# SEO Rockstars 2026: Day 1 - Shaun Mitchell

**Video URL:** https://www.youtube.com/watch?v=Dapzc7ft9Z4

---

## Transcript

**[00:04]** Hey guys,

**[00:07]** uh I don't know if you saw my last uh

**[00:09]** talk that I gave. I was talking about AI

**[00:12]** uh guard rails and stuff like that. This

**[00:15]** is going to kind of uh continue on that

**[00:17]** theme. Um, and so what I what I want to

**[00:21]** go over is basically some uses of AI

**[00:24]** with very complicated systems to make

**[00:26]** them actually easy and really kind of

**[00:30]** how we

**[00:32]** radically changed up um a a lot of stuff

**[00:36]** in our agency and are we're getting

**[00:39]** better content um than ever. Like so the

**[00:43]** thing that I keep seeing is a lot of

**[00:45]** people are just like AI slopping. so

**[00:47]** much. It's just, hey, we're just going

**[00:49]** to mass page this or hey, we're just

**[00:52]** going to kick kick out a ton of freaking

**[00:53]** AI content, you know? Oh, we're going to

**[00:56]** throw in maybe like a little bit to, you

**[00:58]** know, try and improve it. Oh, we're

**[00:59]** going to say, hey, have some listicles

**[01:01]** in it or some crap. No. Um, so our

**[01:05]** victim for the day, and I do want to say

**[01:07]** I live here in Dallas. So all of you SEO

**[01:10]** testers,

**[01:13]** doing all your tests, uh, this is not a

**[01:16]** client, but what I did is, uh, you're

**[01:18]** going to see some documents, they are

**[01:20]** from my clients, and I just flipped the

**[01:22]** name out to make it easier. Um, and it's

**[01:26]** So I picked this, uh, guy over House

**[01:29]** Plumbing, who says he's in Keller.

**[01:34]** But wait, Google says he's in Fort

**[01:36]** Worth. So, we're dealing with the real

**[01:38]** genius over here. Just to start things

**[01:40]** off real quick. Um, and really like if

**[01:43]** you look at this site now, I mean, not

**[01:46]** being super critical, it's not terrible

**[01:49]** to, you know, you start looking around

**[01:51]** and seeing that the menu is not, you

**[01:53]** know, it's jacked. There's a whole bunch

**[01:55]** of stuff I discovered on this. But um so

**[01:59]** this is who we're going to kind of look

**[02:01]** at and house plumbing. Uh this is their

**[02:04]** homepage. We're going to pick on a a

**[02:07]** just wellthoughtout SEO strategy page. A

**[02:11]** water heater page. Wait, a water heaters

**[02:14]** water heater dash repair. So like we got

**[02:18]** a real SEO strategist working on on this

**[02:21]** page.

**[02:22]** What

**[02:24]** I have always you guys will constantly

**[02:27]** hear in SEO, oh this old thing that used

**[02:29]** to work is working again. You know, you

**[02:32]** hear that a lot with the parasites.

**[02:34]** Well, the parasites never went um away.

**[02:36]** And one of the things that like has me

**[02:38]** concerned as an SEO is uh you guys are

**[02:42]** like it almost seems like the wanting to

**[02:44]** write all the content that's topical. I

**[02:46]** completely agree with the topical

**[02:48]** because I'm in the brand camp. I'm on

**[02:51]** SEO everything. Uh, but like if you're

**[02:54]** only doing like topical without doing

**[02:56]** foundational SEO,

**[02:58]** I think you're going to have issues,

**[03:00]** especially when we're dealing with like

**[03:01]** a money page. And this is right here.

**[03:03]** This is a money page. The best SEO tool

**[03:07]** um for onsite is Kora. Is everybody in

**[03:10]** agreement on that?

**[03:11]** >> Is every is there anybody that's never

**[03:13]** seen a Kora report?

**[03:16]** Okay. All right. So a Kora report uh

**[03:20]** this is Kora. Okay. The way Kora

**[03:23]** operates is it's basically doing

**[03:25]** statistical analysis based off of

**[03:27]** keywords that you've put in and then you

**[03:30]** select um your competitors and you put

**[03:33]** in who you want. It's scraping off of

**[03:36]** their sites and every little piece of

**[03:39]** that code it's scraping to do a

**[03:41]** statistical analysis. And then Ted's an

**[03:44]** uber genius and he's figured out all

**[03:46]** these ranking factors and it spits those

**[03:48]** out. So a core report takes a while and

**[03:51]** uh for those of us who have done core

**[03:53]** reports know they are heavily headache

**[03:56]** inducing.

**[03:58]** Okay. This is not something

**[04:01]** you can really just take this and in

**[04:04]** general hand it to a VA. Would you guys

**[04:06]** agree? Okay. I'm going to change that

**[04:09]** right now. Okay. So, I've come up with a

**[04:11]** system that with basically with a core

**[04:15]** report, you can easily throw this in uh

**[04:18]** show a a VA how to use the core system

**[04:21]** and in a heartbeat, you can trust your

**[04:24]** dumbest VA to crank out really freaking

**[04:27]** good content. But to get that content,

**[04:29]** we need guardrails, right? Who here is

**[04:33]** in uh local?

**[04:36]** Okay. Raise your hand if you're in local

**[04:38]** and you're getting intake forms for your

**[04:41]** clients.

**[04:43]** Okay. Very, very good. Okay. Is there

**[04:46]** anybody that's not getting an intake

**[04:48]** form?

**[04:49]** Okay. So, with one of the things that I

**[04:52]** discovered is that I intake forms

**[04:56]** completely drastically change your your

**[05:01]** content. Okay? Because the thing that

**[05:04]** Google is looking for is unique content.

**[05:07]** When you're creating a cloud project and

**[05:09]** you are adding a pricing grid, an intake

**[05:12]** form, um, and a and a site map, and we

**[05:16]** could keep going on and on and adding

**[05:17]** more things. The content is unique to

**[05:21]** that person and it indexes so much

**[05:24]** better. Uh, it ranks so much better and

**[05:27]** it is that helpful content that Google

**[05:29]** is looking for. Are you guys getting

**[05:31]** pricing grids from your clients?

**[05:34]** >> You are getting pricing grids.

**[05:36]** >> We have a couple of clients that send

**[05:38]** out a pricing.

**[05:39]** >> Okay. Uh is anybody else getting pricing

**[05:42]** grids?

**[05:43]** >> Pricing grids.

**[05:44]** >> Yeah. Grids. Yeah.

**[05:46]** >> Yeah. Here. Simple pricing grid. So,

**[05:48]** this is a this it's this is my actual

**[05:51]** client. It's flipped out with House

**[05:53]** Plumbing. So, he came up with a pricing

**[05:56]** grid um and then I ran it through the

**[05:58]** system. He sent it over in a PDF. You

**[06:00]** just tell Claude just turn it into

**[06:02]** structured data. Boom. But what's great

**[06:05]** is is now I have service price range and

**[06:08]** uh time range for how long it is. So any

**[06:11]** content that I create now uh I just tell

**[06:15]** Claude to refer over to this. Okay. My

**[06:18]** intake forms are let me see if this the

**[06:23]** is the naked one. Um,

**[06:26]** typically the uh the t that if I don't

**[06:31]** do this with a client, it usually takes

**[06:33]** them two weeks to get this back to me

**[06:35]** because this is a lot of freaking

**[06:38]** information. It's huge because when we

**[06:41]** look at what a filled out intake form

**[06:45]** looks like

**[06:48]** and then a lot of the I do a lot of

**[06:51]** multilocation local so this kind of

**[06:53]** changes up. So,

**[06:56]** I had a uh one client I sent over the

**[07:00]** intake form. I didn't even ask for this.

**[07:01]** He sent over every single service and

**[07:04]** then what brand of tool he uses for

**[07:06]** that.

**[07:07]** That's going to be some damn good

**[07:09]** content. Really freaking good. So, not

**[07:12]** only were we using this on the money

**[07:13]** pages, um I follow Marino's methods for

**[07:16]** blogging. So, we do a lot of PAAs. All

**[07:19]** these same files on every single PAA.

**[07:22]** All of them. So, when we're blogging,

**[07:24]** we're putting in those pricing grids.

**[07:25]** We're putting in in the stuff from the

**[07:27]** intake form, what we do, which brands we

**[07:30]** work with, referrals over to the uh if

**[07:33]** they're using REM, why not link over to

**[07:35]** the RE actual water heaters that they're

**[07:37]** doing, include that into your intake

**[07:39]** form. The more that you put into an

**[07:41]** intake form, I think the better that

**[07:44]** your content is going to rank because

**[07:46]** it's going to be more informative, more

**[07:48]** helpful. There's just so many different

**[07:50]** freaking angles that you can go into.

**[07:53]** This is not actually really kind of like

**[07:55]** I typically when I do geos, I like to do

**[07:58]** geos on the intake form. So you have a

**[07:59]** tier one. These are your top ones. These

**[08:01]** are your your primary and then maybe

**[08:03]** your next door. Then I have my tier two

**[08:05]** and my tier three. You know, like here

**[08:07]** in Dallas, you should say, "Hey, you

**[08:09]** know, I get a client um that's in Fort

**[08:11]** Worth, but says, "Hey, I want to appear

**[08:13]** in Dallas." I'd be like, "Well, you're

**[08:15]** good luck with that. That's just not

**[08:17]** gonna happen." So you got to keep the

**[08:19]** the the geos um

**[08:22]** uh within reason.

**[08:25]** So

**[08:27]** all of that gets paired together with a

**[08:30]** core form. And so with the Kora we have

**[08:34]** up here we have all of the what's called

**[08:38]** variations. Okay. So what you do with

**[08:41]** your variations, you're going to simply

**[08:42]** just take that and you're going to pop

**[08:45]** it into a simple notepad and proof them.

**[08:48]** Okay. So, once you gone through there,

**[08:51]** you're going to put it into a form. And

**[08:54]** then you see how I have all of this

**[08:56]** stuff that is just

**[08:59]** giving you a headache. It's a ton of

**[09:02]** freaking crap. Okay? It's too much.

**[09:05]** Okay? So, I'm going to show you the

**[09:08]** basic. You can go way farther than what

**[09:10]** I'm about to do. Um, I have just created

**[09:14]** a simple prompt.

**[09:17]** I need you to write content to improve a

**[09:19]** service page. So, you change this up if

**[09:21]** you're writing a new page. Does that

**[09:23]** make sense? Okay. Uh, we have some

**[09:25]** content already. I gave them the link.

**[09:27]** We want a full rewrite of this content.

**[09:29]** Please scan the page first. Using this

**[09:32]** as an H1, okay? Because I designate what

**[09:35]** I want my H1 is. The rest of it, I could

**[09:37]** care less. Okay. Uh, the theme of the

**[09:40]** page. Uh, use semantic ICO. This is

**[09:43]** where you can get into your prompting.

**[09:45]** It doesn't really like matter what you

**[09:47]** guys are putting in. Um, if you want to

**[09:50]** put in, you know, semantic triples or

**[09:52]** or, you know, things about u bird or

**[09:56]** whatever, um, you can put those those

**[09:59]** things in there. I do designate because

**[10:02]** this is for Keller. I only want it for

**[10:04]** Keller, Texas. Okay? But Keller, Texas

**[10:06]** has little like neighborhoods and

**[10:08]** everything. That neighborhood content is

**[10:10]** extremely important. Always put

**[10:13]** neighborhood content when you're writing

**[10:15]** money pages. Okay? Even just doing your

**[10:18]** blogs, it's good to even have that

**[10:19]** neighborhood content on there. And you

**[10:21]** simply just need to tell um uh the uh AI

**[10:27]** that you want that neighborhood content.

**[10:29]** Now, another thing is we have banned

**[10:31]** chat GPT in our agency.

**[10:34]** We don't allow its use for writing or

**[10:37]** anything. It's utter [ __ ]

**[10:40]** And so we we completely just we don't

**[10:42]** even allow like simple lookups of things

**[10:44]** anymore in chat GPT. Uh I have such a

**[10:47]** low opinion of it.

**[10:48]** >> We use Claude, we use Perplexity. Uh

**[10:51]** sometimes we mess around with Manis. Um

**[10:54]** the only time I'm ever using a chat GPT

**[10:57]** product would be if I if I'm by coding

**[11:00]** and I'm checking security. So then I'm

**[11:02]** actually using codeex. So like regular

**[11:05]** chat GPT, nope, not using it. I've just

**[11:08]** seen too many hallucinations. So much

**[11:11]** like wording is just so awful. It's such

**[11:14]** a night and day difference between

**[11:15]** claude and and and chatt you can get

**[11:19]** into like local models if you're using

**[11:21]** some local models and stuff like that.

**[11:22]** There's lots of great freaking local

**[11:24]** models, but if you're using a Frontier

**[11:26]** Cloud's definitely where you got to be.

**[11:28]** Um so all right. So, um,

**[11:32]** >> yeah,

**[11:41]** >> never on the not never never on the

**[11:42]** money page.

**[11:47]** >> Do you have a prop that will pull the

**[11:50]** any sort of external authority like

**[11:52]** something about plumbing codes?

**[11:53]** >> Oh, so the It's funny you say that. The

**[11:57]** Can this hear me still? the the client

**[11:59]** that this is uh uh actually taken from

**[12:03]** his intake form does have that because

**[12:04]** he's big into uh water filtration.

**[12:08]** And there is a site out there that is a

**[12:11]** public site um by I I don't know if it's

**[12:15]** a nonprofit or if it's the government.

**[12:17]** They have all the contaminants and

**[12:19]** carcinogenics and everything in the

**[12:20]** water. We went psycho on that content.

**[12:24]** We pulled a ton and basically we were

**[12:26]** running it in for for each single city

**[12:29]** and we ran that and then ran it uh

**[12:32]** through claw because it read it's a

**[12:34]** scient too scientific. It's terrible and

**[12:37]** so we ran that through claw to make it

**[12:39]** like read better and then we also did a

**[12:41]** refer out link to that.

**[12:43]** >> Yeah.

**[12:48]** >> Yeah. Yeah. It's their their water

**[12:51]** filtration uh SEO is improved by quite a

**[12:55]** bit. So, but I try and keep it um I mean

**[13:01]** you can have a page that is hey we

**[13:04]** service these areas and that's a

**[13:05]** different page but we're talking about

**[13:07]** water heater repair Keller here. So,

**[13:09]** we're not we're going to stay true to

**[13:10]** the entity and we're not going to me

**[13:12]** mention any sing I don't even I don't

**[13:14]** even want DFW on the page. Would you

**[13:17]** link up to say something ongo?

**[13:19]** >> Yes. Yeah. Yeah, we actually do the it

**[13:22]** will on their site. Yes. This claude you

**[13:26]** can do a simple prompt. Um when I was

**[13:29]** building their site uh three four months

**[13:31]** ago just a simple prompt to get external

**[13:34]** stuff it had that all through the

**[13:36]** content that we were writing.

**[13:38]** >> Yeah. And it took no effort. I just off

**[13:40]** a whim asked it to do that and we we

**[13:43]** discovered oh wow that's really awesome.

**[13:45]** and it went to town on that. So, it it

**[13:48]** really is a lot of playing around with

**[13:51]** your prompt to get what you want. Uh the

**[13:54]** main thing, you know, that I'm wanting

**[13:56]** to show you guys is is that you can use

**[13:58]** use Kora in a very simple way u to to

**[14:02]** write a page now very very quickly. All

**[14:05]** right. So, um you notice I highlighted

**[14:09]** refers over to the intake form, the

**[14:10]** pricing grid, and this project if

**[14:13]** needed. So, I'm just making suggestions.

**[14:15]** Also include neighborhood content. So,

**[14:18]** this I'm not writing a book about like,

**[14:20]** hey, I need all this neighborhood stuff.

**[14:21]** And I'm making just like a basic

**[14:23]** suggestion. Hey, if you feel like it,

**[14:26]** check out the pricing grid. If you

**[14:28]** don't, cool. You know,

**[14:31]** um, okay. So, I'm going to supply some

**[14:33]** important uh info from a technical SEO

**[14:36]** document. Okay. So, I've got the

**[14:38]** variations. Okay. And you can see that

**[14:41]** basically

**[14:43]** all I did was literally go in to the

**[14:47]** Excel spreadsheet

**[14:49]** and I went through here and I looked for

**[14:52]** things that just said variations.

**[14:55]** That's it.

**[14:58]** VA can do this.

**[15:00]** Two years ago, VA could not do this. I

**[15:03]** mean, well, preI. Um,

**[15:08]** so the next thing is we go and now we

**[15:11]** have our entities. Um, are you guys is

**[15:15]** anyone here not familiar with entities?

**[15:18]** Shut up.

**[15:22]** Uh, in case you guys didn't know, Simon

**[15:24]** is probably like one of the best people

**[15:26]** when it comes to like language,

**[15:27]** understanding language. Um, he does

**[15:30]** crazy things with just like how good he

**[15:32]** can actually write a page. Uh he did a

**[15:36]** page in he um in Plano and somehow got

**[15:41]** that within what couple weeks or was it

**[15:44]** a week into Dallas. Um and I was like

**[15:49]** how in the hell and he did no links. It

**[15:51]** was just entities and how the language

**[15:54]** and the prompts of and it was the

**[15:57]** uniqueness. That's what I'm trying to

**[15:58]** get to is that it's the uniqueness of

**[16:00]** the content that you're providing. So,

**[16:02]** we're getting that uh by doing the

**[16:05]** outsour the the outside sources such as,

**[16:08]** you know, the city of Keller or things

**[16:10]** like that. We're doing that by doing a

**[16:12]** very insanely detailed intake form. Go

**[16:15]** psycho on your intake form. Um, pricing,

**[16:19]** you guys know that Google's eating that

**[16:21]** up. Is everybody here mess around with

**[16:23]** the PAAAS? Are you guys all familiar

**[16:25]** with that? The first one's always

**[16:27]** freaking cost. I mean, come on. So, you

**[16:30]** call the client up, you're like, "Dude,

**[16:32]** I got this really sucks. Google really

**[16:35]** wants you to start talking about

**[16:37]** pricing." He's like, "I don't agree with

**[16:39]** it. I think it's stupid." He's like,

**[16:41]** "Yeah, it is stupid, but

**[16:44]** I see it's ranking, so I know it's

**[16:46]** stupid, but what if we were to do

**[16:48]** something like I get a range?"

**[16:51]** H, now this conversation like with a

**[16:54]** client could literally take a couple

**[16:56]** months. Just keep the conversation every

**[16:58]** time you talk to the client. Keep

**[17:00]** bringing it up. You'll eventually wear

**[17:02]** them out. They'll send over a range.

**[17:04]** >> Okay. So, I'm paranoid when it comes to

**[17:08]** as a lawyer.

**[17:09]** >> Okay. Texas is messed up because they

**[17:11]** because the state bar keeps messing with

**[17:13]** your guys' websites. Yeah. It's

**[17:16]** >> But I think we can say it, right?

**[17:18]** >> I don't know. As far as attorneys,

**[17:24]** >> like if I'm doing a DWI, well, there's a

**[17:26]** whole different gamut. It could be a

**[17:28]** low-level DWI, could be a high level.

**[17:30]** So,

**[17:32]** >> put the ethics part side. I can deal

**[17:34]** with that, right?

**[17:35]** >> Are you saying that

**[17:37]** >> if you the I mean, what I'm saying is is

**[17:41]** I know how bad like the state of Texas

**[17:43]** is on lawyers. They're absolutely awful.

**[17:46]** Um,

**[17:49]** but on the other side, Google really

**[17:50]** wants you to talk about price.

**[17:52]** >> So,

**[17:52]** >> so you got to find that in between.

**[17:54]** >> The gamut would be I don't want to lo I

**[17:58]** don't want to be tied down to it. I want

**[17:59]** somebody to say, well, you put X, but

**[18:02]** Google might be okay with

**[18:04]** >> range. Yeah, I'm doing ranges. Um, they

**[18:09]** would prefer X. Uh but uh the range

**[18:13]** seems to be working uh fairly well.

**[18:16]** >> How about start?

**[18:19]** >> That's a that

**[18:21]** I I think I you can do that. But then I

**[18:25]** think some of the owners would even push

**[18:27]** back because that's a singular price.

**[18:29]** They know that once some yahoo gets that

**[18:31]** in their head, they're going to be like,

**[18:32]** "Well, it said $100, you know." So, this

**[18:36]** is a this is an insanely hard

**[18:38]** conversation to have with your clients

**[18:40]** trying to get them to put prices in

**[18:42]** their blogs and in their content.

**[18:43]** >> How many people?

**[18:46]** >> I don't even SEO my site. I live out in

**[18:48]** the sticks.

**[18:52]** >> Yeah. So,

**[18:53]** >> we had this conversation internally

**[18:55]** about putting or having pricing

**[18:58]** available for certain clients but like

**[19:02]** not having it page and like when we're

**[19:06]** talking to a client in person putting a

**[19:09]** QR code on something so they could

**[19:12]** goirectly.

**[19:12]** >> Do you mean a orphan or do you mean a no

**[19:14]** index page? That's right. You would

**[19:16]** almost have to know index it.

**[19:18]** >> Well, I assume if you no index it,

**[19:21]** Google is going to should ignore it.

**[19:24]** >> Should but you know how bad the AI bots

**[19:26]** are right now.

**[19:27]** >> Yeah.

**[19:29]** But I'm just wondering like if it

**[19:31]** doesn't if you create this page in a

**[19:33]** blog post, people can navigate to it and

**[19:36]** they can see it, but it can put it in a

**[19:38]** place that doesn't link from anywhere.

**[19:39]** >> Google is so hungry to display price.

**[19:43]** They're psychotic.

**[19:44]** >> So they'll still pull it up.

**[19:46]** >> They will I would almost guarantee

**[19:47]** they're going to find it. I mean, it's

**[19:50]** Yeah, they they love the pricing grid. I

**[19:54]** mean, it's the easiest way, easiest

**[19:55]** thing to hit in in SEO is the pricing

**[19:58]** grid because nobody wants to do it. They

**[20:00]** don't want to be first and the owners

**[20:02]** hate it, so they're not doing it. So,

**[20:04]** it's everybody's in a stalemate. So, the

**[20:05]** second somebody throws out a pricing

**[20:07]** grid, they just go freaking nuts.

**[20:09]** >> We get ranked because our our surgeons

**[20:11]** won't put their price out.

**[20:12]** >> Yeah.

**[20:13]** >> So, we we have a one sense that seems to

**[20:16]** work great and it's, you know, the

**[20:17]** average cost of LASIC in city name gives

**[20:20]** between X and Y,

**[20:22]** >> right? And we rank like for every city

**[20:25]** we've done that in if you search cost of

**[20:27]** basic city name they rank for that and

**[20:30]** then in the article explained it could

**[20:32]** vary for you coming

**[20:33]** >> that is exactly what we are doing that's

**[20:35]** that's the method that I think you can

**[20:38]** that so I think you start off on the

**[20:40]** hard sale start off on the hard sale

**[20:42]** with your clients on hey they want

**[20:44]** pricing and he's going to him and haw

**[20:46]** and then the next conversation you go

**[20:49]** what about range I'm seeing a lot of

**[20:51]** people doing the range

**[20:52]** That's what you I like the range better.

**[20:55]** I think it's easier fit. I mean, you got

**[20:58]** to put yourself in their shoes. They

**[21:00]** don't want to do it just X. I think I

**[21:03]** think the range is a lot better. And

**[21:05]** then, you know, it's subject to change,

**[21:07]** you know, and stuff like that. You know,

**[21:08]** I got a garage door client. They can't

**[21:10]** even do X, you know, because it's the

**[21:13]** depth of the of the um of the garage and

**[21:16]** there's a ton of freaking like factors

**[21:18]** on a simple freaking garage door, let

**[21:20]** alone a freaking like putting in a water

**[21:22]** heater or slab leak. I mean, who

**[21:24]** freaking knows how much a slab leak? I

**[21:26]** there's you don't know until you get out

**[21:29]** there, but Google wants to know and

**[21:32]** they'll rank you and then hopefully the

**[21:34]** client can get out there.

**[21:37]** Have you tried putting pricing in like a

**[21:39]** zero toggle or like an

**[21:45]** >> okay so that brings up a another thing

**[21:49]** there is uh un I don't know if you guys

**[21:51]** are aware of some of the controversy

**[21:53]** going around accordians

**[21:55]** there's there's a lot of issues with

**[21:57]** some some accordians you have to look at

**[21:59]** how the accordion is coded it is

**[22:01]** extremely important because uh some of

**[22:04]** the accordians the way they are built

**[22:07]** the uh AIS are not actually reading

**[22:09]** them.

**[22:15]** >> Yeah. Yeah. It's something to do with

**[22:17]** the rendering and how how it's built

**[22:19]** out. Uh accordians are um in fact I've

**[22:23]** seen other elements even the way they've

**[22:26]** been built by specific builders that the

**[22:29]** AI won't pick up on them.

**[22:33]** >> Gemini analyze one of my pages and

**[22:37]** just pointed that out that the accordion

**[22:39]** they wanted to showing the answers.

**[22:41]** >> And what builder were you using?

**[22:42]** >> Um that one it was break dance.

**[22:45]** >> Break dance.

**[22:47]** Yeah. Um and some people think I'm like

**[22:50]** psychotic, but I like Divvy. Divvy has a

**[22:53]** lot of issues. However, Divy Divvy 5 is

**[22:56]** about to be released and it's a

**[22:58]** breakthrough. Um we're looking at this

**[23:00]** core report. You guys know inside this

**[23:02]** Corora report, uh Ted's always been

**[23:04]** asking that even in the div tags and the

**[23:06]** style tags, hey, stuff some keywords in

**[23:09]** that stuff. You can't do that with with

**[23:12]** the other builders out there. Divvy 5,

**[23:14]** you can you can actually change the div

**[23:16]** tags to stuff them with keywords and

**[23:18]** stuff like that now. So, I mean, there's

**[23:20]** a lot of change going on.

**[23:23]** >> We talked about this in the room last

**[23:25]** night for anybody that wasn't there. Uh

**[23:27]** when you add a service in the back of

**[23:29]** the GMBB, it asks you do you want to

**[23:33]** list it as free

**[23:35]** uh range or a price, right? Or or none.

**[23:40]** I've been using the range for ever even

**[23:42]** though I didn't want to, right? It was

**[23:45]** something that my VA did years ago by

**[23:47]** mistake and absolutely crushes it

**[23:49]** because it sends the justice case and

**[23:51]** Google picks it up. So, even if you

**[23:53]** don't want to mention it on your

**[23:55]** website, put it on the GMV and then ask

**[23:58]** your client, right? This is my

**[23:59]** suggestion. Hey, what's the minimum

**[24:02]** amount you'll do this type of work for?

**[24:05]** >> 500 or there's your minimum. What's the

**[24:08]** maximum it'll cost? You don't have to

**[24:09]** put that number so you don't scare

**[24:11]** people away. Part two, because of the

**[24:13]** legality, Sean, right? Do not bait and

**[24:17]** switch. If you charge $1,000, don't put

**[24:20]** the [ __ ] range as 200 to,200 because

**[24:23]** you're lying and you're playing yourself

**[24:24]** and Google's gonna

**[24:25]** >> I mean the main thing to stress is I if

**[24:28]** you start adding price, you have no idea

**[24:29]** how much easier life is going to be.

**[24:32]** Okay, now here's the third step. So we

**[24:34]** went through hard cell, then we went to

**[24:36]** range, then we talked about the GMBB.

**[24:39]** Okay, then we the final step you have

**[24:41]** with the client is it doesn't freaking

**[24:44]** matter anymore. It does. I don't [ __ ]

**[24:46]** care what you say. Guess what? When

**[24:49]** somebody leaves a review now, they're

**[24:52]** leaving price. So whether you want to

**[24:54]** freaking leave talk about price or not,

**[24:58]** the way Google reviews work now, they're

**[25:00]** putting the damn price, dude. So we

**[25:02]** better get a hold of this thing, get in

**[25:04]** front of it, and we say what our range

**[25:07]** is before we let the Google Google

**[25:09]** reviews manipulate the pricing.

**[25:12]** >> To that very point, if someone does

**[25:14]** leave a review with It's manipulating

**[25:16]** stuff 100%.

**[25:18]** >> I mean, so I pay $600,

**[25:21]** whatever it was,

**[25:21]** >> right? If you copy paste the review and

**[25:23]** have an external link in a quote markup,

**[25:25]** let's say in WordPress, whatever it is,

**[25:26]** and you're just happy customer review

**[25:28]** about water heater and Keller, you leave

**[25:30]** a review, you're not even saying the

**[25:32]** price, but could that be readable if

**[25:33]** it's just copy paste text with an

**[25:34]** outbound link back to the individual URL

**[25:36]** in a markup post because you're quoting

**[25:38]** someone else's review citing the

**[25:40]** original source. Could you at least pick

**[25:42]** up that price point at that point

**[25:43]** without really revealing price?

**[25:45]** >> Yeah, you could do something like that.

**[25:47]** >> Is that is that a if you have it, is

**[25:49]** that at least a workaround in the short

**[25:50]** term? Yeah, I mean you could you could

**[25:52]** do that. The the whole the whole taking

**[25:56]** of a Google review and using it

**[26:00]** there's weirdness around that too,

**[26:02]** right, Mike?

**[26:03]** >> Yeah,

**[26:04]** >> there's there I would be careful about

**[26:06]** doing that.

**[26:07]** >> Um,

**[26:09]** >> so I currently I I use trust trust

**[26:12]** review as a big plugin, but like as far

**[26:16]** as I know, the AIS can't read it. But

**[26:18]** then I was talking to another guy and

**[26:20]** he's like, "No, I pulled it up the other

**[26:21]** day. The AI Reddit." So, I don't know

**[26:24]** what's going on with that plugin. I

**[26:26]** think it's one of the best looking

**[26:27]** plugins. I love the plugin. It's a great

**[26:29]** way to display the reviews. Uh, it's

**[26:31]** cheap, especially when you're an agency

**[26:33]** because I I love the thing, but the

**[26:36]** problem is is that the AI can't read it.

**[26:39]** What we're personally doing is doing

**[26:40]** some extra reviews in the in the

**[26:42]** homepage that is just pure just HTML.

**[26:45]** So, we know that like Google is seeing

**[26:48]** reviews because that's in the helpful

**[26:50]** content. And then the problem is when

**[26:52]** you run uh DG's helpful content system,

**[26:55]** it's not reading that trust review

**[26:56]** thing. I haven't got it able to read it.

**[26:59]** So, it doesn't exist. Those reviews on

**[27:01]** your site don't exist. If you're using

**[27:03]** trust

**[27:03]** >> reviews on your home aren't on the GMBB

**[27:07]** link maybe to read more.

**[27:09]** >> Yeah, we're we are doing that because

**[27:11]** it's it that way we are hitting the

**[27:12]** helpful content update. Um so um all

**[27:16]** right so back to uh all right so

**[27:19]** everybody's familiar with entities um

**[27:21]** this is a very important step in the

**[27:23]** Kora uh because the problem with um Kora

**[27:27]** is it's badass when it comes to entities

**[27:30]** but you have to really understand um

**[27:34]** topic and what an entity should be okay

**[27:37]** so this guy is in uh Keller who thinks

**[27:41]** that DFW area should be on that page

**[27:46]** Good. No. Um, so it's throwing a whole

**[27:51]** bunch of this stuff out. In fact, okay,

**[27:54]** we have a competitor, Milestone

**[27:56]** Electric. Um, we got FAQ. We have

**[28:00]** Dallas. We obviously know that shouldn't

**[28:01]** be on there. Um, but then we also have

**[28:04]** things that are kind of like leak

**[28:07]** detection. Well,

**[28:11]** well, gray.

**[28:14]** Um, heating. Well, heating definitely

**[28:16]** needs to be on the page because it's

**[28:17]** water heating. Uh, noise.

**[28:22]** Um, shut off valve. There's going to be

**[28:26]** >> Yeah.

**[28:28]** >> All right. Entity master.

**[28:32]** >> Yep.

**[28:35]** Maybe DFW. So careful about how you say

**[28:39]** we service DFW or DFW. I might mean we

**[28:43]** service the airport. So if you care what

**[28:46]** >> Yeah. So like my me and Simon talk like

**[28:50]** every few weeks or so. We're very close.

**[28:52]** I remember what is it was a couple years

**[28:54]** ago I was working on a client. Okay. And

**[28:58]** uh my VA had written the the content and

**[29:01]** they were a plumber in Plano and it said

**[29:03]** located in the heart of plumber. Simon

**[29:05]** saw that and said what the [ __ ] are you

**[29:07]** doing? Is this is this a heart surgeon?

**[29:10]** What are you doing? Don't have heart

**[29:13]** there. It's like what? It's just that's

**[29:15]** just that's how people No, that's how

**[29:18]** big of a deal entities are. You need to

**[29:21]** like trim the fat and be very precise

**[29:24]** with with the words that you use on the

**[29:26]** page. That's why Simon is is a language

**[29:28]** master and he can do an just do a page,

**[29:32]** no backlinks, nothing and rank it like

**[29:34]** freaking crazy and like Dallas.

**[29:37]** So, garbage disposal. So, you see in

**[29:40]** this it's picking up this stuff because

**[29:43]** some of the competitors have created

**[29:46]** water heater page obviously with garbage

**[29:48]** disposal and water supply and all that

**[29:50]** crap on there. Okay. So you ne you need

**[29:52]** to pair down that list simply to go

**[29:55]** through the notepad and

**[29:57]** uh get rid of it. And so I paired the

**[30:00]** list down quite a bit.

**[30:03]** And then it's simple. We go and we get

**[30:06]** the entities. Okay. So this was a pretty

**[30:11]** easy one. Usually a core report's going

**[30:13]** to have a lot more of this crap than

**[30:16]** this one. Uh but this page really

**[30:19]** honestly it wasn't terrible.

**[30:23]** Two questions. One, I don't know if you

**[30:25]** know the answer, but can you tell me

**[30:27]** what that um those those two Can you go

**[30:30]** back to the core page, please?

**[30:31]** >> Yep.

**[30:35]** >> Do you want to go to the home or the

**[30:36]** entity page?

**[30:37]** >> This is the one. So, I wasn't confused

**[30:40]** about what relevance and confidence

**[30:41]** means. That's the first question. And

**[30:43]** then the second one is when you do your

**[30:46]** knows about schema, do you tend to use

**[30:48]** the wiki link for that? I use Wiki and I

**[30:52]** use Productology.

**[30:56]** >> Yep. That's old school stuff and I still

**[30:59]** use it. Yeah. Don't just use Wikipedia.

**[31:02]** Use Productology also. Any other hot

**[31:05]** tips on that subject, Simon?

**[31:07]** >> Wiki data.

**[31:08]** >> Oh, yep. That one too. I mean, you're

**[31:12]** when you're doing schema, your schema

**[31:14]** can be way more content than your actual

**[31:16]** content of the page.

**[31:17]** >> So, it can get on the recording. Oh,

**[31:19]** >> Simon said what the question was.

**[31:22]** >> Oh, wiki data. Simon also said uh wiki

**[31:25]** data. So we have wiki data, we have

**[31:27]** regular Wikipedia and also

**[31:29]** producttologology for schema.

**[31:34]** Um and I would venture to guess that

**[31:37]** Graedia

**[31:39]** I'm sure people are already out there

**[31:40]** testing something like that.

**[31:43]** Um everybody been on Graipedia yet?

**[31:47]** It's It's pretty good. Uh it's pretty

**[31:50]** good resources.

**[31:52]** Um okay. So once you get the entities

**[31:56]** and you've got the entities, you'll

**[31:58]** notice that there's some like cleanup

**[32:00]** stuff. Uh and so I just write this is

**[32:03]** more info from the technical SEO

**[32:05]** document. So I'm right back to pasting.

**[32:09]** So, number of heading tags, 85. Number

**[32:11]** H3 tags, number of of uh word count, uh

**[32:16]** number of sentences, 49.

**[32:18]** Okay, so we have all that. Um the way I

**[32:25]** show a couple other things. So, I didn't

**[32:27]** have a site map. Um so, all I did was I

**[32:32]** wanted it to create a project. So I said

**[32:36]** I want to add the site map with all the

**[32:37]** pages this site has to the project. You

**[32:40]** can find it at that. And then it made a

**[32:45]** very nice

**[32:50]** markdown file.

**[32:52]** And look at how it organized it.

**[32:57]** There's no excuse not to have this on

**[32:59]** every writing content that you're doing.

**[33:02]** Okay? Whether you're blogging, you're

**[33:04]** doing money pages, I don't freaking

**[33:05]** care. This should be it built into every

**[33:08]** project. Um, I have a client that was

**[33:11]** taking literally three months to get an

**[33:13]** intake form. And so I did I said, "Fuck

**[33:15]** it." I gave it my blank form, showed it

**[33:19]** the blank form, and I and I gave them

**[33:21]** the website. It filled out the whole

**[33:23]** intake form, came up with a whole bunch

**[33:24]** of other [ __ ] got me ahead, and I sent

**[33:27]** it over to the client saying, "Hey, this

**[33:29]** is 90% there. I just need you to look

**[33:31]** over it and approve it." so that we can

**[33:32]** actually start running with it.

**[33:35]** And so that cut down on a lot of time.

**[33:37]** So we just figured out that trick not

**[33:39]** too long ago. So we'll probably actually

**[33:41]** be doing and filling out our intake

**[33:43]** forms for our clients for the most part

**[33:45]** and then just seeing, hey, send this

**[33:46]** over for approval and then be adding it

**[33:48]** to the project.

**[33:50]** Um what else?

**[33:55]** Uh,

**[33:59]** okay. So, once we did that, I paste

**[34:03]** pasted all that stuff in there

**[34:07]** and it spit out this page.

**[34:11]** So, all I did is I always do them in a

**[34:14]** word document. I would not do this in

**[34:16]** freaking uh uh Notepad uh because you

**[34:20]** need to like the tables and everything.

**[34:21]** So, always do these if you're if you're

**[34:23]** messing around with Kora. Um, you can

**[34:25]** use Pop also. Uh, POP's a really good

**[34:28]** product. It's just Kora has way more. I

**[34:29]** mean, if you're going to go bottles

**[34:30]** deep, might as well use Corora. Um, so

**[34:34]** it's literally you just copy all this,

**[34:36]** you throw it into the prompt.

**[34:38]** And then this took quite a while. It The

**[34:42]** great thing, another reason why I I

**[34:45]** stopped the U GPT is because I want to

**[34:49]** show you what happened when I threw this

**[34:50]** in.

**[34:53]** When I pasted that stuff,

**[34:57]** it came back

**[35:00]** and asked me

**[35:03]** all of these questions. Okay. What's

**[35:05]** your word count target

**[35:10]** brand?

**[35:12]** Okay. What's up with the FAQs? What are

**[35:15]** you talking about on triplicates? Call

**[35:16]** to action. I hate call to action. I I

**[35:18]** want to do my own call to action. I

**[35:20]** don't want freaking Claude doing my call

**[35:22]** to action. So I always tell them to not

**[35:24]** do that. Um and so um I answered that

**[35:28]** and it spit it spit this stuff out.

**[35:33]** So I answered uh let's look let's uh let

**[35:36]** loose write as much as you want. Two was

**[35:39]** yes uh primary but we want to work on

**[35:42]** all brands. So the question about ream

**[35:43]** is it said hey only talk about re

**[35:46]** by telling it and you setting up your

**[35:48]** cloud system you should have something

**[35:50]** inside of your cloud system that

**[35:52]** whenever you put something in it's still

**[35:53]** going to ask you another question to

**[35:54]** redefine what you're you're saying

**[35:58]** and then output raw HTML I only want the

**[36:01]** uh body

**[36:04]** and so we're left with this

**[36:09]** now this one spit out a little too much

**[36:11]** freaking

**[36:12]** uh content. This it went really

**[36:15]** overboard. We're at like 7500.

**[36:18]** Um but I should not have said let loose.

**[36:25]** Uh but um quite often like if I'm

**[36:28]** writing a new page, I will actually say

**[36:31]** let loose because there's going to be a

**[36:33]** whole bunch of stuff in there that I may

**[36:37]** not want. The other thing is I do not

**[36:40]** typically uh get into repetitive or

**[36:42]** redundancy in the first prompt. I do

**[36:45]** that in the second one. Okay. So, if I

**[36:48]** go in here and I say uh remove anything

**[36:50]** repetitive or redundant, this is

**[36:52]** instantly like from 7,000, it's going to

**[36:54]** like take paragraphs and a whole bunch

**[36:56]** of stuff. Have you guys done that in CLA

**[36:57]** before?

**[36:59]** >> Okay.

**[37:01]** >> One question. you have a software

**[37:04]** you would after you use the redundant

**[37:05]** and clean it up a little bit you'd be

**[37:07]** embedding YouTube videos or images or

**[37:10]** just to break it up because the average

**[37:11]** person is still going to want to see

**[37:14]** that broken up with the multime media

**[37:16]** shop in addition to header tags

**[37:17]** >> okay so I'm part of Merino's group I

**[37:21]** have and I do I do tons of freaking

**[37:24]** YouTube videos I have massive libraries

**[37:26]** for every single client we rank through

**[37:28]** YouTube uh we push maps I mean it's We

**[37:31]** have so much freaking video stuff that

**[37:33]** we do. Um, and it's it's done in

**[37:36]** Marino's specific way. So, what I would

**[37:38]** do is if I'm wanting to add in to

**[37:41]** YouTube, um, you would just add that

**[37:44]** into your project file. All this stuff

**[37:47]** should just keep getting added, keep

**[37:48]** making your project file bigger. If if

**[37:51]** let's say like your VA all he does is

**[37:53]** blogs, well, why not have like a file

**[37:56]** for like the HTML structure that you

**[37:58]** want, you know?

**[38:01]** you're using Elementor, you know, oh,

**[38:05]** you could do some like pre embeds and

**[38:07]** stuff like that. So, there's a lot of

**[38:10]** stuff that you can do. Uh,

**[38:13]** keep adding more. You know, I don't even

**[38:15]** have helpful the helpful content in the

**[38:17]** uh what's the other one? The uh

**[38:21]** >> huh?

**[38:24]** >> No. Uh quality evaluators guidelines and

**[38:28]** the uh helpful content. So this is

**[38:30]** Diggity uh came up with this last year

**[38:34]** and I I threw it in Claude. Um so this

**[38:37]** is basically his EAT system. So I could

**[38:41]** add this uh quality evaluator guideline.

**[38:44]** Um another thing if you want to go

**[38:46]** really freaking psycho.

**[38:51]** The I think the most impactful person

**[38:54]** for SEO last year was Sean Anderson uh

**[38:57]** from Hobo. he went through um a fine

**[39:00]** tooth comb over the Google lawsuit and

**[39:03]** the stuff the content that he's been

**[39:04]** putting out has been utterly amazing.

**[39:07]** There's so many things that you can take

**[39:10]** from the stuff that he's put out. This

**[39:12]** is his um 2025

**[39:15]** uh strategic uh plan. He's got tons of

**[39:22]** um where's I keep you

**[39:24]** >> his name?

**[39:25]** >> Uh Sean Anderson. on that document.

**[39:31]** >> Yeah. So, here's his I think this was

**[39:34]** his first big post was talking about the

**[39:37]** uh

**[39:40]** So, there's so much of this and there's

**[39:43]** ways that you can run this through

**[39:45]** claude and turn it into MD files also if

**[39:50]** you're getting bits. So, now you're

**[39:51]** giving it tips and tricks for SEO

**[39:54]** guidelines and things like that.

**[39:58]** So, if you have not had a chance to go

**[40:01]** through what Sean's put out, uh it's I

**[40:04]** would say it's it's critical. Um and

**[40:06]** there's tons of stuff that you can use

**[40:08]** from what he did to add it to your

**[40:10]** project files to write better content

**[40:12]** and review your SEO. So, um that's

**[40:17]** pretty much it. Um

**[40:23]** but this is just the way we are doing

**[40:25]** our blogs and our content and an easy

**[40:29]** way now with just using cloud to write

**[40:32]** um

**[40:34]** uh to use Kora because I mean it was a

**[40:37]** lost cause before. See all this the

**[40:39]** other thing is is this is all the got a

**[40:41]** lot of FAQs and um uh some people are

**[40:44]** still unaware that you can just do your

**[40:46]** schema in line. Um, I don't trust my my

**[40:50]** employees to do schema for my website,

**[40:52]** but I do trust them to do FAQ schema in

**[40:55]** line that Claude spit out. There's

**[40:57]** nothing wrong with that. That's easy

**[40:58]** stuff. Uh,

**[41:00]** >> you're going through all those entities

**[41:02]** yourself before you do your VA.

**[41:04]** >> Oh, I mean the V is you get you got to

**[41:08]** teach your VAS a little bit on like what

**[41:10]** entities and how laser focused it needs

**[41:12]** to be. Um, I haven't I have a I have one

**[41:16]** VA that's been with me for two years.

**[41:18]** His attention for detail is not the

**[41:20]** greatest

**[41:22]** and he still has no problem going

**[41:24]** through the list. I rarely catch him uh

**[41:26]** making a mistake. So,

**[41:29]** anything else?

**[41:32]** >> Uh quick real quick on entities on that

**[41:34]** what what you just were discussing. I

**[41:36]** personally until I trust whoever is

**[41:39]** going to do that and I've spent time

**[41:41]** with them on Zoom or they you know like

**[41:43]** it's one of my [ __ ] ninjas. Not

**[41:45]** happening bro. You do that because you

**[41:47]** [ __ ] up that entity report and start

**[41:49]** doing [ __ ] wrong.

**[41:50]** >> Entity pollution is

**[41:51]** >> pollution.

**[41:52]** >> Yeah. I would say almost every single

**[41:55]** website that you look at in the local

**[41:57]** space, every single one has entity

**[41:59]** pollution. They don't understand it how

**[42:02]** laser focused it has to be. It's rare

**[42:04]** that you ever come across a site that's

**[42:05]** not entity polluted.

**[42:09]** >> Are you finding that code or is putting

**[42:12]** in same multiple pieces of content and

**[42:15]** do you care for?

**[42:17]** >> So there are some there are some ways to

**[42:19]** get around that if you are having it and

**[42:21]** it's a little bit annoying. So the first

**[42:23]** thing that you can do if you keep seeing

**[42:25]** that the same FAQs are coming through um

**[42:28]** the lay you can do that by keeping the

**[42:31]** same chat context going.

**[42:34]** So I would continue writing um I would

**[42:37]** do blog one from here blog two and then

**[42:40]** if you simply just say you know don't

**[42:43]** make it redundant. Okay then you would

**[42:47]** within this context window you would cut

**[42:50]** that down. Okay, you could do another

**[42:52]** thing of uh sharing the FAQ and you

**[42:55]** could carry over to the next context

**[42:56]** window by sharing some of the FAQs that

**[42:58]** you have. Um, another one would be uh

**[43:01]** doing something like you're you're um

**[43:04]** training employees to actually use VS

**[43:06]** Code. That's going to keep you going a

**[43:08]** lot longer. Um, what I am showing you

**[43:11]** off of the cloud projects, yes, we do

**[43:14]** that, but we do a lot of stuff actually

**[43:17]** by by VS Code. So we write in VS code

**[43:20]** our content and that's why we get way

**[43:24]** more markdown documents and stuff going

**[43:26]** on way more there's a whole bunch of

**[43:29]** crap it's whole new world when you start

**[43:31]** using VS code

**[43:32]** >> so you write your content here and then

**[43:34]** you give it to Divy or then do you say

**[43:36]** and is this no

**[43:39]** >> so right so when they when when it

**[43:41]** outputs I'm always looking for like when

**[43:43]** I when the content gets done I'm only

**[43:45]** look for the stuff in the body and then

**[43:47]** the VA just takes that and puts that

**[43:49]** puts it in and then we make it look

**[43:50]** pretty and everything. So uh we have

**[43:52]** some uh we have some clients on

**[43:54]** elementor uh we have one client on a

**[43:57]** data um and then uh some most of our

**[44:01]** clients are on uh divvy five not four

**[44:04]** >> on divy will you go in and be sure divvy

**[44:07]** will you use on a local machine take the

**[44:10]** content from code put it local then push

**[44:13]** it or do you separate here's my content

**[44:16]** and then you have somebody make it

**[44:18]** pretty

**[44:19]** >> no they the VA our VAS do They make it

**[44:21]** pretty. They just taking it because it's

**[44:23]** in the um it's in the raw format. Um

**[44:27]** it's I mean it's not that Where is it?

**[44:31]** >> No, no, no. Yep. I I just didn't

**[44:33]** understand if you were taking it and

**[44:35]** then like having Claude look at the

**[44:37]** page.

**[44:38]** >> No, I'm not pushing it. It's an easy

**[44:40]** like it's really easy to push it from

**[44:42]** WordPress. It's very easy to do. We're

**[44:44]** just not at that stage. And the other

**[44:46]** thing is is we have a lot of different

**[44:47]** builders. uh some of the builders. Um

**[44:51]** well, for one, I mean, Divvy 5 is just

**[44:53]** coming out of beta.

**[44:55]** >> Sorry, would you do that?

**[44:57]** >> Can you show Claude real quick?

**[44:58]** Projects. I want to answer his question

**[45:01]** a little bit more just for anybody

**[45:02]** because a lot a lot of folks here are

**[45:04]** not using Claude

**[45:06]** >> projects. Not even Yeah.

**[45:08]** >> Oh, wait. You want my VS code or you

**[45:09]** want my project?

**[45:10]** >> Right here. Right here. Okay. to answer

**[45:12]** your question. Um,

**[45:14]** sir, I want I want to give a little bit

**[45:17]** more context to what you asked. If let's

**[45:19]** say you were you set this up and you're

**[45:21]** using it and it's giving you the same

**[45:23]** FAQs, right? What I would suggest is set

**[45:26]** up another project for the same client

**[45:28]** and don't be scared to have multiple

**[45:30]** clawed projects for one client because

**[45:34]** one client has how many services?

**[45:36]** >> Yeah. Each service should have its own

**[45:40]** project because then when you feed it

**[45:42]** the files, there's no wrong entities.

**[45:45]** There's no wrong blog posts. There's no

**[45:48]** wrong PDN incoming links. There's no P

**[45:50]** uh press release date. You understand?

**[45:52]** It's all about the one service. Then

**[45:55]** when you come in here and ask it for

**[45:56]** everything, you put in the prompt in the

**[45:58]** instructions, hey, once you cover one

**[46:01]** PAAF FAQ, do not cover it again. And

**[46:05]** that's it. it'll just start cycling

**[46:07]** through whatever you feed it. Remember,

**[46:09]** garbage in, garbage out. And the more

**[46:11]** you try to do because it's exciting, you

**[46:14]** start losing quality. When you stay

**[46:15]** laser focused, so again, multiple

**[46:18]** projects for housing plumbing, then you

**[46:20]** go one per service if you need to.

**[46:23]** >> Here's another thing is our prompt for

**[46:26]** doing the PAAAS before we started

**[46:28]** implementing the intake form was two to

**[46:31]** three pages just to write a freaking

**[46:32]** blog. Was out of freaking control. I

**[46:35]** looked at this stuff and I was like,

**[46:36]** "Yeah, that's a damn good freaking like

**[46:38]** prompt." But holy [ __ ] Now, literally,

**[46:41]** because we are putting this stuff in

**[46:44]** into the project, dude, it's it's like

**[46:48]** it's like a paragraph, maybe two at

**[46:51]** most. It's like you've reduced the load

**[46:55]** and you've made it so much less complex

**[46:58]** for the VAS. I mean, just because like

**[47:01]** you're writing writing these prompts

**[47:03]** that are holy crap and now you're doing

**[47:06]** the blog, you're just you're writing a

**[47:08]** prompt and you're saying, "Oh, take a

**[47:10]** look at the intake form. Oh, take a look

**[47:12]** at our pricing." So, you don't have to

**[47:13]** put all that crap in there. He we were

**[47:15]** literally he I caught one. He literally

**[47:19]** put like all the freaking links for he

**[47:21]** did the site map inside of the prompt. I

**[47:23]** was like, "Dude, God, Lord save me." I

**[47:27]** was like, you put a prop with like you

**[47:29]** just probably ate up the context window

**[47:31]** off your damn prop.

**[47:34]** >> You didn't even get to load.

**[47:35]** >> Yeah, you didn't even get to load.

**[47:38]** >> Oh man.

**[47:40]** >> Guys, any questions?

**[47:42]** >> No.

**[47:44]** >> That's the goal.

**[47:45]** >> I love questions.

**[47:50]** I guess some stuff is crazy. you should

**[47:54]** figure that out. This test a lot of

**[47:56]** things that he's dropping in his papers.

**[47:59]** The other thing is he he was feeding

**[48:01]** kind of that stuff grow. So you just

**[48:04]** talk with Gro and Gro will give you a

**[48:07]** lot of insights about

**[48:08]** >> Yeah, Sean's Sean's pretty active on X.

**[48:11]** Um I like I'm a news junkie so I love X.

**[48:15]** Um also a hot tip if you want to get

**[48:18]** into vibe coding it's on X, not

**[48:21]** Facebook. Uh, counter to that, my

**[48:24]** opinion of the SEOs that are on X is

**[48:28]** really bad. If you think it's bad on

**[48:31]** Facebook, the advice and some of the

**[48:33]** things that are said on X about SEO is

**[48:36]** holy freaking crap. It's nightmare fuel

**[48:41]** and it's some of these guys think

**[48:43]** they're really famous.

**[48:46]** >> So, yeah.

**[48:47]** >> Any questions, guys? How many people

**[48:49]** using claw code or claw pro, you know,

**[48:52]** are you using claw desktop? Do you have

**[48:53]** custom projects set up? If not, man, get

**[48:56]** to it.

**[48:57]** >> It's critical.

**[48:58]** >> Like, get to it. Then you get to

**[49:00]** terminal down the line. But bare [ __ ]

**[49:02]** minimum, bro. If you don't set this up,

**[49:05]** [ __ ] chat GPT. Don't talk to me about no

**[49:08]** custom chat GPTs. I'm tell you right

**[49:09]** now.

**[49:10]** >> Yeah. Second I say I see GBT, I'm like,

**[49:12]** well,

**[49:14]** >> stuff, right? Someone mentioned it.

**[49:17]** Grock is amazing when you want to know

**[49:19]** what's going on in the world for

**[49:20]** sentiment,

**[49:22]** >> right? Like what's HAPPENING REAL TIME

**[49:23]** NOW

**[49:24]** >> BECAUSE you can turn on that feature

**[49:26]** when you search with Gro, especially if

**[49:28]** you're using the CLI. Where do you want

**[49:30]** to search the actual, you know, like go

**[49:33]** to the internet or the internet end

**[49:34]** threads or just threads or X? What

**[49:37]** they're called threads, right? X.

**[49:38]** >> No, no, it's just X. Yeah, tweet. I

**[49:40]** don't I just still call it Twitter.

**[49:42]** >> Twitter. I mean,

**[49:43]** >> it goes through Twitter and then gives

**[49:45]** you the report, right? So like tech

**[49:47]** stack wise being that we're talking

**[49:48]** about this stuff. Google Gemini huge

**[49:51]** million context when go research

**[49:52]** everything plus it's Google. I want that

**[49:55]** info from Google like we were talking

**[49:56]** about last night right for the real time

**[50:00]** social media like what's going on cuz

**[50:02]** bro trust me uh Elon got that [ __ ]

**[50:04]** tight. And then the thing you mentioned

**[50:05]** the groipedia. Yeah right. If y'all

**[50:07]** haven't checked that out go check it

**[50:09]** out. Don't listen and not do it. Right.

**[50:12]** Grock deep research plus it's Google

**[50:14]** giving you Google [ __ ] Grock to go get

**[50:17]** the uh Twitter and social media and

**[50:20]** everything else.

**[50:23]** >> Claude Cole, I'm telling you guys, man,

**[50:25]** don't play this. And Madness. Madness is

**[50:27]** Claude Cole made easy.

**[50:29]** >> Yeah.

**[50:30]** >> Right. Because I love Madness and I use

**[50:31]** it. I'm running it right now.

**[50:33]** >> But that's only when I want something

**[50:35]** pretty fast because I don't feel like

**[50:36]** going in VSC.

**[50:37]** >> We Someone asked, I think the other day,

**[50:39]** how many of you are vi coding? And

**[50:41]** almost everybody in the room said yes.

**[50:43]** You have to get on Twitter. The advice

**[50:46]** and stuff that are going on. Who knows

**[50:47]** what Ralph Wigan is.

**[50:51]** We've got three hands and that scares me

**[50:53]** to death. You guys like need to know

**[50:55]** this stuff. you've got to get on Twitter

**[50:57]** and because the Ralph Wigum is basically

**[51:00]** it's a loop system that you basically

**[51:02]** you do a whole bunch of planning and

**[51:04]** like it's for more of a massive type

**[51:06]** project and you literally you put the

**[51:08]** plan in and you walk away and it's an

**[51:11]** infinite loop of running sprints. So

**[51:13]** there's when when programmers are doing

**[51:15]** things they have okay we have this task

**[51:18]** this sprint this sprint this sprint okay

**[51:20]** well this sprint could be connected to

**[51:21]** this sprint and that sprint okay and

**[51:23]** what happens is the Ralph Wigan loop

**[51:25]** runs in parallel and does all the

**[51:28]** activities and you just you know you

**[51:30]** just run off and leave it for a couple

**[51:32]** days and you come back to a program

**[51:33]** that's done on complicated programs

**[51:38]** >> it's no it's still it's still you run it

**[51:41]** through terminal or

**[51:43]** or uh VS Code. No. So, yeah.

**[51:46]** >> How do you get the interface?

**[51:48]** >> I do every I like VS Code. What are you

**[51:50]** using?

**[51:50]** >> I use VS Code.

**[51:51]** >> I like VS Code. Yeah. I'm not using bold

**[51:54]** cursor. I'm not using anything. I'm just

**[51:56]** doing straight VS code. But like the

**[51:58]** stuff and the and the changes and what's

**[52:00]** going on. Um you learn so much about

**[52:04]** what's going on with with AI from X. It

**[52:07]** is definitely

**[52:09]** SEO is awesome on Facebook. Um, I think

**[52:12]** that's the place to be to get network

**[52:14]** and make your groups. But the same thing

**[52:16]** is going on on Twitter for for AI and

**[52:18]** vibe coding and stuff. You will learn

**[52:20]** just a shitload.

**[52:21]** >> You're going to be mad at me. Number

**[52:22]** one, I told you, yo, bro, just code us

**[52:24]** up a [ __ ] thing that takes tweets and

**[52:26]** makes a updates us. I forgot to show it

**[52:29]** to you.

**[52:30]** >> Yeah.

**[52:30]** >> Twitter. Twitter, bro. You could just

**[52:32]** catch a tweet, throw that [ __ ] into one

**[52:34]** of these, and then it gives you a prompt

**[52:37]** or a tool. That's what I got. I promise

**[52:39]** you. Get some one of you out here. You

**[52:40]** want to test me? See if I'm [ __ ]

**[52:42]** lying, send me something that has actual

**[52:44]** here's how you do something. I'll throw

**[52:46]** it in my project and spit you out a live

**[52:48]** vers.

**[52:51]** >> Did we?

**[52:51]** >> Yeah. Remember

**[52:52]** >> Oh, yeah. Yeah. Yeah. Oh, that's right.

**[52:55]** Yeah. And I actually oneshotted a couple

**[52:57]** things on that call. One shot is when

**[52:59]** you could like set things up, give it

**[53:01]** one prompt or a PRD, walk away, come

**[53:04]** back, and pretty much it's done. Maybe

**[53:05]** you need to talk back and forth or edit

**[53:07]** it a bit, but you did the work up front.

**[53:11]** Uh AI is a hallucinate. It's a

**[53:14]** predictive model. Sorry, I said the

**[53:15]** wrong word. So, if you don't give it

**[53:18]** >> all the [ __ ] you can so that it's not

**[53:20]** guessing, you're not winning as much as

**[53:23]** you can. You don't want it to be

**[53:24]** guessing. You want it to just follow the

**[53:26]** SOP, so to speak, right? So, like on our

**[53:29]** mastermind call when Matt Diggity shared

**[53:31]** that post, I grabbed it. I downloaded

**[53:34]** the documents. I put it in claw project.

**[53:36]** I started and then I got a next

**[53:38]** mastermind call. I showed it with them

**[53:40]** and they've been all

**[53:41]** >> diggity. You know that diggity thing. It

**[53:43]** was only on Twitter. He didn't put it on

**[53:45]** Facebook and I found it and I was like

**[53:48]** Mike

**[53:48]** >> I showed you.

**[53:49]** >> I showed it. OH, HERE WE GO. HERE WE GO.

**[53:58]** >> Diggity had a helpful Degity released a

**[54:01]** helpful content thing. it basically. So

**[54:03]** this it it opened the eyes. The reason

**[54:05]** it was so important is you didn't

**[54:07]** realize you could just take a freaking

**[54:09]** PDF and use it as a guide. We didn't

**[54:12]** even think of that. And

**[54:14]** >> the the search quality guide and the um

**[54:17]** Oh man, I closed I closed out of it.

**[54:19]** It's

**[54:20]** >> you know the Claude code where he

**[54:22]** uploaded the files to the projects. You

**[54:24]** can put anything you want there. Matt

**[54:27]** Diggities guy. How about all the Google

**[54:29]** patents?

**[54:30]** >> Yeah, you can do that. You understand

**[54:32]** what I mean?

**[54:32]** >> Any any freaking PDF, just take a PDF

**[54:35]** and that is now like

**[54:37]** >> start talking to Claude about it.

**[54:39]** >> It's not

**[54:40]** >> it's that freaking simple.

**[54:41]** >> Even you were doing that with chat GPTs.

**[54:44]** The whole thing is right the same thing

**[54:46]** you were doing with chat GPT. Feed it

**[54:48]** what you want, you know, make a custom

**[54:49]** GP.

**[54:50]** >> So you can have a conversation about

**[54:51]** >> same thing. Yeah. So Sean's document's

**[54:54]** 200 pages.

**[54:57]** You can put that in. Now, I don't I

**[54:59]** don't really recommend 200 pages using

**[55:01]** that as like a thing. There's

**[55:03]** >> talking just about getting a summary of

**[55:05]** >> Yeah. No, you treat it. Basically, what

**[55:07]** you're doing is you're turning the PDF

**[55:09]** into a chatbot

**[55:12]** because it's in a project. So, that's

**[55:14]** why when we're adding all these

**[55:16]** >> Yeah. So that's why when we're adding

**[55:18]** these other files,

**[55:20]** it basically cloud's becoming our

**[55:22]** chatbot that it's it's we're chatting

**[55:24]** with it and saying, "Hey, I want to do

**[55:26]** this, but why don't you look over here

**[55:28]** at this thing and please look over here

**[55:30]** at my site map and please look over at

**[55:32]** that and oh, here's a Google quality

**[55:34]** raers guide. Can you do something with

**[55:36]** that?" And kind of like when you spit

**[55:37]** this content out, can you look at all

**[55:39]** this stuff and you know, make something

**[55:41]** for me? And it's really that simple.

**[55:43]** just you load [ __ ] up and it spits it

**[55:46]** out and the content is unique because

**[55:48]** the client gave it to you. Um and then

**[55:51]** you can just keep on enhancing it and

**[55:53]** making it more unique and more unique.

**[55:55]** >> Absolutely.

**[55:55]** >> And it it just it's a game changer for

**[55:57]** your SEO because that's what Google's

**[55:59]** looking for.

**[56:01]** Oh. Um,

**[56:02]** >> while he pulls that up, if you interview

**[56:04]** your client

**[56:05]** >> and get their words, their company

**[56:08]** culture, right? Like we talked about

**[56:10]** last night with Notebook LLM. Same play,

**[56:12]** different tool, bro. Upload that [ __ ] in

**[56:14]** there, right? Make the instructions a

**[56:17]** nice prompt like, "Hey, your attorney

**[56:20]** blah blah blah blah. You are going to

**[56:21]** write content blah blah blah blah." You

**[56:23]** know, follow the guidelines I'm

**[56:24]** uploading. Google helpful

**[56:28]** uh content

**[56:31]** guidelines.

**[56:32]** >> You looking for the PDFs?

**[56:34]** >> Yeah. So, there's a PDF out there.

**[56:35]** >> I'll drop it in the back for everybody.

**[56:37]** >> Okay. So, there's the Google helpful

**[56:39]** content guidelines and the search

**[56:42]** quality evaluator guidelines. Literally,

**[56:45]** it it is freaking nuts. You take those

**[56:48]** two PDFs and you run your you can put in

**[56:51]** your site or you run it you tell it to

**[56:53]** use that as a help you write the content

**[56:55]** dude your content changes like that's

**[56:58]** how I knew about the tables and the

**[56:59]** listicles all because that I mean it was

**[57:01]** really like low-key at that time but the

**[57:04]** second you ran it through that it put

**[57:06]** that content instantly in tables and

**[57:09]** bullet points and numbered list all that

**[57:11]** [ __ ] instantly happened by just putting

**[57:13]** those two stupid PDFs in it kicked out

**[57:16]** content that is exactly what everybody's

**[57:18]** talking about right now from the just

**[57:20]** those two.

**[57:22]** >> Yes. Google helpful content guideline

**[57:25]** and search quality evaluator guidelines.

**[57:29]** Okay.

**[57:30]** >> And it's a 200page document that I don't

**[57:32]** want.

**[57:32]** >> Now there is a way to handle that.

**[57:34]** >> I just say if this is true assume this

**[57:37]** is true what ex how should this change

**[57:39]** the way I execute SEO? Right. Right.

**[57:42]** Shrink that thing down.

**[57:43]** >> Yeah. There is. Yeah, there's a couple

**[57:46]** ways. There's to do that. Then there's

**[57:48]** things that are specialized. There's

**[57:50]** specialized plugins for 200 like long

**[57:52]** PDFs where where they are are meant for

**[57:56]** structured data markup of long

**[57:58]** documents. You can run it through it's

**[58:00]** called Llama PDF or Llama something.

**[58:02]** Yeah. And and so that's what I ran it

**[58:05]** through. I ran that 200 thing and it

**[58:07]** spit out a whole bunch of information

**[58:09]** that made it easier to use. And if you

**[58:11]** dissect it because I'm not going to read

**[58:12]** 200 pages.

**[58:14]** from that guy from the other.

**[58:16]** >> Yeah, it's almost like there when you

**[58:18]** get stuff like that, you guys. Um, so I

**[58:20]** guess you could run it through notebook

**[58:22]** LLM, but there's there's there's

**[58:25]** specific um LLMs for long PDFs that spit

**[58:30]** out the information a little bit better

**[58:32]** and easier. They're structured for that

**[58:33]** type of situation.

**[58:34]** >> But to your point, you said projects,

**[58:37]** he's got a table of contents. So you

**[58:39]** could just take a section, drop it in,

**[58:41]** say, "Make me a project to use with my

**[58:43]** clients with this. How will this apply?

**[58:45]** Can you add to the voice of the client?"

**[58:47]** Bam. That's project one.

**[58:49]** >> Yeah. Yeah. Yeah.

**[58:51]** >> Absolutely.

**[58:51]** >> So there there is

**[58:53]** >> I'm going to drop I'm gonna leave I'm

**[58:55]** going to give you one final if I can

**[58:56]** find it. Let me I'm gonna try and find

**[58:59]** my uh personality. Let me

**[59:02]** >> I've been looking for mine.

**[59:04]** >> Yeah.

**[59:06]** What is it you're trying to get to

**[59:08]** online as well?

**[59:08]** >> Okay, so I'm gonna I'm gonna So my

**[59:11]** Claude acts drastically different than

**[59:14]** most other people's clouds. Yours acts

**[59:16]** like mine. I'm pretty sure. When you ask

**[59:17]** claw a question like yours?

**[59:19]** >> Well, we've talked about this before,

**[59:22]** but when you ask Claude something, what

**[59:24]** does what happens? It asks you a bunch

**[59:26]** of questions.

**[59:28]** >> When I ask Claude something?

**[59:29]** >> Yes. Does it come back with you with a

**[59:31]** bunch of questions?

**[59:32]** >> That phase is kind of over. It was going

**[59:34]** through all that a lot, but yes.

**[59:35]** >> Okay. So, all

**[59:37]** >> update some [ __ ]

**[59:38]** >> So, I added ask clarifying questions to

**[59:41]** make sure that you understand what I'm

**[59:43]** asking for. So, there will be no

**[59:45]** assumptions prior to response. Think

**[59:47]** carefully and please consider best

**[59:49]** practices. Now, this is where you can go

**[59:51]** off the rails. I said when writing uh

**[59:54]** code, because I code a lot, uh SEO or

**[59:56]** content, only perform actions based on

**[59:58]** tasks requested. It drastically changes

**[01:00:01]** the way my cloud works because have

**[01:00:03]** that. You have that in your claw MD

**[01:00:04]** file, right?

**[01:00:05]** >> No, I have that in my personality.

**[01:00:06]** >> Your person Oh, you mean on the Okay,

**[01:00:08]** got it.

**[01:00:09]** >> Yeah. So, on my profile, I have that and

**[01:00:12]** I did it to um our employees also. Their

**[01:00:14]** Claude instances have that also. So,

**[01:00:17]** what happens is is when you're prompting

**[01:00:18]** Claude every time it it is freaking

**[01:00:20]** annoying. uh it asks you a whole bunch

**[01:00:23]** of questions, but it turns out it's it's

**[01:00:26]** a saving grace because you didn't even

**[01:00:28]** think about this or that and your

**[01:00:30]** content just became better, more

**[01:00:32]** >> unique.

**[01:00:35]** >> Uh

**[01:00:37]** exact because I'm doing what you said at

**[01:00:39]** the beginning of it. The only thing that

**[01:00:40]** I have that's different at the end of it

**[01:00:42]** is in order to execute the task with a

**[01:00:45]** 95% success rate.

**[01:00:47]** >> Yeah, that's good. I like that.

**[01:00:51]** >> So, we talked about I need to have a to

**[01:00:54]** get this on the recording or

**[01:00:56]** >> I think so.

**[01:00:57]** >> Yes.

**[01:00:59]** >> You're my partner in crime. Anyways, you

**[01:01:01]** want to hop up here?

**[01:01:02]** >> In fact, we should just wrap this up

**[01:01:03]** into panel.

**[01:01:05]** >> Oh, okay.

**[01:01:12]** >> You got something better than the llama?

**[01:01:13]** Because that thing was pain in the butt.

**[01:01:16]** >> It was awful.

**[01:01:23]** Just stand behind the

**[01:01:26]** >> This is for sound.

**[01:01:27]** >> That's for the recording.

**[01:01:28]** >> So, I hold them both.

**[01:01:29]** >> Yes.

**[01:01:30]** >> Oh, okay. All righty. So, with with PDF

**[01:01:34]** files, what they're trying to do is

**[01:01:36]** basically OCR. What all the LLMs are

**[01:01:39]** trying to read. Basically, it's OCR.

**[01:01:41]** They might get it wrong. It's not 100%.

**[01:01:44]** So what I did and I didn't come up with

**[01:01:46]** this I found it or come across wasn't

**[01:01:49]** Twitter somewhere else I don't know

**[01:01:50]** where but if you take your PDF file and

**[01:01:54]** upload it to your choice of wherever

**[01:01:56]** chat GPT or whoever have it turn it into

**[01:01:59]** a MD file so it's a text file right it

**[01:02:03]** will shrink it down from 200 whatever

**[01:02:05]** megabytes to a couple of K maybe 50k

**[01:02:09]** 100k so do that and then upload that

**[01:02:12]** because an MD file is a heck of a lot

**[01:02:14]** easier to read by an LLM than a than a

**[01:02:17]** PDF file. So that's your one minute tip

**[01:02:20]** for big PDFs. You can also say, "Hey,

**[01:02:23]** I'm my my client is a plumber. Only

**[01:02:27]** extract stuff from this PDF which is

**[01:02:29]** plumbing related." So if it's not YM L,

**[01:02:33]** then you can do that. So cool.

**[01:02:36]** >> Yeah. Yep.

**[01:02:38]** >> You want to though you're not You