
it's.
Ciaran (07:14):
What's clever about it is when they put the word agent
on it, it sounds really cool.
But is this really that differentfrom Gemini's Deep research?
Daniel Rowles (07:23):
It is wildly different because deep research
can go off and do things.
It can look at things and readthings, but this can actually fill
forms in navigate through a website.
You know, it's not justreading the text of the page.
It can interact with the page.
So if we go beyond just usinga web browser, I mean, that's
the main functionality.
It can use a web browsernow like a human can.
(07:43):
But also if you give it like it usedto be, you'd give it a spreadsheet,
it would analyze it, and then itwould give you some sort of output.
Now you can give it the spreadsheet andit can put the answers in the spreadsheet
So it's got the abilityto utilize things that it.
We'll put it in the show noteswhere target.com/podcast, the
functionality it has at the moment.
But the point being is that beforeI could look stuff up, I couldn't
(08:07):
actually interactively do things.
Now I can go to a logged in website.
I can say to it, go and find me arecipe for this many people that
this many dietary requirements,and then log in to my Tesco
account and buy that stuff for me.
And it's able to go through and do that.
And as and if it doesn't got aparticular product in stock, makes
some alternative options for me.
And there'll be human points.
I have to help it log in.
(08:28):
I have to say yes to purchaseand things like that.
But the, what's really interesting isall the marketing videos they created,
they gave the agent a task and thenthey shut their laptop and walked away.
There was a very big push in the marketingand then what happens with chatt PT is
you get a notification on your mobiledevice when that task is finished.
So it's like it's doing stuff for you.
(08:48):
It is your agent, it's going offand completing tasks for you.
So, and it, you're absolutely right,it is deep research mode combined
with the ability to use a browserand some other stuff as well.
So it's got more tools, it isgot more tools at it's disposable
and therefore it's disabled tochoose which tools it's using
Ciaran (09:03):
It's quite fun to watch it while it works.
Daniel Rowles (09:06):
Massively.
Ciaran (09:07):
a and actually I could see it was at one, think I'm, I
shared this in a previous episode,but I bought a laptop using the,
a agent feature with in chat GPT.
And actually at one point I couldsee it was writing a whole bunch
of code to compare some of thedata that it scraped and exhausted
and do something clever with it.
But you have to watch it.
(09:27):
It happens quite quickly.
Daniel Rowles (09:29):
Well, lemme give you another example.
This is amazing.
Ciaran (09:31):
Yeah.
Daniel Rowles (09:32):
It basically, I asked, I gave it an image and I'd done this
three months ago and it was an imageof a, with 99 dog faces on one picture.
And I said, how many dogs in the picture?
And three months going,sorry, I can't answer it.
I can't work it out.
It's too confusing.
This time it went right.
I'll go through and I will use.
An algorithm for facial recognitionand then it says, oh, it doesn't
work because the facial recognitionis set up for human faces.
(09:54):
But what I'm gonna do is quicklywrite some code that uses it
just to count the number of eyes.
'cause I know that dogs have gottwo eyes and it counts the number
of eyes and it got the answer right.
So essentially what we've got issomething now that has a number
of tools, it can do web scraping,like looking stuff off the web.
It can go to its large language model, itcan write code, it can use a web browser.
It's using those tools to make a decision.
(10:16):
What's the best way ofsolving this problem?
Now it's not.
Ciaran (10:18):
Game changer for summer fates when you need to know how many
jelly beans are in the kilner jar.
Daniel Rowles (10:22):
So, you know, the ability to do things quickly, effectively, in
a more creative way is certainly there.
What's interesting is that thisreplaces tasks that we do every day now.
It's not that I'm interactingwith it backwards and forwards.
It's like, right, go and do this, andit'll go off and do it, and I get back.
The outcome is what I need it to be.
It's not giving me somethingI can then do something with.
(10:43):
It's actually doing thething for me in many cases.
Ciaran (10:45):
So my experience with this is yes, and it gives some very impressive results,
you sure as hell have to double check.
Sense, check what it's done, or you'regonna get egg on your face pretty quick.
And I only know that because I'veproduced some true, truly wonderful
things using this, that I'm very proud of.
I know I, I'm very fortunate I workwith somebody who really, their head
(11:10):
is totally in detail and you make anyclaim and they'll double check it all.
'cause that's how they, that'show their brain's wired.
And it's brilliant.
'cause actually it means you,you're forced to sense check stuff.
Like I've
Daniel Rowles (11:20):
Right.
Ciaran (11:21):
land mine too many times.
And I'm not having to kirinDid you absolutely check this?
Like, remember when we were firstplaying around with chat GBT and
it was given, we were asking it forlatest news in SEO and it was given
us all sorts of absolute rubbish.
This is
Daniel Rowles (11:34):
Yeah.
Ciaran (11:35):
back.
Right?
But you know, I still think that.
That element is there, especially whenyou're playing around with different
models you've not used before and youhave to sort of sense check it and
work out where are the boundaries,what's it good at and what isn't it.
So that was gonna be my questionto you, like if people are
wanting to test this out, what
Daniel Rowles (11:50):
Yeah,
Ciaran (11:52):
would you recommend people, because I know if you haven't
used it before, you're definitelygonna wanna play with this.
What would you recommend Daniel People do?
Daniel Rowles (11:57):
well this, that's a really good question 'cause it leads us
into the prompting piece, which is thatI think a lot of people use this first
of all and went, oh, it's all right.
But it's not great because likethey did like a one line prompt,
Go to my LinkedIn.
Tell me what my best posts are.
Okay, so it'll go, but are yourbest poster ones that got the
most engagement are the onesthat drove your desired outcomes?
(12:19):
Are they ones that got the most views?
You know, you haven't reallydefined what best means,
But it will be able to log into yourLinkedIn and then it will go to each
of your posts and it'll try and getthe stats from and download them for
you and it'll, so actually that's thekind of thing you can get it to do.
That's, I would try it out withthose, kind of get it to do something,
you have to log in to do something.
So to a social media profile, toa shopping website, to, you know,
(12:41):
something like go to Amazon to buyyou something, whatever it might be.
Ciaran (12:44):
so I haven't gone that far with it.
But how did, how do you do that?
Do you give it thelogin details within the
Daniel Rowles (12:49):
No.
So you just give it, you say goto this website and do this thing.
And then when it gets there, it will,it'll open up its browser and it will
get to the login page and go, okay,well I think I need you to log in and
it'll give you a button to take over.
Ciaran (12:59):
Right.
Daniel Rowles (13:00):
then you take over its browser, you put your login details,
and then you hand the browser back.
And at that point.
Then does it, but it will rememberthat's what you've gotta be careful of.
So now it knows my login for LinkedIn,and if I ask it to go LinkedIn
again, it's able to repeat that step.
So, and depending on whichtool you're using, but what,
Ciaran (13:17):
Terrifying because far as I'm aware, we don't have two
factor authentication on our.
Language learning model logins.
Daniel Rowles (13:26):
no.
So what will happen?
That's why you want twofactor authentication.
Switch on everything.
Ciaran (13:30):
Yeah.
Daniel Rowles (13:31):
Because you can switch it on with CHATT P and
stuff like that now as well.
Ciaran (13:34):
Yeah.
Daniel Rowles (13:35):
So you can have two factor, because what you want it on chatt pt, and
you want it on all of your tools becausethen if someone, if you accidentally
are doing something with your agent, orif you know somebody's accessing your
agent, you want the security of knowingon your mobile device, oh, that's not me.
I don't want 'em to log in.
So yeah, security becomes evenmore important before there
are some mess up with this.
But what I'd say is that.
(13:57):
I would look at what you do every day inyour day-to-day role, and I would think
about which of those tasks are repeatableand which of those tasks are purely admin.
Right, as opposed to creative thinking.
So I know when we create this podcast,there's a number of steps I go through.
I take the podcast, I put it intoDescrip, I edit it in Descrip.
(14:17):
That is not something I'm gonna get anagent to do because there's loads of
nuance and subtlety and what's that?
Then I'm gonna go throughand I'm gonna create the.
Show notes that go onto the website.
Well, actually I can get AI to do that'cause I've trained AI to go through,
take the transcription, put out thekey points, structure it in a certain
way, find any links, put those up.
(14:37):
So actually that's astep I could go through.
Then from that, I'vegotta log into my website.
I've gotta put that onto a webpage,create the page, fill in all the
standard bits I've gotta then go throughinto Libsin and publish it in Libsin.
There's a lot of stuff in there.
I can get the agent to dofor me automatically, but
it's so business critical.
There's some bits I'm not gonna do.
I'm not gonna let it publish it toLibsyn without double checking it
(14:59):
and all those kind of things as well.
It's not a hundred percent at the moment.
It will do a task, get alittle bit confused sometimes,
and then just kinda stop.
Because it's not, it's got a bitconcerned about what's going on
or what have you as well, but thiswill iterate and this will change.
Now the important thing from thisis that one, this is gonna help you
be more effective and efficient.
So that's great.
Hopefully it gives you more timeto focus on the creative, the
(15:21):
strategic, the critical lateralthinking, whatever it might be.
But what I know it's gonna mean isthat as this gets more advanced,
what's gonna happen to junior roles?
I was being a little bit facetious, right?
You know, it will replace a marketingexec, but if it can replace 80% of what a
marketing exec does, do you think peopleare gonna still gonna keep recruiting
(15:43):
and everyone's going, yeah, 'causeyou know you need a talent pipeline.
You do.
But we know what companies that are likebig companies about efficiency already.
There's a report in theTimes, well ads, show notes.
You can see I'm not making this up.
Junior roles across the uk andthis is, there's a global pitch
that goes with it as well.
60% drop in recruitment.
For a couple of reasons.
(16:03):
One, 'cause the stuff that's goingin the economy, but also because the
fact they're going, well actuallyI might be able to get an agent.
I might be able to get AI to do this.
Do I need that role at the moment?
Where do I need to focus my efforts?
So people are just takinga bit of a step back.
It's not necessarily replacingthem, but it is making people
think about recruitment.
Take this six months down theroad, this will radically change.
(16:23):
Actually a lot of the stuff thatwe've got junior roles doing, we'll be
able to do and be able to do it withthese agentic tools really easily.
So it's a game changer for the workplace,and it's gonna change all of our roles.
It is gonna replace, I'vebeen preaching for ages.
Oh, it's not gonna replace roles.
And you know, it's onlycertain roles it'll replace.
It's mostly, it's gonna change roles.
I think this is gonna replace lotsof things that people are doing
Ciaran (16:45):
I just, you've just got me thinking about the
security of this though.
Daniel Rowles (16:50):
right.
Ciaran (16:50):
Like here we've got a back door.
Which could any, anybody in your teambe hooking it up to all your business
critical systems, giving it logins?
And actually the, that'soutside of your organization.
You don't even knowthey're using this tech.
bet there's a lot of you out there thatare in that instance, and you won't
(17:12):
even know like how, and that's a that'sa hacker's best dream right there.
How that's actually quite hardfor IT teams to even manage.
Daniel Rowles (17:23):
Well let's talk about something with that then.
'cause I had a really interestingconversation with a very senior
person in a large retail organizationand it's a great conversation.
'cause they said, look, what they'reconcerned about is at the moment I want
to go and buy something and I'm gonna goto the retailer website and I'm gonna.
(17:44):
'cause I trust them, I like them.
Product quality's great,all that kinda stuff.
I'm gonna go there and I'm gonna buythat stuff Instead, now I go to my
agent and say, go and buy me this thing.
Go and find me the best one of thisbased on user ratings, and then buy me
if it's less than 10 pounds or whatever.
It's deciding which retailer it goes to.
So suddenly there'sdisintermediation in terms of like.
(18:05):
You know, we had this thing where, oh,I could go directly to the retailer, or
I could go to Amazon, and the retailer'strying to get you to come directly to
them, and some people are going to Amazon.
Well, now there's a wholenother tier where the AI is
deciding where you shop from.
Ciaran (18:18):
Well,
it's more nuance than that.
'cause actually what I found when itwas doing research, it was unable to
get information from certain websites'cause they were structured to prevent
a gentech bots from accessing the
Daniel Rowles (18:30):
They're not part of the conversation suddenly.
Ciaran (18:32):
I know, right?
But then it, the, thatmeans the data's skewed.
Daniel Rowles (18:36):
Hundred percent.
Ciaran (18:37):
And you're not necessarily getting the best results, but you don't
know that because it's all happeningunder the hood and you've closed your
laptop, gone off to make a coffee.
Daniel Rowles (18:44):
Right, so, so you've got two big issues already.
You've got the security issue where, youknow, if someone in your organization
has connected their agent and theirchat to BT or their Google Gemini, to
their OneDrive and to their SharePointand to their Outlook and to their
Gmail, and they've given it logins fortheir different platforms they use.
That's a way in to lotsof different things.
(19:04):
There's a huge security implication.
So suddenly these agents are decidingwhich websites that they use.
So this is a huge thing as well.
And then the third thing that we werethinking about, which is a real strategic
game changer is at the moment certainbrands have huge data assets, right?
So if.
You've been coming to my website for along time, like Amazon, for example, knows
(19:26):
what I like, what I don't like, what I'mlikely to buy, what I'm not like to buy.
And that's a closely guided secret 'causethey use that to, to target things at you.
Now an agent that I allow tolog into Amazon is not allowed
to take that data and put itinto their large language model.
I'm not suggesting that, but whatthey do learn over a period of time
is what I like and what I don'tlike within my Amazon account.
(19:47):
And for me, that agent is ableto give me personalization.
Across a whole range of websitesbased on my behavior across
a whole range of websites.
So actually now, like Amazon givesme good personalization to, it
might be a good reason to go there.
Well actually, if my agent learns loadsand loads about me over time, it can
personalize stuff about me a whole,across a whole range of websites.
And suddenly that data that Amazonhave got isn't valuable to them anymore
(20:11):
because yes, they can personalize, butI can personalize via my agent as well.
So there's this whole thing of like, is.
Do these people that have lots ofdata about us have really good value?
Well, yes they do.
With the large language models likeit, yes they would, but sooner or
later they're not gonna need itbecause what's gonna happen is that
they'll learn all that stuff themselvesand it'll become irrelevance.
(20:33):
So suddenly the AI agents and AI platformshave a big security portals, so there's
a huge level of risk that goes with that.
They're actually learningmore about us in one place.
Without really needing to use cookiesand all that kind of stuff, that was
limiting people like Google previously.
And actually they're taking kindof commercial organizations out of
(20:54):
the conversation unless they chooseto put 'em in the conversation.
So if, for example, I am sort of arogue AI company and everyone's using
my platform and I decide I don't likethat brand, I'm never gonna recommend
anyone go to their website and thereforeI could hit their commercial revenues.
And that's been the samewith Google for a long time.
Remember?
That Google can decide what showed up inthe search rankings and didn't, and have
(21:15):
had lots of legal cases about, they've hadto prove they're as unbiased as possible.
It's the algorithm and so on as well.
Ciaran (21:20):
Yep.
Daniel Rowles (21:21):
So these are really big conversations that
have come about really quickly.
And then on top of thatis the impact on jobs.
Yeah.
Ciaran (21:28):
The problem is these conversations are happening post-event not before.
Daniel Rowles (21:33):
Yeah.
Ciaran (21:33):
You know what I was saying in the other episode about, you know, wisdom
Thinking about the implicationsof what you do before you do
it, mapping that out a bit.
You know, and I just see thisall the time, it is just a
race to beat the competition.
And what I think is really interestingis that even the organizations that
report on this are struggling to keep up.
(21:53):
So, you know, I've been looking at someof the reports and data on this and
you know, people like McKinsey and allthe major US banks are reporting on
this, but the pace of changes so fast.
Daniel Rowles (22:04):
Right.
Ciaran (22:05):
information that's going to the decision makers is way out of date.
We're waiting on the next iterationof chat GPT, and there'll be
no news or evidence of whatit's gonna do or how it works,
Daniel Rowles (22:15):
And today does it?
Ciaran (22:17):
and then it changes everything and it's massively
ahead of any legislative curve.
Daniel Rowles (22:22):
Well, can I give you the only solution to this then
You've got a couple of theories in this.
One is regulation.
The problem is the regulators can'tmove as fast as the technologies.
That's not gonna happen.
You'd have to haveuniversal global agreement.
That's not gonna happen.
So the only opportunity you reallyhave here is one that we've spoken
about in digital transformation for along time, which is the organization
that is the most agile wins.
(22:42):
And I'd come back to the thing I'vebeen saying for a while, which is that
you don't need to know everything.
You need to know a bitmore than your competitors.
And that's why.
Podcasts, you know, live learningsessions a learning culture.
Now, I would say I'm ina training company, okay?
But the point being is unless you'rein better culture of learning at your
organization, you're gonna have a problem.
(23:03):
So iterative, always learning.
Always learning, always trying testing.
Ciaran (23:06):
Yep.
Daniel Rowles (23:07):
Creating an organization that is agile enough
to change quickly, and the biggerand more complex your organization
is, the harder it is to do that.
But that's the same of allyour competitors if they're
all large and complex.
Where it becomes tricky is if you'vegot competitors that are smaller
and agile, more nimble than you,they will take advantage of this.
So I think that we're gonna see thedestruction of large organizations quicker
(23:30):
than we will smaller ones, which hasbeen the case for a long time anyway.
But those large organizationsalso have the clout.
To go off and try and investin this and so on as well.
The thing I'd also bring you to is as wetrust everything less and less, 'cause
everything is deep fake and everythingis, you know, is it an agent, is it real?
Whatever.
That leaning into our humanity thingthat I've come to a few times before,
(23:50):
which like, what makes our brand human,that's the stuff that's really important.
Still, let's really kind of focus on that.
I would think about that, soI wouldn't panic about it, but
I would be highly concerned.
I would just say, what is it that I canadopt from this that's gonna be useful?
What don't I want to adopt?
But just looking at the fact that itwill really change the market, and
(24:11):
actually I think we're in it for aperiod of disruption in the fact that.
We will come to adjust to thesethings and kind of learn what
we want and what we don't want.
But in the interim, stuff's gonnahappen that isn't been planned.
We are gonna stop recruiting junior roles.
Potentially not we, but you know, anumber of organizations will, some people
will go through ruthless efficiencyand they'll cut loads of roles and
then we'll go, well, what happens whenyou haven't got a talent pipeline?
(24:34):
What happens when you haven't got adiversity of think in your organization?
What impact does that have in thelong term of the organization?
So I think it's making sure our short-termthinking and our long-term thinking are
actually married up a little bit andtrying to do some strategic planning.
Ciaran (24:47):
There's a danger in all of this.
I had a conversation with my bank,and I was frustrated a PayPal
payment hadn't been made, And PayPalwas telling me it had been made.
Daniel Rowles (24:59):
Right.
Ciaran (25:00):
the bank's telling me it's not.
So I'm like, how do I sort this out?
So I was having a chat conversation withmy bank and I said, look, this situation,
can you do anything to help me with it?
I just casually mentioned in the chatyeah, I'm a bit stressed about this
because I wanna get it sorted so I'm,and suddenly the person that I thought I
was speaking to clearly is an AI agent.
(25:20):
starts going off on one about, you know,do you need to talk about the stress?
in a way it
Daniel Rowles (25:25):
Oh, I see.
Right.
Ciaran (25:26):
right?
It's just unnatural.
And you just think, gosh, how manyof, how much of that's going on?
No one's looking at it.
How much damage is that being done?
Because I'm like, I'm not actuallyeven speaking to a, so they've driven
efficiency there for sure, and they were
Daniel Rowles (25:38):
But they've damaged their brand.
Ciaran (25:39):
but they've damaged their brand.
And I think that's a goodargument for actual people.
I've had it for a long timewith you know, customer service.
It's, there's nothing like beinggiven great human customer service,
Daniel Rowles (25:50):
Right.
Ciaran (25:50):
and I think if you slash and burn too quickly there, you will
reap the negative awards for sure.
You know, there, there's no reasonwhy a machine should be able to
outperform an actual human and a tasklike that, but I know everybody, every
company seems to be vying that as areal ripe of use for this technology.
Some of the voice AI now is incredibly
Daniel Rowles (26:11):
Wow.
You just look at 11 labs at the moment.
They've got the new voices whereyou can go say it, sarcastically,
whisper it, and things like that.
It's epically good.
Ciaran (26:17):
can hear the breath,
Daniel Rowles (26:19):
Yeah,
Ciaran (26:19):
Which is like, so, so, so much better.
But is it as good actually?
Where does this position you, isthere a a, an advantage within
the marketplace of actually havingreal people that genuinely care?
Daniel Rowles (26:30):
there is.
There is.
And I let's come to that a little bit.
I'm getting lots of people saying tome, in a lot of my students, a lot
of people that listen to the podcastsaying, I'm really worried about this.
I'm, you know, I'm a junior marketer.
How am I gonna get afoothold in my career?
Ciaran (26:43):
It is.
The first thing anybody does whenyou show them Agen AI is they start
testing it to see could this replace me?
Daniel Rowles (26:49):
Right,
Ciaran (26:50):
The fear is their lies.
And when they realize, actually, no itcan't because they're a bit more senior
or a bit more strategic, then they relax.
But that is short term thinking
Daniel Rowles (26:59):
right.
Ciaran (27:00):
opinion.
Daniel Rowles (27:01):
So what I've been saying to.
Students and people like that are kindof study marketing, getting into this,
it's not a bad career to get into.
Don't panic from the point of view thatwe're getting this technology first.
But what you need to be reallygood at doing is critical
thinking, problem solving.
You need some strategy frameworks.
You need the cause of, so thosefundamentals of business, fundamentals
(27:23):
of marketing become really important.
The ability to prompt really effectively,and I, everywhere I go still.
A really good approach to promptingis really rare still, you know,
that whole piece had right, andwe will do an upcoming episode on
some prompting techniques and goodprompts and so on as well, just to
give you some deeper tips on that.
But I think that critical thinking,problem solving, understanding,
(27:47):
you know, the economics of howthe world works, these are kind of
core skills and it's always beeninteresting to me that marketers
don't really understand economics.
Actually they really should becausethey're gonna see how does their
organization fit into the broader world?
How does that fit?
So there's those kind of core businessskills I think are really important.
So we will do a follow up episode on,you know, what skills are important
in the marketplace right now.
But I do think that organizations needto be looking at what they wanna do this
(28:10):
in the short term, but what will be thelong term implications of doing that?
And unless we're able to do that, we'regonna cause ourselves lots of problems.
And also, like with any of this stuffon the surface, the shiny, like,
oh, I can do this, it can do that.
Test it, refine it.
Now, one tip with AI agents,get it to do something.
See where it goes wrong.
Rewrite your prompt.
Start it again.
Okay.
(28:30):
Remember, every conversation in anAI is in a context window, meaning
it's remembering what you'vesaid before in that conversation.
Ciaran (28:38):
Yep.
Daniel Rowles (28:38):
very often when you have a conversation, it goes off at a tangent.
It gets it wrong.
Learn from it, improve your prompt.
Start the conversation again.
Okay.
So kind of go through that process.
The other question I've been havingis that as these things are more
and more powerful, they are usingmore and more processing power.
And actually there is a carbon impact ofusing this stuff as well, which there is.
(28:59):
And if you do a very simple AIprompt, you know it, it's using
a certain amount of carbon.
If you're doing a really complex prompt,it's going to use more carbon as well.
But what I would say about all of that.
Is the, and I have abig debate about this.
It's a really interesting organizationrecently, is that yes, that does use
more carbon and if you're just sittingthere willy-nilly playing with it,
you're just using it for the same time.
(29:19):
You sitting on Netflix for eight hoursat the weekend and watching Netflix
will obliterate that in terms of theamount of carbon you're using, right?
And I've got a chart and I will put thechart into the show notes in terms of
like, what does a Google search use?
What does an AI query use?
What does watching TikTok videosfor three hours a day use?
What does streaming Netflix use?
That video streaming isoff the charts comparison.
(29:40):
So I think that we've been a littlebit unfair in honing in on AI and
saying, oh, it uses load of card.
We shouldn't be doing this now.
There are environmental impacts,there are data centers that are needed
that's using huge amounts of water.
There is environmental damage.
I'm not denying that, but I thinkit needs to be a conversation
that's held in context as well.
If this is something that's of interestto you, please message in and we'll
do an episode on it, because I've gota ton of research we've been using
(30:00):
for a couple of clients recently.
So short term, amazing.
Long term.
What's the impact gonna be?
If you haven't played with theAI agent in Chate, go and do it.
It's gonna blow your minda little bit as well.
Think about how it's gonnaimpact your organization.
We have got a half day masterclassfor Target internet members coming
up on effective use of agents.
(30:21):
So if you're not a member,get signed up for that.
Also, we've got an update session forour newsletter subscribers and members.
So target internet.com/newsletter.
You've got those one hour updatessessions every month with me.
And they're a lot of fun.
We get a really big audience.
Everyone's coming together for thosea lot, so it's a lot of fun as well.
So.
Go off and have a play with it.
We'll do another session soon.
But there's a lot to think aboutand a lot to think about how it's
(30:42):
gonna impact your organization, yourrole, and the future of marketing.
For more episodes, resources toleave a review or to get in contact,
go to target internet.com/podcast.
Welcome to the Digital Marketing Podcast, brought to
you by target internet.com.
My name is Kieran Rogers.
Daniel Rowles (00:06):
And I'm Daniel Rolls.
Ciaran (00:08):
And today, Daniel, we are discussing AI agents.
Daniel Rowles (00:20):
Okay, so let's just define what an AI agent
is, talk about the evolution andthen talk about why we are here.
This very specific period in time.
Ciaran (00:30):
Why so profound Yeah.
Daniel Rowles (00:32):
Well, it's gonna get a bit profound in this one actually.
So, definition of a agen AI oran agent-based AI is it uses
sophisticated reasoning anditerative planning to autonomously
solve complex multi-step problems.
Okay?
So this started when chattwo BT and Google Gemini.
Release deep research modes.
I mean, it started way beforethat, but that's what's kind
(00:54):
of brought it to everyone.
So deep research mode.
If you're not familiar, you goin, you've in chat, for example.
You've got your little plus buttonas you have in Gemini, and you
could select deep research mode.
You'd give it a task, maybe to do someresearch and it would go off and ask
you a series of questions, alright,what target market are we talking about?
What's the time periodyou're interested in?
Et cetera, et cetera.
(01:15):
And then it would go off andcomplete a task for you autonomously.
Now the thing is, that's not quick.
It might take 10, 15, 20, 30 minutes to gooff and do that thing, but that was their
first kind of move into agent-based ai.
Then what you also had was chatt,BT were building this thing called
operator, and operator was theability of chatt BT to use a browser.
(01:39):
Not to search in a browser, but to, youknow, log into a website to go through to
fill forms and all those kind of things.
What's happened and we've just done anews episode on this is that chat two.
BT released their agent, so chat two.
BT agent and chat.
Two.
BT agent suddenly takes theresearch, the ability to use a
browser to manipulate documents.
(02:00):
Gives you this new thing thatcan go off and do stuff for you.
It is a game changer.
Now, I should say I got myself intotrouble with this, so I got hold of this
a bit earlier, so it's, by the time thiscomes out, it would've been out for about
10 days chat agent and I got, hold itreally early on and I started playing with
it and I started creating lots of testcases and I tested the heck out of it.
(02:24):
I then went, look, what about I wantyou to research content, create content,
log into my profile on LinkedIn andpost it how I you to go into all my
LinkedIn posts, and I want you toanalyze all of 'em and tell me what
content I should do in the future.
How about I use the Canva connector inChat two piti, if you're not familiar,
there are lots of connectors in chatthat can connect to external tools.
(02:47):
And I was able to connect itto my Canva account and I could
do this through the agent.
Either I could get it to log in to myCanva account, take a load of content it
created, and create a load of carouselposts for me, and then take those carousel
posts and put them into LinkedIn and.
Ciaran (03:03):
What could possibly go wrong?
Daniel Rowles (03:05):
Well, it didn't, well, the thing was nothing went wrong.
It did it.
It did a really good job.
But there was a risk.
Right.
Now, what I should say, with allof this AI agent stuff, there
are human intervention points.
So for example, when it goes to myCanva account, it doesn't own my
password, so I have to take over.
So I click the button, I take overthe browser for a moment, and I put
my password and my username in, and itlogs in and it carries on my behalf.
(03:28):
So let's just take a step back from that.
How this LinkedIn post got meinto trouble was that I said,
this is game changing technology.
I reckon I can replace a marketingexec with seven decent prompts.
Okay.
And my, what I went on to say was thatmeans you'll have huge cost savings.
You won't, you know, theseagents don't need holidays.
They don't want bonus pay.
(03:49):
Their dogs don't get sickand all this kinda stuff.
And then I did add, which is thepeople, the bit that no one seemed
to read was that you'll also haveno talent pipeline 'cause you won't
have any junior people anymore.
You're gonna have no diversityof think in your organization.
You won't have any youngpeople potentially, you won't
have a working culture and.
That's gonna have quite abig impact on businesses.
What do you think businesseswill do about this now?
(04:11):
Classic social media.
Nobody read the bottom bit tosee that I was being a little bit
facetious about it and was saying,
Ciaran (04:16):
You had
Daniel Rowles (04:16):
yeah,
Ciaran (04:17):
we don't need,
Daniel Rowles (04:18):
we don't need market execs anymore.
Yeah.
And my favorite one was that commentsaid, you clearly didn't really test this.
And it just like, you know, myprofessional integrity was called
out to quite a large extent.
But anyway I, what I normally dowhen I get angry about a social
media post is I refuse to respondto it until the next day, and I've
normally calmed down by the next day.
Ciaran (04:35):
Always
Daniel Rowles (04:36):
so, so I went through and
Ciaran (04:37):
edge.
Daniel Rowles (04:38):
very politely pointed out to everyone that, actually,
I wasn't saying it was a hundredpercent, but let's, so, okay.
There's a couple ofthings to pick out there.
First of all.
I can now get this to do things Icould never do before, like log into my
LinkedIn and analyze all of my posts.
I don't have to export it first.
So there's phenomenallyuseful things to do.
I can get it to log in andpost social media stuff for me.
(05:00):
Is that a good idea?
Probably not.
It's kind of risky.
But I can get it to log inand buy something for me.
So I don't think we're very far away fromthe first legal case of saying chat PT
spent all my money and I told it not to.
So, which is, yeah, there are lots ofsafety features in place, but my point.
This is game changing because we,this is the first iteration and we
(05:20):
know how quickly this stuff's moving.
Ciaran (05:22):
Yeah.
Daniel Rowles (05:22):
In six months, this will be radically better than
it is now because it's learning.
Okay.
And therefore you will be able to verysimply say, okay, what do I do every day?
What does my marketing exec do every day?
Let's turn that into a reallygood step-by-step, prompt.
Go to this website, do this thing,download this, go here, do that,
(05:43):
and then just take my role andbreak it down into that series of
tasks, and I can automate those.
Now, we've had robotic processautomation for a while, but it falls
over because of the fact that actually.
You know, when the website changes,it can no longer handle it.
Well, actually now the AIcan kind of overcome that.
The first time you use, by the way,chat BT agent and you sit there
(06:05):
watching it, browsing your websiteand telling you what it's doing.
Oh, well that didn't work.
I'm gonna try this.
Oh no, that didn't work.
I'm gonna do this.
It's quite a mind blowing experience.
I thought, I felt it was like,I'd really, you feel like you
experienced something very new.
It feels very science fictionally thefirst time you kind of do it as well.
Ciaran (06:20):
We, well, we were chatting on text, weren't we last week?
Have you tried it?
I'm like, no.
And you were like, well,have you got access to it?
I'm like, I dunno.
And I went in and looked I've becomea bit obsessed with this, but it's the
little icon that they shoved in there,kind of looks like a little robo print.
So I'm seeing it.
A wonky mushroom,
Daniel Rowles (06:38):
Righty muscle.
I'm sure that wasn't theirintent from a design perspective.
Ciaran (06:41):
I know it's, well, I think what they're trying to, like the icons, trying
to like illustrate the point that it canopen a web browser and click and explore.
Like that's what it
Daniel Rowles (06:50):
right.
Ciaran (06:50):
like to me.
But it does look like a bit like asquare foot, like paw print to me.
Like if
Daniel Rowles (06:55):
So to,
Ciaran (06:56):
pauses
Daniel Rowles (06:57):
to me it's just a screen with a pointer clicking on it, but Okay.
Ciaran (07:00):
it's a pore look, it's a paw.
You can sit, it's a pore.
Right?
Daniel Rowles (07:04):
No,
Ciaran (07:04):
if you agree with me right in.
Daniel Rowles (07:06):
it's a screen, a pointer and some like little dashes that,
that are showing movement, I think.
But anyway,
Ciaran (07:12):
it's
Daniel Rowles (07:12):
let's,
Ciaran (07:13):
I,
Daniel Rowles (07:14):
it's.
Ciaran (07:14):
What's clever about it is when they put the word agent
on it, it sounds really cool.
But is this really that differentfrom Gemini's Deep research?
Daniel Rowles (07:23):
It is wildly different because deep research
can go off and do things.
It can look at things and readthings, but this can actually fill
forms in navigate through a website.
You know, it's not justreading the text of the page.
It can interact with the page.
So if we go beyond just usinga web browser, I mean, that's
the main functionality.
It can use a web browsernow like a human can.
(07:43):
But also if you give it like it usedto be, you'd give it a spreadsheet,
it would analyze it, and then itwould give you some sort of output.
Now you can give it the spreadsheet andit can put the answers in the spreadsheet
So it's got the abilityto utilize things that it.
We'll put it in the show noteswhere target.com/podcast, the
functionality it has at the moment.
But the point being is that beforeI could look stuff up, I couldn't
(08:07):
actually interactively do things.
Now I can go to a logged in website.
I can say to it, go and find me arecipe for this many people that
this many dietary requirements,and then log in to my Tesco
account and buy that stuff for me.
And it's able to go through and do that.
And as and if it doesn't got aparticular product in stock, makes
some alternative options for me.
And there'll be human points.
I have to help it log in.
(08:28):
I have to say yes to purchaseand things like that.
But the, what's really interesting isall the marketing videos they created,
they gave the agent a task and thenthey shut their laptop and walked away.
There was a very big push in the marketingand then what happens with chatt PT is
you get a notification on your mobiledevice when that task is finished.
So it's like it's doing stuff for you.
(08:48):
It is your agent, it's going offand completing tasks for you.
So, and it, you're absolutely right,it is deep research mode combined
with the ability to use a browserand some other stuff as well.
So it's got more tools, it isgot more tools at it's disposable
and therefore it's disabled tochoose which tools it's using
Ciaran (09:03):
It's quite fun to watch it while it works.
Daniel Rowles (09:06):
Massively.
Ciaran (09:07):
a and actually I could see it was at one, think I'm, I
shared this in a previous episode,but I bought a laptop using the,
a agent feature with in chat GPT.
And actually at one point I couldsee it was writing a whole bunch
of code to compare some of thedata that it scraped and exhausted
and do something clever with it.
But you have to watch it.
(09:27):
It happens quite quickly.
Daniel Rowles (09:29):
Well, lemme give you another example.
This is amazing.
Ciaran (09:31):
Yeah.
Daniel Rowles (09:32):
It basically, I asked, I gave it an image and I'd done this
three months ago and it was an imageof a, with 99 dog faces on one picture.
And I said, how many dogs in the picture?
And three months going,sorry, I can't answer it.
I can't work it out.
It's too confusing.
This time it went right.
I'll go through and I will use.
An algorithm for facial recognitionand then it says, oh, it doesn't
work because the facial recognitionis set up for human faces.
(09:54):
But what I'm gonna do is quicklywrite some code that uses it
just to count the number of eyes.
'cause I know that dogs have gottwo eyes and it counts the number
of eyes and it got the answer right.
So essentially what we've got issomething now that has a number
of tools, it can do web scraping,like looking stuff off the web.
It can go to its large language model, itcan write code, it can use a web browser.
It's using those tools to make a decision.
(10:16):
What's the best way ofsolving this problem?
Now it's not.
Ciaran (10:18):
Game changer for summer fates when you need to know how many
jelly beans are in the kilner jar.
Daniel Rowles (10:22):
So, you know, the ability to do things quickly, effectively, in
a more creative way is certainly there.
What's interesting is that thisreplaces tasks that we do every day now.
It's not that I'm interactingwith it backwards and forwards.
It's like, right, go and do this, andit'll go off and do it, and I get back.
The outcome is what I need it to be.
It's not giving me somethingI can then do something with.
(10:43):
It's actually doing thething for me in many cases.
Ciaran (10:45):
So my experience with this is yes, and it gives some very impressive results,
you sure as hell have to double check.
Sense, check what it's done, or you'regonna get egg on your face pretty quick.
And I only know that because I'veproduced some true, truly wonderful
things using this, that I'm very proud of.
I know I, I'm very fortunate I workwith somebody who really, their head
(11:10):
is totally in detail and you make anyclaim and they'll double check it all.
'cause that's how they, that'show their brain's wired.
And it's brilliant.
'cause actually it means you,you're forced to sense check stuff.
Like I've
Daniel Rowles (11:20):
Right.
Ciaran (11:21):
land mine too many times.
And I'm not having to kirinDid you absolutely check this?
Like, remember when we were firstplaying around with chat GBT and
it was given, we were asking it forlatest news in SEO and it was given
us all sorts of absolute rubbish.
This is
Daniel Rowles (11:34):
Yeah.
Ciaran (11:35):
back.
Right?
But you know, I still think that.
That element is there, especially whenyou're playing around with different
models you've not used before and youhave to sort of sense check it and
work out where are the boundaries,what's it good at and what isn't it.
So that was gonna be my questionto you, like if people are
wanting to test this out, what
Daniel Rowles (11:50):
Yeah,
Ciaran (11:52):
would you recommend people, because I know if you haven't
used it before, you're definitelygonna wanna play with this.
What would you recommend Daniel People do?
Daniel Rowles (11:57):
well this, that's a really good question 'cause it leads us
into the prompting piece, which is thatI think a lot of people use this first
of all and went, oh, it's all right.
But it's not great because likethey did like a one line prompt,
Go to my LinkedIn.
Tell me what my best posts are.
Okay, so it'll go, but are yourbest poster ones that got the
most engagement are the onesthat drove your desired outcomes?
(12:19):
Are they ones that got the most views?
You know, you haven't reallydefined what best means,
But it will be able to log into yourLinkedIn and then it will go to each
of your posts and it'll try and getthe stats from and download them for
you and it'll, so actually that's thekind of thing you can get it to do.
That's, I would try it out withthose, kind of get it to do something,
you have to log in to do something.
So to a social media profile, toa shopping website, to, you know,
(12:41):
something like go to Amazon to buyyou something, whatever it might be.
Ciaran (12:44):
so I haven't gone that far with it.
But how did, how do you do that?
Do you give it thelogin details within the
Daniel Rowles (12:49):
No.
So you just give it, you say goto this website and do this thing.
And then when it gets there, it will,it'll open up its browser and it will
get to the login page and go, okay,well I think I need you to log in and
it'll give you a button to take over.
Ciaran (12:59):
Right.
Daniel Rowles (13:00):
then you take over its browser, you put your login details,
and then you hand the browser back.
And at that point.
Then does it, but it will rememberthat's what you've gotta be careful of.
So now it knows my login for LinkedIn,and if I ask it to go LinkedIn
again, it's able to repeat that step.
So, and depending on whichtool you're using, but what,
Ciaran (13:17):
Terrifying because far as I'm aware, we don't have two
factor authentication on our.
Language learning model logins.
Daniel Rowles (13:26):
no.
So what will happen?
That's why you want twofactor authentication.
Switch on everything.
Ciaran (13:30):
Yeah.
Daniel Rowles (13:31):
Because you can switch it on with CHATT P and
stuff like that now as well.
Ciaran (13:34):
Yeah.
Daniel Rowles (13:35):
So you can have two factor, because what you want it on chatt pt, and
you want it on all of your tools becausethen if someone, if you accidentally
are doing something with your agent, orif you know somebody's accessing your
agent, you want the security of knowingon your mobile device, oh, that's not me.
I don't want 'em to log in.
So yeah, security becomes evenmore important before there
are some mess up with this.
But what I'd say is that.
(13:57):
I would look at what you do every day inyour day-to-day role, and I would think
about which of those tasks are repeatableand which of those tasks are purely admin.
Right, as opposed to creative thinking.
So I know when we create this podcast,there's a number of steps I go through.
I take the podcast, I put it intoDescrip, I edit it in Descrip.
(14:17):
That is not something I'm gonna get anagent to do because there's loads of
nuance and subtlety and what's that?
Then I'm gonna go throughand I'm gonna create the.
Show notes that go onto the website.
Well, actually I can get AI to do that'cause I've trained AI to go through,
take the transcription, put out thekey points, structure it in a certain
way, find any links, put those up.
(14:37):
So actually that's astep I could go through.
Then from that, I'vegotta log into my website.
I've gotta put that onto a webpage,create the page, fill in all the
standard bits I've gotta then go throughinto Libsin and publish it in Libsin.
There's a lot of stuff in there.
I can get the agent to dofor me automatically, but
it's so business critical.
There's some bits I'm not gonna do.
I'm not gonna let it publish it toLibsyn without double checking it
(14:59):
and all those kind of things as well.
It's not a hundred percent at the moment.
It will do a task, get alittle bit confused sometimes,
and then just kinda stop.
Because it's not, it's got a bitconcerned about what's going on
or what have you as well, but thiswill iterate and this will change.
Now the important thing from thisis that one, this is gonna help you
be more effective and efficient.
So that's great.
Hopefully it gives you more timeto focus on the creative, the
(15:21):
strategic, the critical lateralthinking, whatever it might be.
But what I know it's gonna mean isthat as this gets more advanced,
what's gonna happen to junior roles?
I was being a little bit facetious, right?
You know, it will replace a marketingexec, but if it can replace 80% of what a
marketing exec does, do you think peopleare gonna still gonna keep recruiting
(15:43):
and everyone's going, yeah, 'causeyou know you need a talent pipeline.
You do.
But we know what companies that are likebig companies about efficiency already.
There's a report in theTimes, well ads, show notes.
You can see I'm not making this up.
Junior roles across the uk andthis is, there's a global pitch
that goes with it as well.
60% drop in recruitment.
For a couple of reasons.
(16:03):
One, 'cause the stuff that's goingin the economy, but also because the
fact they're going, well actuallyI might be able to get an agent.
I might be able to get AI to do this.
Do I need that role at the moment?
Where do I need to focus my efforts?
So people are just takinga bit of a step back.
It's not necessarily replacingthem, but it is making people
think about recruitment.
Take this six months down theroad, this will radically change.
(16:23):
Actually a lot of the stuff thatwe've got junior roles doing, we'll be
able to do and be able to do it withthese agentic tools really easily.
So it's a game changer for the workplace,and it's gonna change all of our roles.
It is gonna replace, I'vebeen preaching for ages.
Oh, it's not gonna replace roles.
And you know, it's onlycertain roles it'll replace.
It's mostly, it's gonna change roles.
I think this is gonna replace lotsof things that people are doing
Ciaran (16:45):
I just, you've just got me thinking about the
security of this though.
Daniel Rowles (16:50):
right.
Ciaran (16:50):
Like here we've got a back door.
Which could any, anybody in your teambe hooking it up to all your business
critical systems, giving it logins?
And actually the, that'soutside of your organization.
You don't even knowthey're using this tech.
bet there's a lot of you out there thatare in that instance, and you won't
(17:12):
even know like how, and that's a that'sa hacker's best dream right there.
How that's actually quite hardfor IT teams to even manage.
Daniel Rowles (17:23):
Well let's talk about something with that then.
'cause I had a really interestingconversation with a very senior
person in a large retail organizationand it's a great conversation.
'cause they said, look, what they'reconcerned about is at the moment I want
to go and buy something and I'm gonna goto the retailer website and I'm gonna.
(17:44):
'cause I trust them, I like them.
Product quality's great,all that kinda stuff.
I'm gonna go there and I'm gonna buythat stuff Instead, now I go to my
agent and say, go and buy me this thing.
Go and find me the best one of thisbased on user ratings, and then buy me
if it's less than 10 pounds or whatever.
It's deciding which retailer it goes to.
So suddenly there'sdisintermediation in terms of like.
(18:05):
You know, we had this thing where, oh,I could go directly to the retailer, or
I could go to Amazon, and the retailer'strying to get you to come directly to
them, and some people are going to Amazon.
Well, now there's a wholenother tier where the AI is
deciding where you shop from.
Ciaran (18:18):
Well,
it's more nuance than that.
'cause actually what I found when itwas doing research, it was unable to
get information from certain websites'cause they were structured to prevent
a gentech bots from accessing the
Daniel Rowles (18:30):
They're not part of the conversation suddenly.
Ciaran (18:32):
I know, right?
But then it, the, thatmeans the data's skewed.
Daniel Rowles (18:36):
Hundred percent.
Ciaran (18:37):
And you're not necessarily getting the best results, but you don't
know that because it's all happeningunder the hood and you've closed your
laptop, gone off to make a coffee.
Daniel Rowles (18:44):
Right, so, so you've got two big issues already.
You've got the security issue where, youknow, if someone in your organization
has connected their agent and theirchat to BT or their Google Gemini, to
their OneDrive and to their SharePointand to their Outlook and to their
Gmail, and they've given it logins fortheir different platforms they use.
That's a way in to lotsof different things.
(19:04):
There's a huge security implication.
So suddenly these agents are decidingwhich websites that they use.
So this is a huge thing as well.
And then the third thing that we werethinking about, which is a real strategic
game changer is at the moment certainbrands have huge data assets, right?
So if.
You've been coming to my website for along time, like Amazon, for example, knows
(19:26):
what I like, what I don't like, what I'mlikely to buy, what I'm not like to buy.
And that's a closely guided secret 'causethey use that to, to target things at you.
Now an agent that I allow tolog into Amazon is not allowed
to take that data and put itinto their large language model.
I'm not suggesting that, but whatthey do learn over a period of time
is what I like and what I don'tlike within my Amazon account.
(19:47):
And for me, that agent is ableto give me personalization.
Across a whole range of websitesbased on my behavior across
a whole range of websites.
So actually now, like Amazon givesme good personalization to, it
might be a good reason to go there.
Well actually, if my agent learns loadsand loads about me over time, it can
personalize stuff about me a whole,across a whole range of websites.
And suddenly that data that Amazonhave got isn't valuable to them anymore
(20:11):
because yes, they can personalize, butI can personalize via my agent as well.
So there's this whole thing of like, is.
Do these people that have lots ofdata about us have really good value?
Well, yes they do.
With the large language models likeit, yes they would, but sooner or
later they're not gonna need itbecause what's gonna happen is that
they'll learn all that stuff themselvesand it'll become irrelevance.
(20:33):
So suddenly the AI agents and AI platformshave a big security portals, so there's
a huge level of risk that goes with that.
They're actually learningmore about us in one place.
Without really needing to use cookiesand all that kind of stuff, that was
limiting people like Google previously.
And actually they're taking kindof commercial organizations out of
(20:54):
the conversation unless they chooseto put 'em in the conversation.
So if, for example, I am sort of arogue AI company and everyone's using
my platform and I decide I don't likethat brand, I'm never gonna recommend
anyone go to their website and thereforeI could hit their commercial revenues.
And that's been the samewith Google for a long time.
Remember?
That Google can decide what showed up inthe search rankings and didn't, and have
(21:15):
had lots of legal cases about, they've hadto prove they're as unbiased as possible.
It's the algorithm and so on as well.
Ciaran (21:20):
Yep.
Daniel Rowles (21:21):
So these are really big conversations that
have come about really quickly.
And then on top of thatis the impact on jobs.
Yeah.
Ciaran (21:28):
The problem is these conversations are happening post-event not before.
Daniel Rowles (21:33):
Yeah.
Ciaran (21:33):
You know what I was saying in the other episode about, you know, wisdom
Thinking about the implicationsof what you do before you do
it, mapping that out a bit.
You know, and I just see thisall the time, it is just a
race to beat the competition.
And what I think is really interestingis that even the organizations that
report on this are struggling to keep up.
(21:53):
So, you know, I've been looking at someof the reports and data on this and
you know, people like McKinsey and allthe major US banks are reporting on
this, but the pace of changes so fast.
Daniel Rowles (22:04):
Right.
Ciaran (22:05):
information that's going to the decision makers is way out of date.
We're waiting on the next iterationof chat GPT, and there'll be
no news or evidence of whatit's gonna do or how it works,
Daniel Rowles (22:15):
And today does it?
Ciaran (22:17):
and then it changes everything and it's massively
ahead of any legislative curve.
Daniel Rowles (22:22):
Well, can I give you the only solution to this then
You've got a couple of theories in this.
One is regulation.
The problem is the regulators can'tmove as fast as the technologies.
That's not gonna happen.
You'd have to haveuniversal global agreement.
That's not gonna happen.
So the only opportunity you reallyhave here is one that we've spoken
about in digital transformation for along time, which is the organization
that is the most agile wins.
(22:42):
And I'd come back to the thing I'vebeen saying for a while, which is that
you don't need to know everything.
You need to know a bitmore than your competitors.
And that's why.
Podcasts, you know, live learningsessions a learning culture.
Now, I would say I'm ina training company, okay?
But the point being is unless you'rein better culture of learning at your
organization, you're gonna have a problem.
(23:03):
So iterative, always learning.
Always learning, always trying testing.
Ciaran (23:06):
Yep.
Daniel Rowles (23:07):
Creating an organization that is agile enough
to change quickly, and the biggerand more complex your organization
is, the harder it is to do that.
But that's the same of allyour competitors if they're
all large and complex.
Where it becomes tricky is if you'vegot competitors that are smaller
and agile, more nimble than you,they will take advantage of this.
So I think that we're gonna see thedestruction of large organizations quicker
(23:30):
than we will smaller ones, which hasbeen the case for a long time anyway.
But those large organizationsalso have the clout.
To go off and try and investin this and so on as well.
The thing I'd also bring you to is as wetrust everything less and less, 'cause
everything is deep fake and everythingis, you know, is it an agent, is it real?
Whatever.
That leaning into our humanity thingthat I've come to a few times before,
(23:50):
which like, what makes our brand human,that's the stuff that's really important.
Still, let's really kind of focus on that.
I would think about that, soI wouldn't panic about it, but
I would be highly concerned.
I would just say, what is it that I canadopt from this that's gonna be useful?
What don't I want to adopt?
But just looking at the fact that itwill really change the market, and
(24:11):
actually I think we're in it for aperiod of disruption in the fact that.
We will come to adjust to thesethings and kind of learn what
we want and what we don't want.
But in the interim, stuff's gonnahappen that isn't been planned.
We are gonna stop recruiting junior roles.
Potentially not we, but you know, anumber of organizations will, some people
will go through ruthless efficiencyand they'll cut loads of roles and
then we'll go, well, what happens whenyou haven't got a talent pipeline?
(24:34):
What happens when you haven't got adiversity of think in your organization?
What impact does that have in thelong term of the organization?
So I think it's making sure our short-termthinking and our long-term thinking are
actually married up a little bit andtrying to do some strategic planning.
Ciaran (24:47):
There's a danger in all of this.
I had a conversation with my bank,and I was frustrated a PayPal
payment hadn't been made, And PayPalwas telling me it had been made.
Daniel Rowles (24:59):
Right.
Ciaran (25:00):
the bank's telling me it's not.
So I'm like, how do I sort this out?
So I was having a chat conversation withmy bank and I said, look, this situation,
can you do anything to help me with it?
I just casually mentioned in the chatyeah, I'm a bit stressed about this
because I wanna get it sorted so I'm,and suddenly the person that I thought I
was speaking to clearly is an AI agent.
(25:20):
starts going off on one about, you know,do you need to talk about the stress?
in a way it
Daniel Rowles (25:25):
Oh, I see.
Right.
Ciaran (25:26):
right?
It's just unnatural.
And you just think, gosh, how manyof, how much of that's going on?
No one's looking at it.
How much damage is that being done?
Because I'm like, I'm not actuallyeven speaking to a, so they've driven
efficiency there for sure, and they were
Daniel Rowles (25:38):
But they've damaged their brand.
Ciaran (25:39):
but they've damaged their brand.
And I think that's a goodargument for actual people.
I've had it for a long timewith you know, customer service.
It's, there's nothing like beinggiven great human customer service,
Daniel Rowles (25:50):
Right.
Ciaran (25:50):
and I think if you slash and burn too quickly there, you will
reap the negative awards for sure.
You know, there, there's no reasonwhy a machine should be able to
outperform an actual human and a tasklike that, but I know everybody, every
company seems to be vying that as areal ripe of use for this technology.
Some of the voice AI now is incredibly
Daniel Rowles (26:11):
Wow.
You just look at 11 labs at the moment.
They've got the new voices whereyou can go say it, sarcastically,
whisper it, and things like that.
It's epically good.
Ciaran (26:17):
can hear the breath,
Daniel Rowles (26:19):
Yeah,
Ciaran (26:19):
Which is like, so, so, so much better.
But is it as good actually?
Where does this position you, isthere a a, an advantage within
the marketplace of actually havingreal people that genuinely care?
Daniel Rowles (26:30):
there is.
There is.
And I let's come to that a little bit.
I'm getting lots of people saying tome, in a lot of my students, a lot
of people that listen to the podcastsaying, I'm really worried about this.
I'm, you know, I'm a junior marketer.
How am I gonna get afoothold in my career?
Ciaran (26:43):
It is.
The first thing anybody does whenyou show them Agen AI is they start
testing it to see could this replace me?
Daniel Rowles (26:49):
Right,
Ciaran (26:50):
The fear is their lies.
And when they realize, actually, no itcan't because they're a bit more senior
or a bit more strategic, then they relax.
But that is short term thinking
Daniel Rowles (26:59):
right.
Ciaran (27:00):
opinion.
Daniel Rowles (27:01):
So what I've been saying to.
Students and people like that are kindof study marketing, getting into this,
it's not a bad career to get into.
Don't panic from the point of view thatwe're getting this technology first.
But what you need to be reallygood at doing is critical
thinking, problem solving.
You need some strategy frameworks.
You need the cause of, so thosefundamentals of business, fundamentals
(27:23):
of marketing become really important.
The ability to prompt really effectively,and I, everywhere I go still.
A really good approach to promptingis really rare still, you know,
that whole piece had right, andwe will do an upcoming episode on
some prompting techniques and goodprompts and so on as well, just to
give you some deeper tips on that.
But I think that critical thinking,problem solving, understanding,
(27:47):
you know, the economics of howthe world works, these are kind of
core skills and it's always beeninteresting to me that marketers
don't really understand economics.
Actually they really should becausethey're gonna see how does their
organization fit into the broader world?
How does that fit?
So there's those kind of core businessskills I think are really important.
So we will do a follow up episode on,you know, what skills are important
in the marketplace right now.
But I do think that organizations needto be looking at what they wanna do this
(28:10):
in the short term, but what will be thelong term implications of doing that?
And unless we're able to do that, we'regonna cause ourselves lots of problems.
And also, like with any of this stuffon the surface, the shiny, like,
oh, I can do this, it can do that.
Test it, refine it.
Now, one tip with AI agents,get it to do something.
See where it goes wrong.
Rewrite your prompt.
Start it again.
Okay.
(28:30):
Remember, every conversation in anAI is in a context window, meaning
it's remembering what you'vesaid before in that conversation.
Ciaran (28:38):
Yep.
Daniel Rowles (28:38):
very often when you have a conversation, it goes off at a tangent.
It gets it wrong.
Learn from it, improve your prompt.
Start the conversation again.
Okay.
So kind of go through that process.
The other question I've been havingis that as these things are more
and more powerful, they are usingmore and more processing power.
And actually there is a carbon impact ofusing this stuff as well, which there is.
(28:59):
And if you do a very simple AIprompt, you know it, it's using
a certain amount of carbon.
If you're doing a really complex prompt,it's going to use more carbon as well.
But what I would say about all of that.
Is the, and I have abig debate about this.
It's a really interesting organizationrecently, is that yes, that does use
more carbon and if you're just sittingthere willy-nilly playing with it,
you're just using it for the same time.
(29:19):
You sitting on Netflix for eight hoursat the weekend and watching Netflix
will obliterate that in terms of theamount of carbon you're using, right?
And I've got a chart and I will put thechart into the show notes in terms of
like, what does a Google search use?
What does an AI query use?
What does watching TikTok videosfor three hours a day use?
What does streaming Netflix use?
That video streaming isoff the charts comparison.
(29:40):
So I think that we've been a littlebit unfair in honing in on AI and
saying, oh, it uses load of card.
We shouldn't be doing this now.
There are environmental impacts,there are data centers that are needed
that's using huge amounts of water.
There is environmental damage.
I'm not denying that, but I thinkit needs to be a conversation
that's held in context as well.
If this is something that's of interestto you, please message in and we'll
do an episode on it, because I've gota ton of research we've been using
(30:00):
for a couple of clients recently.
So short term, amazing.
Long term.
What's the impact gonna be?
If you haven't played with theAI agent in Chate, go and do it.
It's gonna blow your minda little bit as well.
Think about how it's gonnaimpact your organization.
We have got a half day masterclassfor Target internet members coming
up on effective use of agents.
(30:21):
So if you're not a member,get signed up for that.
Also, we've got an update session forour newsletter subscribers and members.
So target internet.com/newsletter.
You've got those one hour updatessessions every month with me.
And they're a lot of fun.
We get a really big audience.
Everyone's coming together for thosea lot, so it's a lot of fun as well.
So.
Go off and have a play with it.
We'll do another session soon.
But there's a lot to think aboutand a lot to think about how it's
(30:42):
gonna impact your organization, yourrole, and the future of marketing.
For more episodes, resources toleave a review or to get in contact,
go to target internet.com/podcast.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
My Favorite Murder with Karen Kilgariff and Georgia Hardstark
My Favorite Murder is a true crime comedy podcast hosted by Karen Kilgariff and Georgia Hardstark. Each week, Karen and Georgia share compelling true crimes and hometown stories from friends and listeners. Since MFM launched in January of 2016, Karen and Georgia have shared their lifelong interest in true crime and have covered stories of infamous serial killers like the Night Stalker, mysterious cold cases, captivating cults, incredible survivor stories and important events from history like the Tulsa race massacre of 1921. My Favorite Murder is part of the Exactly Right podcast network that provides a platform for bold, creative voices to bring to life provocative, entertaining and relatable stories for audiences everywhere. The Exactly Right roster of podcasts covers a variety of topics including historic true crime, comedic interviews and news, science, pop culture and more. Podcasts on the network include Buried Bones with Kate Winkler Dawson and Paul Holes, That's Messed Up: An SVU Podcast, This Podcast Will Kill You, Bananas and more.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.