LLM Visibility vs SEO KPIs

AirOps. Uh thank you for joining us on

the growth leader series this time this

week with uh Crystal Carter from Wix. Uh

say hey Crystal.

Hello everyone. Welcome to the webinar.

Yes. So glad that everyone is here. If

this is your first webinar, if you could

just put a star in the chat, that would

be great. And if you've been with us for

a few times, if you could put Oh, let's

see. What's your favorite emoji,

Crystal?

Uh, I like the one with the party hat.

That's my favorite one. Okay, so if

you've been with us before, put a party

hat. I'm gonna try to do a party hat

right now.

I don't know how far. Hey. Oh, thanks

everybody. There we go. We got several.

Oh, we got one star from Andrew. Bunch

of party hats. Thank you. Me and Crystal

were just chatting about um country

music actually. But you're a huge music

fan, isn't that right? I'm a big music

fan. Lots of lot I listen to lots of

different types of music. I just feel

like if they're if they're getting into

it, then I can get into it. Like so

yeah, that's that's how I feel about

these. Yeah. And I think we have a few

um Cowboy Carter references in the deck

today if I you know I mean I'm Mrs.

Carter. She's Mrs. Carter. Like it's

like true. You know I have been called

the Sasha Fierce of SEO. I might have

said it but that's not the point. No,

I'm

kidding. No. Yeah. So, a few Easter eggs

there for everyone to to check out. So,

uh let's give it one more minute. Oh,

we're two after the hour, so we can go

ahead and kind of kick off things

here. Today, we're talking about LM vis

LLM visibility and SEO

metrics. So many issues, so much

confusion I would say on these two

different things. Everyone has a million

questions about it. So Crystal is going

to bring her knowledge from Wix and all

of her experience to help us bridge this

gap. I really like that idea because we

are in both worlds in a lot of ways and

I'm sure we will dive into that today.

So let me give the official intro here

to Crystal um head of SEO communications

at Wix, digital marketing expert for

over 15 years. Um Crystal, why don't you

tell us a little bit about your

experiences at Wix and and your other

SEO experiences as well?

Yeah. So, I've been at Wix since 2022,

which is a long time ago in marketing in

the marketing world. Um, pre that's BC

before Chat GPT. Um, was sort of big on

the scene. And, um, and in that time,

I've done a lot of SEO um, in the

enterprise space. I've done um, and I've

done a lot of supporting folks who are

users on Wix supporting folks who are

considering considering joining Wix. And

then in my previous space I've also or

my previous uh work I worked as a

digital marketer freelance. I worked for

local authorities. I worked for um I

worked for agencies. I worked in house.

I worked with big companies like Tomy

and Disney um and those folks. I worked

with um smaller companies as well. I was

a very busy lady. And um and in that

time I work with people across lots of

different platforms and uh and helping

them with primarily with SEO but also

you know occasionally social as as

anybody who's ever done freelance

marketing knows like if the client needs

a hand you just help them with whatever

they need help with. Um so we did a lot

of that but today I work with Wix and I

work with Wix Studio and it's something

I really love because we have 250

million users across the globe. um sort

of like 22 different languages and we

get to be you know as in my role I get

to be front row seat for some of the

really cool stuff that we get to do.

We're very AI forward as a team and we

have it across lots of our lots of our

tooling and built into the SEO tools

built into the marketing tools um you

can even build a website with AI and we

also um enable folks across our teams to

adopt AI in order to um in order to

facilitate you know better workflows and

and things like that. So um 360 360

adoption uh from our team and and it's

something that I think has really helped

us um uh through this through this

fantastic and fascinating and

interesting and and complicated time as

marketers. Yeah, I mean you have such a

front row seat just seeing so many

different sites and how people are using

web and reporting all of those things.

So let's look into our agenda today. Um

going to be talking again about the

metrics that matter for for LLMs and uh

and in SEO and how those metrics

compare, bridging the gap between those

things. Um tactics for everyone to kind

of take away and use on their measuring

LLMs and in SEO. Then we're going to

also talk about like tracking and and

some things like that too. Um on the

right side you can see a lot of the

great customers and people who have been

with with Aerops. So I'm glad many of

them are joining here today as well. Uh

just a quick note, um we have a lot of

info. Crystal is just going to like kind

of power through this. We're gonna have

plenty of time at the Q&A at the end. Uh

drop your question in Q&A or in the

webinar chat. Uh Stephen and our team is

back there too. He'll be helping to

facilitate facilitate the conversation.

So if you have a question, uh leave it

there and we'll be we'll try to get uh

to it. So yes, Crystal, I will let you

take it away.

Okay. So yeah, in this section we're

going to talk about which metrics matter

in LLM brand visibility. And this is the

way that I'm referring to what some

people call GIO, some people call LLM O.

Um and and some people get very funny

about like whether it should be called

this or should be called that or should

just be called SEO. Um, but in my in my

experience, I find that there are

certain tactics that you can do that are

specific to LLMs and that that are

unique unique to that experience in the

same way that YouTube SEO has things

that you do on YouTube that are that are

specific to YouTube, but it's not just

related or unrelated to SEO. But I think

one of the things that's really

important about that conversation,

whatever you call it, is that you're

able to measure it. And this is one of

the reasons why I put together this this

um bit of of uh of uh deck uh and

information because I think as as mark

marketers, being able to measure what

you're doing is is absolutely crucial

whenever you're talking to anyone. So um

first of all, when we think about

classic SEO metrics, we think about

things like impressions or share voice.

And uh impressions is something you'll

be able to see in something like Google

Search Console. Share voice might be

something you see when you're looking at

cross competitors. And then we also have

keyword ranking and and position. We

also have pages that are indexed by

search engines, back links, referring

domains, all of these sorts of things.

These are things that we have tools to

help us measure things. And these are

ways that we're able to show whether or

not what we're doing is working. And

it's really important that we have these

and I think that there are some

parallels when we get into uh into the

LLM space.

So things things that are common across

both of them is that within um SEO and

within LL LLM optimization, you need to

be able to show uh show that things are

moving. So you are going to be making

sure that you're growing your traffic um

and conversions via LLM optimization. It

won't be it will not be a giant giant

wave of people um at this stage in the

game. However, you should be seeing it

gradually increase every month and and

people who are investing heavily in

this, particularly those who are in the

enterprise space will should see this

move move forward and they should see

this increase going forward. Um, I think

I've recently run the numbers and the

the LLM LLM use uh in in Q1 of 2025 was

about 15 billion uh visits per over over

the quarter as opposed to fraction a

fraction of that last uh the same time

last year. So, LLM visits are increasing

rapidly and so you should be seeing that

number increase as you're going forward

and the the effort that you invest in

now will help you to to do well when it

becomes more more significant. And in

SEO, you you will see your traffic and

conversions generally. So, when we're

comparing the two, you have your classic

SEO metrics of like share of voice and

things like that. And LLMs that the

easiest equivalent to that is brand

mention. So, this is when you're

mentioned in an LLM response. This may

or may not include a link, but it's

where you're where you're whether you're

mentioned in it. And then you have your

keyword ranking. So your keyword ranking

will be, you know, whether you're

position one or two or three. And then

um for LLM, this is the wrap of

citations. And I'll get into what that

means. It's my own acronym and I'll get

into what that means later, but you can

assess the quality of your of your

citation in an LLM based on that

framework. And then when we think about

classic SEO, your pages that are indexed

is very similar to the pages that are

retrieved or known to the LLMs. And um

not every page will be will be regularly

cited in LLM. So it's worth knowing

which ones are um and just like not

every page will be will be indexed.

Similarly um you have your classic sorry

the last one on the on the back links.

Oh, if you go back yes on the back links

and referring domains you'll have your

um your back links and referring domains

are similar to the website citations and

brand mentions. And I'll get into a

little bit about how you relate those

two uh as you go forward. So going to

the next

slide. So when we think about um about

mentions and citations, it's important

to understand which ones are which. So

um so for instance, if I have this query

of like can you name an influential

country music album for 2024, the

mention here is Cowboy Carter by

Beyonce. The citation though is to

Wikipedia, right? So that citation isn't

going to be.com or cowboycarter.com.

that citation is going to Wikipedia. So

that's important to to understand

whether or not you're getting the um

whether or not you're getting the

citation and the mention or the citation

or the mention. Next

one. Um so yeah, so similar to

impressions, your brand is included in

the information around the topic but

doesn't always uh doesn't always

necessarily link to a click. So or lead

to a click. So, for instance, if

somebody searches for your brand and

they and they see you have an impression

of your of your website, it doesn't

necessarily mean they're going to click

on your website. A brand mention is the

same thing. Your brand is understood to

be relevant to the topic, but it might

not include a link. We go to the next

one. And when you're thinking about your

citations, it's important to understand

the value of the citation. So, you want

to understand whether or not the the you

are mentioned regularly, whether or not

it's accurate. the prominence of your of

your citation and also whether or not

it's positive. So we think about regular

mentions. Um does your brand name appear

brand name and website appear regularly

for relevant LLM responses. So for

instance Wix is a is a CMS. It's a web

builder. So we would expect to see our

our name show up when people say like

what's a good web builder? What's a good

CMS? Where how can I build a website? Um

how can I create a website? Those sorts

of things. we would expect to show to

see Wix showing up regularly for those

kinds of queries accurate. So is are the

the citations and the mentions are they

accurate? Do they accurately reflect

what you do um you're offering? And are

they up toate about the in in the LLM?

So the up-to-date part is really

important because sometimes LLMs are

referencing information from their cache

of knowledge and sometimes they're

referencing information that's that's

active and and um immediate search

information. So, for instance, the other

day I looked up um I looked up something

about um I looked up something about

like is Dolly Parton a a great country

music country western uh singer. And if

you if you um go on Perplexity or Gemini

or something, they'll say yes, yes, she

is. But they won't necessarily give you

any links for that because they just

know that Dolly Parton's a great country

western singer. However, if you say who

has Dolly Parton collaborated with

within the last year, they will

reference recent articles because

they'll need to reference that because

their their training data that's that's

their corpus of content that's their

base knowledge might not be up to date

on that on on the last year, but their

search their search information will

fill that in. Now, the accuracy point is

also really useful because you can

respond to these. So if you see

information and I've done this before,

but if you see information that's

inaccurate, then you can um then you can

do reinforcement learning on the on the

LLM. So if you see on all of the LLM

responses, they have the thumbs up and

the thumbs down. If you see inaccurate

information, you can give a thumbs down

and that's reinforcement training for

for the LM. So if it say if it says that

you know your your home office or your

HQ is in Nashville and actually your HQ

is in is in Los Angeles and you can say

no it's that's wrong. Our HQ is actually

in Los Angeles and you can improve that

and I've seen that be improved across

across LLM. So you want to make sure

that things are accurate so that LLMs

aren't giving you giving false

information across your systems. Josh,

you look like you wanted to say

something on that. Is that a uh I was

just I mean I think everyone wonders

like how you do you handle when they

show inaccurate information like kind of

what is your course of action when when

that happens. Is there anything else

besides just like thumb down? Do they

have like a reporting mechanism or

anything like that? Yeah. So they do

have a reporting mechanism but basically

normally on the thumbs down thing

they'll say why did you give this a

thumbs down and you should give a

detailed detailed information about what

that is. Um, ideally if you can have if

you can you know trigger it a few times

and so that you can reinforce it again

and again to until it's correct I think

it's worth doing that and it's also

worth checking again. So, I've done it

before. Um I I I've shared the link to

it. We'll share the link to it later on,

but um I had a project called Site of

Sights, which started in 2024, spring

2024, and by the summer 2024, it was

showing in LLM. And then I said, "What's

the URL?" And they said, "Oh, it's part

of Behance." And I was like, "It is not

part of Behance. The URL is here at

sight of.co." And then I checked it like

a week later and I said, "What's the URL

for sight of sites?" And it said, "It's

here. It's at sight of.co." So like you

can you can do that. and they and all of

the LLMs ask for like regularly ask for

that feedback. So if you see in accurate

information don't just let it slide out

actually act on it. Yeah, that's good.

Yeah, prominence is really use really

useful as well. So very often if you say

name me the best the best web builder

for for instance they'll give you a list

the one that's on the top that's what

you're looking at like if do you show on

top. Now sometimes and I'll show this

later on sometimes they they have

different criteria for how they're

ranking them but that's essentially

ranking. So a lot of these tools will

have will say you know whether what

position you show in the list and those

the reasons why they show you in the in

the in the list may vary but ideally

obviously you want to be shown on the

top just because that you know

psychology and then around positivity.

So are you being mentioned positively or

at least fairly um with regards to your

brand. So you know are people asking

good or bad questions? are you showing

up as good? And basically, you want to

pull all of these things together to

understand the value of of the citations

and the

mention. To the next one. So, a couple

of examples of this. So, regularity of

mentions. If you click ahead, I've got

an illustration. So, I've got two

different questions about a country

music. Um, so I have, can you name a

country music an influential um country

music album from 2024? And you've got

Cowboy Carter showing there on the first

one. And then I say, can you list some

significant country pop albums by women?

And if you click on the animation,

please click forward. Yeah, you can see

that cowboy carter is also mentioned

there. Now, what's interesting here is

the one on the on the left. Um, so for

both of them, they're showing the one on

the right is showing uh is showing

fifth, but what's interesting here is

that this is

chronological. So, this is this is a

chronological list. So, when we're

thinking about the the wrap of the

citations and things, that's also

something to consider. But when we're

thinking about regularity, we're saying

like these are two related questions.

They're two of the kinds of questions

that someone who is interested in this

topic might ask and you want to be

making sure that you're showing up

regularly in

those next one. And then regularly of of

citations on this. So this has to do

with the URLs again. So if you go if you

click on the animation for me. So we

have one there that's going to

Wikipedia. Then we have one that's going

to a different website. So we have two

citations, but the citations are both

different. And that's that's something

to consider as well. So whether or not

you're seeing the same the same URL show

through is something that you should be

considering when you're thinking about

the regularity of your

citations and then accurate. So can you

name an influential country music from

uh album from 2024 and it says um yeah

here's Cowboy Carter and we go through

that and that's a good one. We can go

forward on the

animation. So that one's showing showing

accurate. Um and then when we think

about the mentions as well and the

mentions and citations um the when we

think about the the ranking of your of

your keywords on particular topics. So

let's say and let's say the topic was

like country music albums or something

or country music albums or by Beyonce or

something to that effect. Um, country

cowboy Carter showing top on the ranking

would be would be, you know, what you

would measure on traditional SEO and it

would measure the position of the pages

that are triggered around the keywords

of like country music Beyonce or country

music album or women like country music,

etc. Your wrap-up citations of whether

or not you've got regular, accurate,

prominent, positive citations. Um, this

will give you an idea of whether or not

the quality of your mentioned citations

are good.

And so this is one that's around

positivity as well. And this is working

well. Here we

go. And so when we think about um so we

think about that that works really well.

And again with the prominence we had we

have the one that's showing on the top

and then we have the one that's showing

up

fifth. Scroll down. So that's showing

first and that's showing fifth. And

again that's something to consider as

well.

So, when we think about bridging the

gap, we can use a lot of the tools that

we've used for traditional LLM

optimization or traditional SEO. Um, and

I and I always cringe when I say

traditional, but classic SEO, shall we

say? Um, we can think about all of those

when we're bridging the gap between like

moving into LLM optimization, adopting

LLM optimization as we go along. So,

when you're thinking about your share of

voice and your brand mentions, this

isn't to say that that these certain

things don't matter. they still do.

They're still very very important. Um,

and when you're thinking about this, you

should think about your mentions within

LLMs similarly as you do with your share

of voice so that you're benchmarking

yourself against competitors um, around

the prompts, for instance, around your

core topics. So, for instance, if I was

looking at like female country music,

uh, country music albums, for instance,

I'd want to look at like where other

folks are showing up, where I'm showing

up, and where that's where that's all

all working in the same way that I would

with share of voice, for instance. So,

there's some SEO tools which will tell

you that, you know, which which

percentage of the of the searches you're

showing up for on that particular topic

and help you pull that through. When

you're thinking about your keyword

ranking, again, that's similar to sort

of the wrap of your um of your brand

mentions and LLM responses. So, you want

to know the quality, not just the number

of your of your LLM uh mentions. And you

can use keywords to to help you research

your prompts and help you help you to

create example prompts. So, a lot of the

LLM tools will help you to see this. So

they'll help you to generate prompts

that you can use um to sort of test. And

it's of course very difficult to um to

completely replicate what a user might

might query in an LLM, but you can get

something in the ballpark. And a lot of

LLMs, as we know from the AI overviews,

will very often distill their LLM

responses into something into something

like fairly similar um that that's

repeatable.

Um, and I think also a lot of these

tools will also give you the opportunity

to ask the same prompt overall or like

again and again, which will allow you to

see how the how their responses evolve

over time. When we think about pages

that are indexed by search engines, the

the LLM equivalent is pages that are

known to LLMs. So, you can identify

which um that you're you can identify

which pages are being seen by LLMs and

then you can identify which pages are

most valuable for LLM. So if if it's the

case that like maybe your homepage isn't

showing up on the LLMs very often, but

this one blog really is really

frequently then that's something to

consider. If it's a case that that you

know when people say is your product

good or bad or like is your product

worth it? That's something that a lot of

people will ask. Um then it's worth it's

worth having a look at um how you are

how you are

identifying h whether or not your your

whether or not your content is showing

there and also whether or not um the

content that you want to show there is

showing there. So it might be that again

it's showing one blog but it's not

showing another one. And the other thing

you want going want to consider is like

whether or not you're adding back links

to those pages as well.

And speaking of backlinks, we think

about the backlinks um overall. And when

you look at your LLMs, you want to look

at the website citations from LLM

responses. So again, if I was to query

something like what's the best country

western uh country music album and like

let's say the CMA uh website showed up,

do I have backlinks to the CMA website?

Do I have partnerships with um with

certain newspapers that are that have

partnerships with LLMs, for instance?

Are you able to audit your backlinks to

understand which ones of the backlinks

and which one of your backlink partners

um is is regularly cited in LLMs or not?

Because it might be that that you're

spending a lot of energy on a particular

website and they're never showing in

LLMs or you're spending a lot of a lot

of energy on a particular rel

relationship and they're they're not

showing in LLMs. But this other these

other folks that you are spending a

little bit of time with, they're showing

LLMs a lot. that might mean that maybe

you should invest more time with that

with that particular outlet and less

time with this with this other

outlet. So that's something to consider

as

well. Okay. So when we think about to

sum it up when you're thinking about

your rankings, ranking is something that

people consider a lot. Um when we think

about SEO for LLMs, again we want to be

making sure we're getting regular

website citations as much as possible.

This is how you will drive traffic to

your to your website via LLMs. And this

is how you will get, you know,

measurable data on on, you know, how

people are interacting with LLMs and

whether or not what the conversion rate

is like for for traffic from LLMs and

which kind of content is is best for for

LLM. So, it's worth worth working uh on

getting those

citations. Go through accurate and

factual. Again, we want to make sure

that they're saying if they're saying

stuff that is relevant and is clear and

is accurate um about what what we do.

LLMs are prone to hallucinate. They're

prone to give information that's not

exactly correct. And if you have

accurate and factual information, that

will help. Also, um, folks who have an

active PR team are also less likely to

see inaccurate information because a lot

of the search enabled LLMs are pulling

heavily from news sources to help shore

up their um, their information. So, when

particularly when there's breaking news

or people are asking questions about

something that happened like yesterday,

then they're going to be checking on

checking the news sources in order to

collaborate or corroborate their the

information they have um, within their

general data store.

Then we go to prominent positioning um

and first listed. Again, we had some of

the content there that was showing us as

as first for like general country music

album and then we had it where it was

showing fifth and then that fifth one

that the reason why it was showing there

was because because it was alphabetical.

So that's not necessarily a panic panic

stations. Um but it does does mean that

we should pay attention to to the

prominent positioning overall. Um even

and and if there are some exceptions

like that one that's showing in

alphabetical or showing in chronological

order then it's worth noting noting

which kinds of queries will will show

that kind of that kind of uh

prominence and positive or fair

sentiment. Again everybody wants to be

thought of well and I think that this is

also something that that comes to mind

with regards to um creating content in

the LLM with within the LLM reality that

we have. So we are seeing a space where

sometimes SEO does isn't giving the same

amount of return that it used to in

terms of in terms of content creation in

terms of like traffic and things like

that. However, this doesn't mean that

you shouldn't make good content. It

doesn't me mean that you shouldn't make

content that's that's um you know good

quality and that that that explains what

you do well. In in fact, it almost means

that you should do more of it because if

you aren't speaking positively, clearly,

articulately about your brand, then when

a some when someone goes to an LLM and

asks a question about it, they're less

likely to actually refer to you. They're

going to be they're going to refer to

somebody else who's who's talking about

your brand. So, um, one of the things

you can do think about uh for a positive

and fair sentiment is to think about

making sure that you're speaking

positively and fairly about yourself. um

which is you know I mean that's that's a

skill for life. So I like that a skill

for life.

Okay. So tactics everyone can

use when we're thinking about um about

uh LLM's KPIs. So the first one is um

common metrics that matter. So traffic

and conversions these are standard

across LLMs and SEO like they and they

will be forever and it's not again you

should set your expectations with LLMs.

you're not necessarily going to get um

you know millions of people flooding

through the door straight away. However,

again, you should see a steady uptick in

the number of people that are coming

that are coming through. So, you should

be looking at the trend going up um as a

as a general rule, but and and you also

want to think about your conversions as

well. So, how can we do that for

everyone? So, conversions from LLM. The

first thing that you can do if you

haven't done this already, this is

something that you can do straight away.

It costs nothing and you should be

tracking these is basically conversions

from LLMs. So lots of people on their

contact forms have a how did you hear

about us um you know selection and you

should have chat GPT or AI um written in

there straight away. So it costs nothing

and from what I understand from people

who have implemented this anecdotally

the the leads tend to be good and even

if you don't have this on the form if

you like let's say you're a freelancer

or let's say people don't you know let's

say you're a smaller business or let's

say pe people don't always fill in the

form it's worth asking this question

directly as well um if you and even if

you have people who are working in

customer customer care it's worth asking

this question as well because you you

know it costs nothing and you it'll get

you some more information and I think

people are people these days I think are

quite proud like oh yes I found it

through AI so that's that's all exciting

so if you go forward you can see an

example of one that they've done from

SER interactive SE interactive are a

great team that have been very very

forward on this for a while and they

added in a section that says um you know

how did you find us you know referral

conference webinar chatbt and just ask

ask the question straight

up and this is um Will Reynolds he's the

um the uh MD and founder of C

interactive and he was saying that

basically seeing traffic from chatbt and

perplexity and really wanted to know how

that was working and so they added the

drop down and it helps them to track the

deal as as they go through. It's also

really useful because you don't always

get a referral link directly from like

they sometimes it shows up as direct

traffic. So it's worth making sure that

you get that when people are filling in

the

form. Um you can also filter your

traffic in G4. So this is a an example

of a reax that um that was set up from

Jess Schultz um a few few months back

and you can filter it. If you go forward

there's a bigger example of the reax I

think. Yes. And you can filter by lots

of different types. Um so I don't think

this one has Grock in it but you can add

that in um for instance. And it just

gives you an example. If you need a

little help with this um Chach can

probably help you with this as well if

you're not sure like what do I do with a

reax? Um, I highly recommend that you

have a have a look at that. And if you

go to that um search engine land

article, there's also information about

how you could imple implement that in uh

with Google um with Google Looker

Studio, which I think we also get into

in this one as well. Next. Yes, there we

go. So, this is how you set it up in the

back end in um G4. And if you go forward

to the next slide,

um yes, you can add that into and if

you're not completely familiar with G4,

we have some great resources on the Wix

uh learning hub and you can uh you can

go and learn some more about G4 and how

to how to get going with that, how to

set up filters, how to set up

explorations. So that's super useful for

you. And then also you can you can use

that to um help you to use Looker Studio

as well to get more um ongoing

information on that. Yeah. Very. And

then the next Yeah. And also you can

track um track AI referers. Um so in Wix

Analytics we have a couple of different

filters for that. We literally just

today um launched a new filter

specifically for AI traffic from from

bots. And so we have a dedicated report

for um for AI traffic um that you can

that you can also see as well. So you

don't have to click the filters, but um

look in your CMS. Hopefully, it's a wick

CMS, but look in your CMS and see check

out your referers and see if they've got

AI um AI filters for for that. And you

might be surprised at the kind of

content that you're getting. The other

thing is is if you are somebody who's

like, "We should be tracking this and

you need to convince higherups um or

stakeholders about whether or not, you

know, this is worthwhile." It's worth

going through your referrals and going

through your referral traffic to see to

evidence um historic traffic to be able

to show them that look we're we are

getting this traffic and if we and if we

invest a little bit more we could get

some more. So go in and have a look at

your AI referers and see if you've got

some traffic

there. And then you can monitor bot bot

traffic. So there's different kinds of

bots. So there's a user bot, there's the

search bot, there's other things that

are going through and there they go

through your website and they have a

look at all all the different different

pages that are going on. The other thing

that's important is that for some of

these they list them explicitly. So chat

GPT for instance lists all of their bots

by IP and some of them are listed by

name. Um some of them are a little bit

trickier to find. But um but yeah, I

think that if you are somebody who gets

technical with your bot traffic, you can

get right up in the bot log reports and

see exactly which which parts of your

site that you're they're crawling. I saw

I saw someone who who um ran with this

and they were able to find, you know,

certain exactly, you know, which

components of their website the LLMs

were crawling and they were able to get

some great um some great improvements as

as a result. Do you have to enable these

bots in any way on your site or or how

does that kind of work? No, so it's the

same as it's the same kind of uh

framework as the bots who will go to so

in crawling crawling an SEO crawl your

website. So bot will crawl your website.

If you go to the next slide, I think

there's some information about it. So

yeah, so what happens is um and this is

comparison between common crawl which

was one of the basis one of the basis

for the original uh for chatbt3 for that

version and versus versus Google mobile.

And you can see the um the big long ones

are um are Google mobile. If you go to

the next slide, I

think oh no, sorry, go back.

Sorry. Um so yeah, what what happens is

they will crawl your website and they

will look to discover new new content.

So um they'll they'll look at

everything. They'll follow links.

They'll go to different places and

they'll try to find the content that is

is most relevant. What very often

happens with um with some of the AI bots

is they crawl in a different way. So

common crawl for instance will crawl all

of the the URLs on on one website rather

than crawling like a few URLs which is

what generally happens with Google.

Google's like oh we will crawl this one

this one and that one but we don't need

to crawl that one we won't look at this

we won't look at that but AI AI tools

will will perform differently but

essentially they're they're doing a

factf finding information um to find to

find out which which pages are available

on your website and which pages they can

serve to users. You don't have to enable

anything. However, um and I've got

another article which I I'll link to in

the next part. Um you can guide what

they crawl if you if you uh choose to do

so. So, in the next one, yeah. So, if

this is this is all completely brand new

to you, um that there's a link in the

middle to my article about SEO for brand

visibility and LLMs and I go through

lots of the sort of basis for um LLM

visibility optimization. And if you want

to get into this KPIs or um information,

I go into a bit more detail in my um

LLM's uh visibility versus co KPIs. And

if you are interested in how to develop

prompts as well, I've got a resource on

the Wix SEO learning hub about that

helps you to sort of shape which kind of

prompts you might want for um for your

LLMs and uh to help you to help you to

sort of measure your measure your

growth, measure your improvement, and uh

measure your performance there.

Yeah.

Yeah. Any final maybe takeaways for

anyone who's just kind of starting or

exploring um LLM visibility and getting

their content in there? Yeah. So, I

would definitely say um to have a look

through your analytics. So, whether

you're using G4 or whether you're using

like a platform based analytics, have a

look at your analytics to see what what

traffic you're getting. Now, you may

very well be surprised. And um you can

have a look at both the traffic and also

the referers and have a have a look at

that information because again you'll be

very surprised and I think a lot of

people are thinking about LLM visibility

but sometimes I say you know we need to

stop just thinking and get involved. So

there's lots of different ways that you

can do that and also test. So, there's

um there's another good tool. Um Elena

Solles has has a great um uh LLM

visibility tool and you can put in your

brand name and it'll ask you a couple of

questions and it'll give you a sort of

quick overall summary of your of your

product. And it's a it's a custom GPT.

Um and and it sort of gives you a quick

overall summary. And again, this is

something that you can take to higherups

to say, "Hey guys, like this is this is

something that's really useful. um you

know we we're not showing up and we

should be or you know we are showing up

and we could be doing better and this is

something that's really useful but I

think that I think that have have some

data have a look at your at your um your

existing information because we are

starting to see starting to see traffic

and have been seeing traffic for about

over a year now from multiple multiple

users and um you know people like or to

tools like perplexity claude um chatgbt

they're seeing millions and millions of

users every every month. So, the chances

that your brand, even if you're a

smaller brand, um aren't showing ever,

are fairly slim. And I think that it's

worth it's worth getting involved um so

so that you can so that you can take

advantage of uh potential conversions.

Yeah, that's great. Um we have a ton of

questions coming in. Um let me pause for

a moment and talk about AirOps and then

we will go through all of these

different questions. So a lot of teams

are using air ops as well to optimize

and to refresh your content and all of

our AI workflows accelerate your content

and your strategy. So you can see here

on the left you can build your own

custom workflows to help you create

content for LLMs but not just create new

net new content but al also optimize uh

the content that you already have. And

we do connect to Chad GBT, Perplexity,

um, Claude, all of those to help you

create and like refine your content. And

then, um, a big part of what we do too

is just scaling with experts in the loop

and humans in the loop. And you can see

our our grid there. Everything goes

through a human review process or we

suggest that everything goes through a

human review process to make sure that

uh, your content is meeting your brand

standards and and where it needs to be.

Um, and so if you are ready to grow your

content and have a great content

refresh, uh, we do have that code on the

right that you can scan. A couple good

examples here recently.

Um, we're seeing a ton of customers use

AirOps to optimize and refresh their

content. I know that's been like a

manual process. uh you know if you have

to change from like 2024 to 2025

updating old um stats and data AOPS can

really help you do that. Uh love this

quote from Dcript. Uh you can see it

there at the bottom. They we just have a

new case study. Um Stephen can drop that

in the chat there of just a exceptional

program that they've done to refresh all

their content. Um so you could be losing

money to your competitors uh if you're

not thinking about how to refresh and

and optimize your content. Um, so feel

free to email us as well, teamupops.com.

Um, mention the webinar in the subject

line and we would be um, happy to to

meet with you and talk about that. Uh,

let's see what we got.

So Q&A, lots of stuff coming in. I have

a few questions, too. Um, one question

that's top of mind for me is because I

was on LinkedIn yesterday like all of

the SEO community seems to be a lot of

the time and I saw like competing ideas

about yes, the strategies you use to

rank in

Google also relate to appearing in LLMs.

And they had a study where they said

like 70% of things on first page of

Google are also usually prominent in

LLMs. And then I saw another report,

another thing that said like actually

these are completely different systems.

Google's search engine results do not

apply or I'm sorry, Google's ranking in

Google, it doesn't necessarily mean that

you also rank in LLM. I'll put it that

way. Um, so what are you just seeing

about maybe that overlap of if you're on

page one in Google, are you more likely

to appear in LLMs or or kind of what

what are you seeing with that?

Um so I think that this is this is a

it's a very early space. Um and I think

that we are in a really really

interesting space. So I think that the

the information and the activity is very

is related but not exactly the same.

Like they're very very much related.

They're not exactly the same. The

toolings behind LLMs work differently

from the toolings behind behind a

traditional search engine. Um so with

LLMs for instance you know they have a

store of knowledge and there the the

knowledge is you know will vary by by

platform. Um and so they have their

their training data and their training

data is updated on a periodic basis with

within a search engine. They're

constantly updating their index all the

time for instance. So like that's

something that that's slightly

different. Um there the way that a

search engine crawls a website is again

also different from the way that an LLM

does. An LLM will crawl will crawl the

text of your website as a large language

model. So it will crawl the words on

your page. However, it won't necessarily

crawl like you know the the other sort

of tech components behind there. Like

for instance, it doesn't directly crawl

the structured data on your page.

However, a search engine will crawl the

HTML, the the structured data, the

JavaScript, like all of that sort of

stuff. It'll crawl all of those

different bits. So there's that. We are

seeing a convergence of a lot of this

technology. So when I first started

looking at this, Claude, for instance,

wasn't search enabled, but they now are

search enabled. And most of the these

LLM tools have search included as part

of them. Um, so they'll include them,

but for a search engine, you always get

links back. Even when there's an AI

overview, you always get links back. But

when you enter in an LLM, if it's a

general knowledge question, if you ask

them, "How do I make guacamole?" They

will, unless you're on perplexity,

they'll just give you a guacamole

recipe. they won't necessarily give you

a link to a guacamole recipe um because

they have different they have different

different goals and they have different

outputs. Now this means this means that

the content that we create for these

things is different. It means that the

tactics are slightly different. And

personally as somebody who's been

looking at this, I've done optimizations

for LLM optimizations that have impacted

LLM visibility and have had no impact on

how we perform on Google. None

whatsoever. but they did have an impact

on how we performed in LLMs. And so to

my mind and based on the evidence and

the things that I've seen, they are

different. They're very much related,

but they're different in the same way

that like, you know, PR and SEO are

related. You know, if you if you have a

TV commercial, um then your SEO will go

up because you had a TV commercial. Like

that happens as well. Um so I think that

they're I think that they're related. Um

but not exactly the same. And to be

honest, like I don't know. I feel like

everybody keeps arguing about it, but I

just feel like just, you know, we just

need more we just need more people

having a go at it rather than arguing

about what we call it. Yeah, that's a

really good point is it's we're still in

an experimental stage. We're still

trying to figure things out. So, you

really you have to like do some of the

stuff on your own and and take in the

information, compare what people are

saying, and then kind of test it out.

Yeah, absolutely.

Uh we had a good question from Lindsay.

Um and this was a question I had too.

She was wondering how do I get my

citations to increase on on LLM search?

Like is there a way to

um you know help your the number of

times you might be cited on a certain

topic or on a certain query?

Yeah. So from what I can see recency

matters um and particularly the recency

around the around the topic. So if it is

a topic that the LLM feels confident

about then they're less likely to as in

like they have a they have a corpus of

knowledge about that particular topic

then then again they're less likely to

they're less likely to include that. So

for instance like guacamole recipes

there's they will have they will have

read 700 million guacamole recipes in 17

different languages or something. Um, so

if you have if your content is like a

guacamole, not going to have um, you

know, great impact there. Um, however,

if you if you have something that that

is about something that happened today,

like I don't know, somebody like a news

anchor went on TV and she had guacamole

on her shirt or something, then that

might be that then you're more likely to

show in there. So I think that things

that are things that are unique

um things that are things that are

unique information and things that are

timely information are more likely to

show to show. So I think particularly

trend and newsp content is more likely

to show um uh particularly for for that

sort of stuff.

Do you have any suggestions about like

product reviews because like you know

what is the best guacamole or something

like that? I mean, I don't know if

there's any guacamole um products on on

our chat here, but I know a lot of

people are in in SAS and enterprise. So,

maybe like how does XYZ product compared

to ABC product? Um what are you seeing

around around you know searches like

that? So, particularly around product

based things like they're they very

often um like I very often see own

brands being being cited in in that

content. So for people who don't have

battle cards for instance on their on

their website like they probably should.

Um and I think that like that's

something that that's something that's

that's really important. So, so you

know, for people who don't have their

own product information on their on

their website, you you probably should.

And I know that it sound that seems

obvious, but I've seen quite a few

people, particularly startups, who don't

have, you know, good corpus of content

around around what their product is and

what their product does. Um, and they

don't invest in in um in uh reviews for

the also useful for reviews um is

Reddit. Reddit gets cited a lot. Um, and

and Reddit, you know, being involved

with Reddit Reddit communities is

something that can be really really

valuable for Google and also um for for

LLMs. Um, and and I think that you can

you can see how how that how that's

working. And I think also it's worth

worth having a look at the the kinds of

sites that are showing. So, for

instance, if you're seeing that um that

you know, Reddit for is showing up

regularly when you see when you see

those reviews, then it's worth worth

maybe investing in that. If you're

seeing the other websites are showing up

regularly, then maybe it's worth

investing in in those ones. Um because

that's what I mean like when you're

looking at the at the back links for

instance like which of the which of the

link partners that you have are being

listed by by LLM are known by LLMs are

regularly cited by LLMs because

everybody ain't like some people are and

some people aren't. So if you're if the

places where you're investing your your

energy on reviews and I know that you

know a lot of big SAS companies do spend

a lot of energy on reviews and getting

good reviews. if they're not showing up

in um in LLM responses and what you want

to do is show up in LLM responses and

you need to, you know, pivot your energy

to to ones that are ones that are

relevant. Wow, that's that's a really

good point to to take away for sure. Um

I think a broad question that everyone

has here in mentioned it specifically is

just like how do we know what keywords

or terms or questions people are asking

in in Chad GBT or perplexity or Claude?

I know there's like um several different

tools nowadays that that can handle some

of that, but maybe give us the like kind

of manual version of that and then maybe

like if you have any tool

recommendations or just some ideas about

how you can just see what people are are

searching for in these in these LLMs.

Yeah. So one of the things you can do um

so in the tool that I shared the prompt

the prompt generator tool the prompt

library tool um and also with chatgpt

for instance like you can and with a lot

of the the tools there's a whole range

of tools available now you can take a

collection of keywords from particularly

like let's say a cluster uh let's say

you had like let's say you had a sauce

cluster you had like guacamole salsa uh

like tatsiki whatever so let's say you

have like those are your keywords and

you like what are things that and and

You would say like you you would say to

an LLM like give me a collection of

questions that that someone who was

interested in um in you know buying

guacamole for instance might want might

ask um and and you can say and you can

mention their audience um you know

somebody who is going to throw a Super

Bowl party um you know and they're based

in Texas or something and then they will

generate a question if you give them

like see give me 20 questions they'll

generate questions for you and then you

can you can go through and try them and

also then and then you can test them

out. So what's useful if you use

something like GPT for sheets for

instance you can run this and you can

change the model so you'll get different

different um outputs per model and

they'll they'll generate the questions

for you and put them in the sheet as you

go along and then you can check them as

as you go along. There's also other

tools that you can you know that have

like big subscriptions that you can that

you can include. Um, I' I've mention

I've reviewed a few of them. Um, and

they do they do lots of different things

there. But I would say that that if

you're not sure, I would take your

keywords and the keywords that you find

drive good conversions and tend to drive

good conversions bottom of the funnel

and think about those. Think about

taking those and asking Chachi PT

because the other thing is ChachiPT has

all of that training knowledge from all

of those things. Ask Chach, ask Gemini,

whatever. um give me some questions

based on based on these keywords and

again I would emphasize bottom of the

funnel because of because they are doing

so much of the of the top of the funnel

stuff right so I would be like you know

what is guacam how do I make guacamole

I'm sorry about the guacamole everybody

um but like I'd be like yeah but I'd be

like I'd be like what is guacamole and

they'd be like guacamole is a dip from

Mexico and I'd be like well how do I

make guacamole and they'll give me the

instructions I go well that's really

complicated who can I get to help me

like where can I buy guacamole because I

don't have time I don't have the

ingredients, etc. And so then that

bottom of the funnel, that bottom of the

funnel question, you know, like what is

what are the bottom of the funnel

questions? That's where you get uh

things. We're talking about Cowboy

Carter. I'm going to the Cowboy Carter

uh concert and I wanted to get um a hat

like Beyonce. And um and so I literally

said like, "Where can I get a hat like

this?" And in the sources they had lots

of lots of other folks, but in the

response they said, "This place, this is

the person who made the original one.

this is the place um you know place

where they used to where they sell it.

This is a place that sells something

almost very similar and I was like

fantastic great you just saved me three

hours of trolling all over the internet

and whoever on their website had like

you know you can buy this cowboy carter

but hat that d you know they were they

were doing really really well. So, think

about think about the keywords that that

are bottom of the funnel because a lot

of the top of the funnel stuff will be,

you know, may very well be um captured

by the LLM itself, but if they know that

you're bottom of the funnel um uh

keywords are there, then they're more

likely to to help uh you know, seal the

deal in the end. Yeah, I think two

things are standing out to me there is

the LLM already have so much training

data. So, if you ask, you know, I used

to work at project management software

tools, what is project management? it's

probably going to tell you and not be as

likely to list out a citation. But then

if it starts saying like how does this

tool compare to this one or what are

some new trends in project management

for example, it's more likely to surface

a citation because there's like trend or

new information. Is that fair? Yeah.

Yeah. New information is super is is

very much your friend. And then another

thing that stood out to me from what you

said was um well now I'm forgetting it.

Oh yes, use chat GBT to to figure out

what questions people are surfacing. So

if you are trying to figure out like do

I need a tracking tool or not? I mean a

really she just gave a really good hack,

so to speak, is just like put in your

ICP and your persona and then say like

what types of things are this person

searching for related to this topic and

then those are almost like your search

queries almost like a Google PAA where

you're just like these are some of the

frequently asked questions of my ICP. I

need to get content for that and then uh

it's more likely to to show up. Yeah,

absolutely. So, so yeah, use the tools

to to help you figure out the tools. And

also, if you and also, you know, if you

don't see, this is another hack that I

saw someone else talking about. I mean,

not necessarily a hack, but um we have

an article that was talking about, you

know, SEO versus versus GEO and all of

that sort of stuff. And someone was

saying that one of the things that's a

benefit of um of LLMs is that if you

enter like, you know, if you enter like

this top project management tool and

let's say you don't show up, you can ask

it why didn't you mention this company?

and it will say because this this this

and this and then you can go back to

your company and go hey guys chachi said

this what you know we should fix this

because of that um or you know or why

did you list that company and over this

company or that that sort of thing so

you can you can it's a two-way

conversation and I think that it's

really important that marketers remember

that throughout the entire process

that that is really great advice and

something I I sometimes forget too um I

believe I don't know who's behind Uh let

me see intellecta marketing I think. Let

me find your question here. Essentially

they were asking if like technical

structure like schema and things like

that matter are different say than than

appearing in in Google or other search

engines like do we need to focus on a

different type of schema or just a

different type of formatting or should

we just kind of focus more on on the

content strategy. So I guess it's kind

of like a technical SEO versus content

marketing marketing type question to to

appear in LLMs. So I think that schema

schema markup is um schema I love schema

markup. I've written a lot of structured

data. I've been team schema markup for a

long time. Love schema. However um the

return on schema has changed um since

2018 uh for instance. So, it used to be

that, you know, you'd write schema

markup and you'd get events uh listings,

but now a lot of a lot of these things

are covered by um AI overviews or

they're covered by other things um and

and stuff like that. So, so um I think

that one of the things that's really

useful for schema markup is to reduce

the lift um so that you still get the

benefits of it, but it doesn't

necessarily take up lots and lots of

time. So for something that's a really

really important page for you, sure,

spend a lot of time like making sure

you've got bespoke, you know, very good,

very detailed, very high quality schema.

For something that's general, at Wix,

for instance, what we have schema markup

built into our page type. So if you make

an event page, we build in schema markup

for it. If you make up a blo if you make

a blog page, then as soon as you publish

that blog, it already has a schema

markup built into it. Um, and videos,

you know, they're all optimized for rich

results as well. So they're all rich

results eligible product pages etc

already built in. So if you can

templatize it so that it's in your

template then that reduces the the

output but still gives you the benefit.

Um and that's because because when you

so all of these um search enabled LLM

they're all grounded by search and if

they are feeding off of search results

which include rich results then it can

potentially impact how you how you show

on how you show on the LLM. So, I didn't

I I've been uh this this debate's been

raging hot in the SEO space for a little

while. Um and I so I did some research

on this and particularly around recipes.

Recipes were like ground zero for for

schema markup from the beginning. Um and

uh and so I looked up like vegan

chocolate cake for instance and there

was nine rich results there. Um and then

I looked up the same question on

perplexity and it was there was quite a

few that were that came through there. I

think it was like seven out of nine or

something were mentioned in the in the

results of perplexity and then I went

through to I think Gemini or Chhatrick

PT and there were a few more that were

that went all the way through from from

Google through to the LLM through to the

other LLM. So schema is not dead but you

don't necessarily need to be investing

lots and lots and lots and lots and lots

of schema on everything for a while.

Like I know everybody was putting FAQs

on every page and how-tos on every page

and all of that sort of stuff. you don't

necessarily need to do that. Um but but

you know for your high quality pages

it's useful and I think that if you're

able to do it in a scalable um scalable

way it should be able to benefit both

your SEO and your um LLM optimization.

Uh I know we're kind of but at time

here. I have two more questions. This

came in um I believe from Tasmine. He

emailed me. But how important is like

the LLM's txt as far as I'm actually a

little confused on on exactly what

that's used for myself, but like how

important is is that for for ranking or

appearing in LLMs.

So that's a that's still a new tool. Um

so adoption is um adoption is like is

increasing. Um but people are sort of

sort of testing it out. I think my my

opinion is that like it is it's a page

that you know would be it's it's a page

that won't be seen necessarily by most

folks. It's you know something that sits

on your you know with your robots.txt.

If it is a low lift for you if your team

has the technology and it would take you

like a couple of hours to put together

why not test it out and see what's going

on. I think again at the moment we are

in a space where everything's super um

super new and it's very very fastpac at

the moment and if you're able to get to

get it going then if they if these LLMs

do start using them and and some of

these LLMs are using them themselves so

they've started testing them clearly

because I think anthropic for instance

has an LLM.ext on their on their

website. Um, so if they do start start

using them more consistently, then um

then you know that's you're in a stay

ready so you don't have to get ready

situation. So my my recommendation is

that if it's a low lift, go for it, test

it, see for yourself. If it is um if

it's you know you're going to have to

raise a ticket and they won't be able to

get to it for seven months and you need

to get in a freelancer and a this and

that to get it done, then maybe don't

worry about it. It's not as important if

it's if it's a huge technical lift like

that. Sure. Um, and then kind of maybe

to to end our conversation here, what is

the SEO community getting wrong about

LLMs in your opinion? What are some

myths or uh that you're seeing out there

um about LLMs and AI visibility that

that maybe you don't necessarily agree

with?

Um, you hear people say that it's just

the same that and like and I just think

it's just not the same. Like there's

there's lots and lots of overlap, don't

get me wrong. Um, but like there's lots

of overlap with digital PR and there's

lots of overlap with um with, you know,

e-commerce SEO or local SEO or or social

media and and I think that there's that

just because there's overlap doesn't

mean that it's not a distinct a distinct

thing that people should be should be

looking at. Um, so I think that it's

it's um I think that it's a fascinating

space and I think that as as an SEO um

who's interested in this, I'm I'm really

excited to see how it develops.

That's a great uh note to end on. Um

really appreciate your time, Crystal.

This was I love the Guacamole

references. Love the um Beyonce

references. Uh that adds a little bit

different flavor to the typical SEO

webinar for sure. Well, I I do love

guacamole. So there you go. Yeah, there

you go. Um, and don't forget to connect

with Crystal. She's on LinkedIn. Um,

she's on Twitter. Lots of stuff that she

has written at Wix. So, um, you know,

Google her or LLM, search her, uh, and

in her articles on Wix for for those

visibility articles. Um, lots of helpful

stuff there, too. Uh, and we have a few

more strategy sessions with Aerops. You

can find the information in the chat or

just feel free to email us at

teamairops.com. Would love to connect

with you. You can find me on LinkedIn.

I'm Josh at Aerops. Um, we're going to

do this again in a couple weeks. I have

a a new report, Crystal. I need to drop

this over to you. It's called uh the

state of content teams report. Surveyed

a whole bunch of SEO and content teams

just about like how they're feeling

about things. Um, so it should be uh a

really good release coming out on

Thursday.

Exciting. Yeah, should be great. Um,

thanks everybody. We will see you in a

couple weeks. Thank you Crystal. Um,

appreciate your time. This was amazing.

Thank you. And thanks to everyone for

being so kind in the in the uh chat as

well. Thank you. Yes, lots of good

stuff. Thanks everyone. See you soon.