Ep. 178: How to Future-Proof Your Content for AI Search Engines

Andreas Voniatis (Founder, Artios.io), who shared proven strategies on how to future-proof your B2B content marketing strategy for AI search engines. Andreas emphasized the need for high-quality, insight-rich content that adds unique value to target audiences. He also highlighted the key differences between traditional SEO and SEO needed for AI search engines, and talked about common pitfalls to avoid.

Episode Transcript

(00:00):
We all know the breakthrough SEO(Search Engine Optimization) is
challenging even for the mostambitious B2B brands today,
there are so many moving partsand so many factors to consider.
So how can B2B brands futureproof their content for AI
(Artificial Intelligence) searchengines? Welcome to this episode
of the B2B Marketers on aMission podcast, and I'm your
host, Christian Klepp. Today,I'll be talking to Andreas

(00:22):
Voniatis, who will be answeringthis question. He's a data
scientist, the founder ofArtios.io and the author of the
book Data-Driven SEO. Find outmore about what this B2B
marketers mission is.
Okay, Mr. Andreas Voniatis,welcome to the show.

(00:45):
Thank you. Thank you for havingme. I'm excited to share how we
can get into the minds of yourtarget buyers.
Absolutely, absolutely great tohave you on the show. I think
our pre interview call wasabsolutely dynamite. I think I
should have recorded that one,but, um, that was that was

(01:06):
giving me a little that wet myappetite a little bit, right? It
was a taste of what's to come.And I'm really looking forward
to this conversation Andreas,because, um, I think we're going
to dive deep today, um, divedeep into, I think, a topic
that's going to be even morerelevant to B2B marketers moving
forward. So let's dive right in.Shall we?

(01:27):
Thank you, Christian, yes, I'mvery excited. It's going to be a
lot of fun, as well assimulating.
Yes, absolutely, absolutely.Okay, so, Andreas, I would say
that you're on a mission to helpB2B companies deliver this is a
mouthful. Just hang in there.Data Science powered content
marketing that delivers therapid winds your C-suite demands

(01:51):
and the long term brand growthyour business needs for this
conversation, why don't we zeroin on a topic that I think has
become part of your professionalmission? How to future proof
your content for AI searchengines. I'm going to repeat
that one. How to future proofyour content for AI search
engines. So why don't we kick upthis conversation with two

(02:11):
questions, and I'm happy torepeat them. Question number one
is, why is breakthrough SEO sohard for ambitious B2B brands
today? And question number twois, what do you believe it takes
for B2B brands to succeed in theera of AI?
Thank you, Christian and bothvery good questions, if you

(02:31):
don't mind me saying so. So whyis it so hard for B2B brands to
achieve breakthrough SEO? Well,for a bit of context, we've had
two decades of SEO playbookswhere you know you would, you
would simply take up your targetkeyword, probably from a tool

(02:53):
like Semrush, and then you wouldproduce an ultimate guide to
that target keyword or topic.But the problem is, is that,
well, if everybody's doing that,then what think about the bigger
picture from a Google or an AIperspective, all we've done is
created a noisy internet, right,where everybody has this low

(03:17):
effort content, there's nothingoriginal, because most, most of
everything is just a betterversion of stuff that's already
out there. It's copy paste, butit's evolved. So you know, when
you're a B2B company that's justadding to the noise? Well, it's
statistically incredibly obviousto any AI or traditional search

(03:41):
engine, that your content is notreally adding any significant
value, and that's that is partof the problem that makes the
challenge highly challenging forB2B companies.
Absolutely, absolutely and thenon to that second part, or that
second question. Rather, what doyou believe it takes for B2B

(04:01):
brands to succeed in this AIera?
Yeah, so AI, it's great news.And I know it sounds a bit weird
for an organic marketer, strokeSEO, content marketer, to say
this, but what AI has done isit's increased the ability of

(04:25):
search technology to reallyunderstand what the content
means or the context around it.Previous attempts have been made
via schema and things like that,and using mathematical analysis
to estimate the probability theuser is satisfied with the
results and all that sort ofthing. But LLMs (Large Language

(04:46):
Models) have taken it to anotherlevel. Now this has several
implications for B2B marketers.
The first thing is, is that,like any PhD, when defending it,
your content has to tell theworld via the Internet something
they didn't know before. Okay,so that means that now it's

(05:10):
about high effort content asopposed to low effort content,
because if you're just producinglow effort content, then you're
simply adding to the noise onthe internet. And because AI,
especially with inflation,rising electricity and energy
costs, the shortage of AI chipsthat are available, it means

(05:34):
that AI is incredibly expensiveto run, even though everybody
wants it and doesn't want to payfor it.
So that means that your contentreally has to do a lot more with
than it did before. You knowyou're not going to succeed in
the AI search error by puttingout low low effort content. So,

(05:58):
so in terms of what high effortcontent looks like. You need to
be putting in some proprietaryvalue, some unique perspective.
So the traditional way of doingit would be to perhaps put out a
survey on a survey report on,you know, say your target topic

(06:23):
and and produce content likethat, because then you're
telling the world something newthat didn't know before. You're
putting together some uniqueinsights that can't be obtained
anywhere else on the internet.
Those are some really fantasticpoints. And I think there was
something that you said earlierthat really jumped out at me. It
was creating a noisy internet,because Ain't that the truth,

(06:47):
right?
Oh, yes.
If you, if you look at theinternet right now, you're just
inundated with all this noise.The, I think, the follow up
question that I had for youAndreas, and I know this is the
field that you are, or the, I'dsay, the lane that you swim in.
But how do you feel aboutcompanies or marketers using AI

(07:07):
generated content, and how doesthat impact the search ability?
So having built my own LLM fromscratch in believe it was 2020
around the same time that GPT 2came out, what I've learned is

(07:28):
that there's nothing worse thana model being trained on model
output. Okay, so this is one ofthe big things that AI search
technology fears, which is totrain their models on AI
generated content, because it'sa bit like taking a cassette

(07:49):
copy of a cassette copy of theoriginal. Alright, so if the
original are your target buyerstalking about having
conversations on on the thingsyou sell, and the search
technology like AI and LLMs aremodeling those conversations.
Then if you're then taking theoutputs of that model output

(08:13):
i.e. The LLM AI generatedcontent, and you're putting it
out there on the web, thenyou're not going to score very
highly from an AI searchperspective, because you'll be
seen as adding noise to theinternet, and they don't want to

(08:34):
train their models on yourcontent, so therefore you're not
going To be very visible in theAI search era, and so that. So
yeah, that that will lead on toa very interesting topic later,
if time allows, on, on how toselect your topics. But yeah,

(08:55):
that's so in terms of to for thesimple answer with a short
answer on AI generated content.Don't do it for that reason.
Absolutely, absolutely. That waskind of like the answer I was
expecting. But yes, I wanted tohear it from you, but, but there

(09:15):
are, there are so many companiesdoing that, though, right? And
you do see, you do see a patternof the type of content they
generate. It's, it's not veryoriginal. It lacks any kind of
like brand tonality, as far asthat goes. And there is, you
know, to your point, if you'retrying to offer a unique and
different point of view, well,guess what? AI generated content

(09:38):
is not going to provide that.
No, you know that. I think frommarketers, marketer we can both
agree that the whole point ofbranding is to be memorable,
distinctive and different. So ifyou're producing something that
is being produced by all of yourcompetitors. Then it flies in

(10:01):
the face of the principles ofbranding. Like, you know that
that it doesn't sound likemarketing thinking to be using
AI generated content.
It's like, it's like, back inthe day, and I'm dating myself
here. It's like we were requiredto study Shakespeare, and some
of the classmates resorted tousing, I think they were called

(10:23):
cliff notes, which is basicallyshowing you the summary of each
of each chapter of the story,and when we would have to write
essays or compositions thatassess that chapter. Well, the
teacher knew right away who wasusing Cliff Notes and who
wasn't.
Interesting. It's funny. You saythat my, my daughter has just

(10:45):
entered the top set for herEnglish class for the year and
and she, 12 months ago, she wasin in the bottom set. Now, you
know, with the right teacher,she's learned to appreciate the
nuances of language, thelinguistic devices and things

(11:05):
like that. So I can't imagine acliff notes would give you the
understanding and the use oflanguage if you're going to
answer those questions verywell. It might, it might be
useful for giving you thesynopsis, so that it gives you
the context. But beyond that.
Exactly, it basically isdesigned similar to AI to a
certain level. It was designedto give you the summary of a

(11:29):
chapter. So instead of having toread.
Yeah.
20 pages of ShakespeareanEnglish, it'll give it to you in
three pages.
Yes, yes.
Yeah. All right, I'm going tomove us on to the next question
about pitfalls that specificallycontent marketers need to avoid,

(11:49):
and what should they be doinginstead?
What content marketers should bedoing first of all is data
mining, like we do your targetbuyer conversations online, you
know. And there's manyplatforms, you know, there's a
reason why ChatGPT or Open AIhas a media deal with the
Atlantic and many other selectpublishers. You know, Google's

(12:12):
relying on Reddit, X is relyingon X, or Grok is relying on X.
And so, you know, this we asmarketers should be doing the
same thing, because where thoseconversations are happening,
there's a reason why AI areusing them. It's because they're
trusted by real, real people.You know, it's not an article

(12:34):
written by their SEO consultant.These are real conversations,
and they're heavily moderatedagainst spam.
I'm not saying spam won't getthrough, but there's much lower
chance of that. So if you'remodeling those conversations,
then you're able to produceinsights much cheaper than

(12:55):
taking out a survey, you know,with Ipsos or a similar agency.
You know that those surveys,especially if you want national
representation, they couldeasily run into the 10s of 1000s
in the US or Canada. But youcould, you know, the scraping
tools out there, and then you doa bunch of machine learning and
AI to to tease out, first ofall, you can, you can subset

(13:23):
those conversations for thetopic you're writing about. You
can even classify or clusterthose conversations into the
topics to build, build out atopic map of what it is you're
you know, around the things thatyou provide as a business, as a
brand. And then within thosetopics you can interrogate, you

(13:43):
can build virtual audiences ofyour target buyer and get their
views on a certain topic andcreate reports. And that is the
type of content that succeeds inAI search.
Interesting. And I guess anotherpitfall to avoid is like, don't
use AI generated content, right?
Yeah, don't. Don't use that. Butalso in the AI search error

(14:06):
tools like Semrush and Ahrefs,they're not going to work very
well, because the problem withkeywords is that keywords then
tell you who's searching againstthat keyword, right? Whereas, by
data mining communities online,you can subset those
conversations by by your ICP(Ideal Customer Profile) or your

(14:28):
target buyer profile, and then,based on that target buyer
profile, you can get their viewson the topic. And when you
produce that content, based onthat research, the content you
produce is is going to correlatevery highly to the fact checking
algorithms of AI, but also it'sgoing to resonate with the
reader when they read it and andwhen you put it on social media

(14:50):
or you or you write a pressrelease and promote it to
journalists, you're going to getpress coverage. So you know,
this is not just about searchand AI, this is about multi
channel. You know this? This ismulti channel value. This is
proper marketing. You know it'sit's not just about trying to

(15:10):
get traffic from the searchengines.
Interesting, interesting. Andsorry, I don't mean to ruffle
anyone's feathers here, but Ijust from what you were saying a
couple of minutes ago. Are yousuggesting that is from your
professional point of view? Isit going to be like the natural
evolution that tools likeSemrush and Ahrefs are no longer

(15:32):
going to be as relevant as theyas they were?
That's absolutely true. In myopinion. I don't think we're
ruffling feathers too much,because Semrush have produced a
tool called Datos, if I'vehopefully I've pronounced it
correctly, and they've got someaudience data. And again, you

(15:54):
know, these tools are thefuture, because, although we
don't provide a tool becausewe've created our own
proprietary technology. Again,we're going to end up with a
kind of software arms race. Atthe moment, we'd like to think
we're winning, even though we'renot a software provider. And the

(16:16):
reason why I say that is becauseit's one thing to profess that
you're producing insights froman audience, but the validation
is key. Like, the question Iwould ask for any of these
software providers is like,Okay, well, what's the, what's,
what's the margin of error for,for your insights, you know,
what is the confidence intervalfor, you know, how the how high

(16:42):
is your confidence interval forthe insights that you're that
you're generating for thistarget buyer when they have
views on a certain subject?
Now that's really important,because, again, if you want your
content to succeed versus acompetitor that's using a
similar tool, then, if your datais better, if your analysis is

(17:05):
better than the othercompetitors, then you're going
to, you're going to be morevisible, because your your data
is going to correlate much morehighly to to the models of AI
versusyou know, alternativesoftware providers.
Absolutely, absolutely. Okay,you kind of brought up some of

(17:28):
these factors already in thepast couple of minutes, but
break it down for us here,traditional SEO versus SEO for
AI search engines, keydifferences and what content
marketers need to be payingattention to.
So the key differences are thatin traditional SEO, there were
10 blue links, okay, and andwhat we also found is that the

(17:57):
tactics were high volume, loweffort content in and also
higher word counts, tended toperform really well. Now in the
AI search era, AI still relieson traditional search results,
and that would exp partlyexplain why Google has started

(18:19):
requiring JavaScript for AIsearch technology, or any
technology to access the searchresults so it could block open
AI from using their searchresults to ground their own
generated answers. So in the AIsearch era, it's all about high

(18:40):
effort content, more word countis not necessarily more, because
if you think about the elect,the energy costs and the chip
processing costs ofunderstanding content, it's much
better to be information rich orinsight rich in your content,

(19:02):
rather than having loads of workcount, because that just costs a
lot more computing flops, whichis a measure of computing power
for the AI to process thatcontent.
So the basic traditional SEOprinciples of making sure your
content is discoverable, makingsure it's unique enough to be

(19:25):
indexed or included in searchengines, and it's got authority,
they still apply, but in the AIsearch era, we just need to up
our game, so to speak. So thatmeans high effort, content,
Insight rich and and multichannel value, because now AI is

(19:50):
a lot more sophisticated to isusing so many different sources.
So it makes sense that if yourcontent is capable of being
liked by humans, not just by AI,which will be surfaced in terms
of signals when it shows up inin publications, publishing or

(20:10):
talking about your content, thenyou know you're winning. So
hopefully I summarize thatreasonably well.
Yeah, yeah, I'm mentally, I'vegot this chart with like, two
columns, right? And then, andthen I'm just trying to, like,
fill them in as you're talking,right? With the traditional and
then the the AI powered search,right? I almost feel also in

(20:33):
traditional SEO people, noteverybody, but a lot, lot of
people, were very obsessed withjust keyword stuffing, right?
Well, I'm glad you mentionedthat, because that was my other
bullet point. Just that I forgotto mention is that traditional
SEO is keyword driven, but AISEO is or GEO is sometimes
called now generationalgenerative engine optimization.

(21:00):
I've just seen that in the lasttwo weeks, that the other
difference is that GEOgeo or AISEO is people driven, okay, so
what, you know, I could see alot of thought leaders within
the SEO space, in the SEOindustry, that they're spending

(21:20):
a lot of time and focus tryingto reverse engineer the AI
search results. Personally, Idon't think that's the best use
of resource or time. I thinkyou're better off reverse
engineering your target buyers,and you should be anyway, not
just, you know, we've been doingthat for decades in market
research. That's what marketresearch is, right? It's getting

(21:41):
data on your target audience,but from an online perspective
and a GEO perspective, we shouldbe reverse engineering our
target buyers in order to workout what content we should be
producing. And with AI, youknow, you can do that at a
fraction of a cost, as opposedto traditional surveys, where it

(22:02):
can run into the thousands.
Well, it's not just about thecost, but it's also about the
time. I mean, just thinkingabout, like, the amount of time
it's required, that's requiredto put together a survey, then
then unleash that, as I call it,unleash it out into the wild,
and then collect all that data,synthesize and aggregate the
data. I mean...
Yeah, 100%. Yeah, no, there'sdefinitely a time aspect. Like,

(22:27):
you know, some of the marketresearchers that I work with,
you know that they willabsolutely refuse to do any kind
of work in the kind of holidayand the Christmas season for
that reason, because they can'tget anybody to actually...
Not in the West anyway.

(22:47):
No.
Fantastic, fantastic. Now,Andreas, you played that out so
beautifully, and it's such agreat segue into the next
question, where I'm going to askyou to, like, break it down a
little bit right, like, so interms of like, how B2B marketers
can leverage this data sciencepowered content for rapid wins.

(23:09):
So talk to us about what are thekey components required, what
are the steps they need to take?
Yeah. So the first thing I wouldrecommend is stitch your SEO
keyword tools that will save youa bit of money. Spend that money
instead on scraping APIs(Application Programming
Interface) and set them on tocommunity forums like Reddit, X,

(23:33):
LinkedIn, any, any I would doall the forums like we do, and
then build a database in thecloud so that you can start
storing quotes and profile withuser profiles against them. And
then do build a neural networkto start feeding that data in so

(23:59):
you can start synthesizing thoseconversations, and then you'll
have to obviously buildsuccessive layers within that
neural network to subset foryour target buyer, because
obviously you're only interestedin in their views. So this is an
infrastructure. It might soundfuturistic to some, but this is

(24:21):
already happening or should behappening. So that's the kind
of, you know, the future needsto be built now effectively.
Yeah.
If you want to get ahead.
It's probably a bit of both,right? Because, I mean, you
know, AI is such a I'm justgonna say it's such a continuous
phenomena. It's...
Yeah.
I mean, what was, what wasrelevant maybe six months ago

(24:43):
has probably changed. I mean,like, look how many versions of
chatGPT there are now, right?
Yes.
And a lot of people are alsolike myself included, are
starting to use perplexity morefor research.
Yeah, 100%.
Google, for example. And thereare some that I know that
stopped using Google forresearch completely.

(25:05):
Yeah, I see I kind of fall inboth camps. If I feel that I'm
about to write something that Ifeel well, I want a bit of an
essay output, then I will tendto use chatGPT or Claude. If I'm

(25:26):
just looking for a kind of quickanswer or I want to navigate to
a web page, then I will useGoogle. So it's...
Yeah.
Google's still in theresomewhere.
Absolutely, absolutely, it's notgoing to go into museum anytime
soon.
I think was about to say, Ithink they're doing a very good

(25:48):
job in terms of keeping up. Ithink, if anything, they really
need to. I know they're trialingtheir own version of chatGPT. I
forget what's called.
I have heard this rumor as well.Yes.
Yeah, so you know, and Google,I've got, you know what Google
does 27 times the amount ofqueries chatGPT does in any

(26:13):
given month. So I don't think,like you said, it's not going to
the museum just yet.
No, I don't think so. Okay, weget to the point in the
conversation where we're talkingabout actionable tips and
Andreas, you've given us plenty.I mean, I've been writing
furiously, oh, in the past 30minutes. But let's just say, for

(26:34):
instance, there is a B2B contentmarketer out there that's
listening to this conversation,and you want, you'd want them
to, like, take away all of thesethings that you've said and
implement them right like, rightnow, not in 12 months, not in
six like, after they listen tothis interview, what are like,
three to five things you wouldadvise them to do, like, right

(26:56):
now.
Start scanning the forums forthe well, to first of all, find
out what it is that your targetbuyers are talking about, you
know. So you know, if you're notdata savvy, as in, you know, you
can't code, or don't have accessto a coder to do all that

(27:19):
scraping, just, just just startreading them. You know, a good,
good PR person, they'reconstantly reading the news of
the journalists or editors thatthey're trying to win coverage
from. Do the same thing, anddon't limit it to forums. Also
read the news. And then, youknow, obviously, open up

(27:41):
spreadsheets. Start, start,start putting them in. And you
can try and feed it into say,chatGPT to, sort of say, right,
look, you know, I've got all ofthese opinions, you know,
cluster, cluster them intotopics oryou know, synthesize

(28:02):
some insights from this to helpme write better content, you
know?
I mean that in itself, isdynamite stuff. And, yeah,
you're absolutely right. That's,that's something that I could
even do after this conversation,right? So...
Yeah, 100%.
It's taking that data,aggregating it quickly, and also

(28:24):
try to produce an end result.Because I think at the end of
the day, that's, that's whatcontent marketers need to do.
Because, you know, it's not thatthey shouldn't do research, but
they shouldn't spend all theirtime researching.
Yeah, yeah, exactly. There's ahappy balance, isn't it? You
know, you can, you could get thedata. I mean, my mindset is I
would always go for the near ifnot perfect solution, and then I

(28:46):
tend to work backwards. So,yeah, absolutely. You know, the
true market, in my opinion,always falls to research first
and then move to strategy andtactics. A lot of there's a big
misconception that you knowsocial media is is or you know

(29:10):
Marcomms (Marketingcommunications) is marketing,
but Marcomms, as we both know,and I'm sure many of our
audience knows that Marcomms isonly 10% of the picture.
Did you hear that? Marcommspeople? No, absolutely,
absolutely, all right, you'vegiven us a lot now, right? We've
talked about, okay, how tofuture proof your content for AI

(29:33):
search engines. We've talkedabout the key pitfalls to avoid,
traditional versus AI searchengines, actionable tips. What
are some and you know, at somepoint these content marketers
are gonna have to show somebodywho is in a non-marketing role,
usually somebody that's higherup in the in the organizational

(29:56):
ladder, they're gonna have toshow them that look, I'm making
progress. Yeah. Yes, this isworking. Look, this is
succeeding. So we're talkingabout metrics here. And I call
this the Love it or hate itquestion, right? So what are,
what are some metrics that youwould advise content marketers
to pay attention to?
Yeah, I think the metrics,there's a couple of metrics,

(30:17):
right? There's the quality ofthe content and the volume of
traffic you get to it. So Ithink traditionally, content
marketing probably hasn't doneas good job at marketing the
tactic itself, in terms of as atool for driving conversions,

(30:40):
whereas the kind of statisticalreports that we produce on any
given topic, I find that theythey straddle that world very
well in terms of, you know, it'slike a hybrid approach. It's not
just top of funnel, but it'salso bottom of funnel. You know,
because when, when your buyersare searching for a solution,

(31:01):
they tend to want to learn fromthe experiences of people who
have been there before, orbought your product or bought a
solution similar to what you'veprovided. And the great thing
about doing these kind ofreports is you can be
accountable for conversions. Youknow, it doesn't have to be
that. That conversation of, ohwell, we just send you traffic.

(31:21):
It's like, oh no, hang on aminute. Let's look at both
traffic and conversions andengagement and shares and
backlinks and press coverage.And that's the beautiful thing
about AI, you know, contentmarketing for the AI search
error era is that you canactually have it all.

(31:42):
Well, that's a pretty boldpromise. You can have it all.
Well, let me, let me give you anexample. So for example, for a
law firm that we did some workfor, we did something around
inheritance issues, and so wewrote a report based on people
that that have been through aninheritance dispute going

(32:06):
through, you know, who was yourwhich family member was your
dispute with? You know, whatwere the assets in question? You
know, who did you go to forprofessional advice, that sort
of thing and you know, I spoketo the department head once it

(32:26):
got published, you know, withintwo months, we were, you know,
being cited in both AI andGoogle and Bing search on page
one. And I said, so you know,how's it working for you? And he
goes, Yeah. We've had sevenpeople, some new clients, walk
in saying they've, they've readthe report. Some said that, for

(32:48):
me, was an aha moment. I wasjust like, Oh, wow. It's not
just top of funnel. This. Thisworks, you know, it was, it's a
surprising benefit.
Yeah, and that's the thing inB2B, right? It's this. It's
these, not necessarilyeducational pieces, but there's
also, like thought leadershippieces. It's these reports, for
example, it's blog articles,it's it's content that helps to

(33:12):
demonstrate, you know, like inyour client's case, that
capability, right? That thatthat competency, but not, not
only that, but it's also theirperspective on things that are
relevant to their potentialcustomers, right?
Oh, definitely. I think the AIera searcher of content
marketing is that we're gettingaway from, you know, the

(33:35):
traditional SEO playbook oftrying to make your clients the
Wikipedia for whatever industrythey're in to actually just
putting out content that youraudience wants to read. You
know, they don't need to knowwhat is disaster recovery.
Chances are the people you'reselling disaster recovery to
already know what sort ofrecovery is. What they want to

(33:58):
know is like, you know, if I buythis, you know what happens, you
know. And if you can give themthat, that assurance and be full
of insight, then you know,that's brand awareness, right
there.
Absolutely, absolutely, allright, sir, get up on your
soapbox.

(34:18):
Okay, here we go.
A status quo in your area ofexpertise that you passionately
disagree with, and why?
I'm gonna go with BrightEdge anddoctor and Semrush.
Okay.
Yeah, because they're puttingout, you know, their tools are,

(34:46):
are built for the last, for thesearch era of the last two
decades, almost the last threedecades. And you know, it's all
keyword driven, and AI has justturned all of that upside down.
It's ripped up the rules to agreat extent. And I see them

(35:11):
bringing a lot of the 10 bluelinks thinking from traditional
search into their new tools. Andthe way they're kind of trying
to create new AI innovationswithin their tech stack, this
SEO software stack for for theusers, but it's just it just it

(35:33):
just does not fit. It just doesnot work. It's not fit for
purpose in the AI search era,and it almost demonstrates a
lack of understanding for forhow AI works and, and I'm not
saying you have to build yourown LLM from scratch, like I
have with some ex Amazonengineers, but yeah, it just the

(36:02):
status quo has to change. AndI'm sure they are changing. I
want to say to their credit if Iactually knew, but I will say it
when I see it absolutely,because I want to be a champion
for others.
Sure, sure. Fair point, fairpoint. But you know, I think
you're also in a in aninteresting space, if I may say,

(36:23):
so that's in a constant state ofevolution, right? Like, it's
sometimes feels a bit transient,doesn't it like things that, I
mean, there are some things likeyou said, like, there's still,
still some principles fromtraditional SEO that still hold
true. But there are, there'sthis other part of the spectrum
that has now completely changed.

(36:45):
Yeah, it's quite weird, becausein the last you know, even as I
observe, like over two decades,you know, the principles have
always been the same of, youknow, making content
discoverable, make making it sofit for search engines, fit for
users and fit for the Internet,okay, because those are the free

(37:07):
stakeholders whose experienceswe need to maximize. But AI has
so I don't think that wouldchange as much. It's the tactics
that always change, and italways has changed, and with AI,
it's really accelerated the needto increase the level of quality
or relevance to the targetbuyer, when it comes to the

(37:32):
tactical execution.
Which is a good thing, right?Because...
It's great, yeah.
You want the quality of yourcontent to improve and not get
worse.
I mean, 98% of the content andSEO industry are probably not
very happy about it, but I'mpersonally very happy about it,
you know. And you know, I thinkthe internet is going to be
better off as a result, youknow. So that, you know, it's

(37:54):
always customer first in theend.
Absolutely, absolutely. Okay,Sir. Here comes the bonus
question. Get ready. So let'sjust say, Andreas, you woke up
tomorrow morning and you said,You know what? Forget all about
all this, like data science andAI stuff. I'm done with it,

(38:14):
right? If you were going to giveup data science, AI search
engines, data science driven,content, all that good stuff.
What other passion would you bepursuing and why?
I'd be a chef.
Oh?
Yeah, really would at theweekends.

(38:35):
Jamie Oliver or Gordon Ramsay.
I would, I would go with GordonRamsay absolutely like I would
probably, you know, take anunpaid job at a Michelin star
kitchen, you know, take all theabuse that they give me just to

(38:56):
soak everything up and learn andwork my behind off. And that
would be, if I wasn't, you know,that's what I would do, even
that my tender age.
Amazing, amazing. So you weregonna say something about the
weekend. Did you? Did you hostthat?
Oh, yeah. I mean, I have over1000 cookbooks, and I, you know,

(39:23):
I'm always my family get a bitfrustrated with me, because they
always want me to cook theirfavorites, and I always want to
cook something new. So thecompromise is to cook the
favorite, but with a twist, youknow.
Ah, that's interesting. That'sinteresting. So what's your
what's your favorite dish tocook? I would say.

(39:48):
I gotta say that I do love Ithink kabsa lamb or lamb kabsa
is one of my favorites. It's aSaudi dish. They serve it at
weddings. I tend to love dishesthat are very...
Exotic?

(40:09):
Exotic, extravagant, you know,used in celebrations weddings.
It's usually a good sign ifthey're served at weddings,
because, you know, it's going tobe really good or either, and
the kabsa is bit like a biryani,except the rice is a bit more
like a risotto as opposed to soit's very but it's really

(40:29):
delicious. So highlyrecommended.
Wow, wow. So not only are you adata scientist, but you're a
chef as well.
Indeed. Well, amateur oneanyway.
Fantastic, fantastic. Andreas,this has been such a great
conversation. Thank you so muchfor your time. Thank you for
sharing your expertise andexperience with the listeners.

(40:52):
So please quick introduction toyourself and how people can get
in touch with you.
Yeah. So I'm Andreas Voniatis,Founder, CEO of Artios, you can
find me at artios.io, onLinkedIn. Feel free to grab me.
Always happy to have aconversation.

(41:12):
Fantastic, fantastic once again.Andreas, thank you so much for
your time. Take care. Stay safeand talk to you soon.
Thank you.
All right. Take care. Bye fornow.