UnHacked - Cybersecurity Made Simple for Small Businesses

Justin Shelley (00:10)
Welcome everybody to episode 90 of Unhacked Pop Quiz. There are two problems with our introduction. What are they? Somebody go fast. Bueller. Bueller. No, what are you talking about? Then there's three problems. What are the other two? Anybody? Nobody. Sentence is still spelled wrong. That's pre-

Joshua Holloway (00:11)
What's it?

Mario Zaki (00:19)
I see a spinning wheel.

Bryan Lachapelle (00:27)
I didn't pay attention.

Joshua Holloway (00:29)
There's nothing else.

Mario Zaki (00:30)
I at this point I

think it's felt fine. Just keep it. Yeah. It'll be our signature.

Justin Shelley (00:34)
I'm in so much trouble over this.

my God. We ⁓ we have several there's there's so many problems today and and Riverside. ⁓ I guess I shouldn't bash them online. Maybe they'll sue me, but ⁓ my god, they did an update and it's been problematic. And then Brian and Mario, you guys had some updates or something going on. So it's like we're we're 26 minutes into the episode and we haven't even started yet. Because we've been

Joshua Holloway (00:37)
I'm seeing a delay in upload.

Bryan Lachapelle (00:59)
Now I know how my clients

feel every time they say, ⁓ the updates, why you guys gotta do that in middle of the day? We don't, by the way, but you know.

Justin Shelley (01:02)
I know. I know. It's true. It's true.

Mario Zaki (01:05)
Ha ha ha.

Joshua Holloway (01:06)
I

couldn't find a link. So it's like wow, here's the guy that can't find the link.

Bryan Lachapelle (01:11)
Ha ha ha.

Mario Zaki (01:11)
Yeah.

Justin Shelley (01:11)
I know.

It's ⁓ and then you guys missed the the main, the most important problem with the introduction, as I said, episode 90. Today's actually episode 89. ⁓ but but ⁓ Mario kind of floated the idea of just skipping an episode to see if anybody noticed, and you guys didn't even notice, so I know that the audience didn't notice. This is ⁓ this is episode three slash four in our mini-series on AI. The initial plan was to do four episodes.

Bryan Lachapelle (01:29)
No, we did not.

Mario Zaki (01:30)
Ha ha.

Justin Shelley (01:39)
on AI basic setup prompting and whatever, and then we would move into integrations and then we would move into app dev vibe coding.

Mario Zaki (01:46)
But see, we used

AI we used the AI to cut twenty five percent of our episodes. So now, you know, we're we're doing going much faster.

Justin Shelley (01:52)
That's exactly right.

Joshua Holloway (01:53)
Eight times A.

Justin Shelley (01:55)
We got so damn efficient we just couldn't even do it in four. We have to do it in three. So ⁓ that's the big spoiler, I guess, for today. the four part has moved to a three sub sub part because we've got a mini. I don't know. Guys, let's get started here. ⁓ let's do introductions. I'm Justin Shelley, CEO of Phoenix IT Advisors, and a little bit of a twist today.

Not my normal introduction. We've been talking about frameworks. I've been obsessed with frameworks for a very long time in the world of technology, like approximately a year. I don't know. ⁓ we had a guest on here a while back who talked about frameworks and how important they are to cybersecurity. And I've not let go of that since. Today we're actually gonna move to putting frameworks around AI. And that's it. That's all I'm gonna say for that. We're gonna finish the introductions. Josh, you went for us last time, so not happening today. Just chill.

Brian, you go first, Mario, and then Josh, you're in the penalty box. You go last.

Joshua Holloway (02:52)
All right. I'll try and bring the heat.

Justin Shelley (02:54)
Okay, thank you. Brian, you're up.

Bryan Lachapelle (02:57)
sorry.

I thought you were talking to Josh. Yes, Brian Lashpo with B4 Networks. he threw me off with located in beautiful Ontario, Canada. ⁓ yeah, that's all I got.

Joshua Holloway (03:00)
No, he's on the penalty.

Justin Shelley (03:10)
See, that's the thing, I've

got mine memories. I say the same thing every time, so I throw mine off and it just screws you guys all up. Mario, bring it like do something do something amazing on your introduction.

Bryan Lachapelle (03:14)
⁓ my gosh.

Mario Zaki (03:19)
Yeah,

I mean Mario Zaki, ⁓ CEO of Mastec IT, ⁓ located in New Jersey. We are a master in cybersecurity, a master in AI, and mastering in podcasts.

Justin Shelley (03:36)
And weirdly enough, my the former name of my company was Master Computing, but you claimed that word more than I did. Josh, what what do you got for us?

Mario Zaki (03:42)
Yeah.

Joshua Holloway (03:43)
Wow, you

used a wave war. Yeah. All right. For me, I'm Joshua Holloway. I'm the CEO for Seventh Eye Technologies out of sunny El Dorado Hills, California. And also an office and Reno. ⁓ our primary for IT is working with companies that require compliance and how to make technology work within those guardrails when technology typically doesn't want to. So if you're looking for compliance, we got you. Framework, framework, framework.

Justin Shelley (04:02)
So you like frameworks too?

Josh, I'm gonna put you on the spot a little bit. What in the hell does seventh DI? Like there's gotta be a little background. I I looked for and I'm gonna I'm gonna throw you under the bus. I looked on your website for your company name, story, and it's not there. And to be fair, neither's mine, but go ahead.

Joshua Holloway (04:19)
Yeah, so

no, it it's all good. So the original company's name was seventh dimension, and the idea behind that was attacking business problems dimensionally, right? So not just looking at technology and being a nerd and talking to technology and saying, Hey, here's the tech that you need. It was more about dimensionally looking at the business, the problem, its issues, the people, the age of the staff, everything and and coming at it from different, you know, degrees of

How do we integrate technology with an older staff that can't necessarily handle the change, or maybe they're ⁓ change adverse? Or how do we take the policies, the procedures, and everything that goes into it to kind of make technology wrap around it? Because at the end of the day, the technology needs to work. And if it and if it doesn't work or if it doesn't help, we get no buy-in. And then what happens? People go back to the same old Excel spreadsheets, the same old manual processes, and then they can just continue to hate IT. This way we

hopefully bring about that desire to not hate T hate IT as much and and see them as a partner that's working alongside them to improve their technology, improve their business processes. And dimensionally is how we're answering it. We're not just saying, and here's your technology, use it. It's more like how do we wrap it all around an entire business? And the business is not a flat, it's not flat at all. So there's a lot of dimensions to it. So there you go.

Justin Shelley (05:37)
That yeah, that that works. And the the key my my key takeaway, Josh, was hate IT less. ⁓

Bryan Lachapelle (05:39)
Fancy.

Joshua Holloway (05:45)
Yes.

Well, I mean, who likes IT? They're only calling IT because they're pissed things aren't working, right? I can I've been doing this for 27 years and I think I've had two or three people call me in 27 years to say, hey, everything's working. And that's awesome. Yeah. And ⁓ actually, nope, today was the fourth. I sat in a meeting today right before coming here. And the guy's like, since we moved to you, we've had one outage in an entire year when we were having outages every other day.

Justin Shelley (05:50)
yeah. yeah.

Good job. Thank you. Yeah. Yeah.

Joshua Holloway (06:16)
prior. And he's like, That's amazing. And we were just sitting in a meeting and I wasn't expecting that, you know, you just you kind of go into that and he popped off with that one. And I was like, Cool, I'm going to I'm going to put that feather in my cap and move on.

Justin Shelley (06:27)
I could take this way off track and and ask how in the actual hell were they having an outage every other day? I've I've never seen that in my 35,000 years in IT, but ⁓ No, no, I'm just saying like that's rough. That's rough. That used to be kind of a sales pitch for managed services, right? It's like, hey, we're gonna prevent your downtime. And now I tell people that and they're like, we don't have any. ⁓ everything's cloud, app, you know, web app, whatever. So that's just

Joshua Holloway (06:37)
I mean, I can go into it if you want to.

Mario Zaki (06:39)
Mm-hmm.

Justin Shelley (06:55)
Take me back, takes me back. And I'm an old man. So yeah, now we're vibe coding. So, guys, let's let's go ahead and jump in. ⁓ we've done two episodes already on AI, and we've kind of started with picking your favorite platform and setting it up securely. And then we were supposed to the second week come back and give actual use cases of here's what you can do to improve your productivity. There's been a little bit of bleed back and forth, overlap, whatever. And you know, as I'm planning the next ⁓

Joshua Holloway (06:57)
And L'5 coding.

Justin Shelley (07:24)
supposed two out episodes where we're gonna be just building on this same concept. ⁓ like I said in the introduction, it just it just wasn't coming together for me. And maybe, maybe it's because I'm like bored to death because I've been doing that for so long. And I hope audience, listening audience, if if that's not true and you want more of a development on this, ⁓ you're in luck because any of us will sit and and talk through this with you like one on one. Just reach out.

the you've got all of our links in the show notes and stuff. So ⁓ we'll walk you through that one on one if you want. But meanwhile we're gonna kind of push this thing forward, wrap up the ⁓ basic platform, and then next week we're gonna come back and we're gonna hit hard with integrations because that is the world that I'm living in right now. Everything I do, I'm like setting up a new damn integration. ⁓ It's a pain in the ass, but once you get it, this thing will just ⁓ it just sings. Yeah. Anyways, Josh,

Mario Zaki (08:21)
Yeah.

Justin Shelley (08:22)
You are the one who introduced this concept of frameworks around AI. I love the word. You're a s you're a compliance guy as well. Talk to us about how do you put a framework on AI prompting? No, I'm not talking about security, right? Usually frameworks are security. This is AI prompting. So take it away.

Joshua Holloway (08:40)
So with the with AI prompting, you have a set structure to it and you have to think before like what do you what do you want your interaction to be like? What is the output and and what is your overall outcome? So one framework you can think of is a PTCF, right? So let's break that down. Peace for persona. I want you to act like a lawyer with twenty five years of ⁓ expertise in corporate, you know, legal.

then you're saying, I'm gonna give you a task. I want you to review this contract and I wanna look for all of the exclusions, any gotchas or language that would leave me and my business out to dry if I don't follow certain ⁓ criteria, right? And then it's the context, right? What's the background, what's the constraints, what's the audience of relevant data, right? So ⁓ I'll probably tell it that I'm an MSP, been open for you know 28 years.

this is what our core fo focuses and we're looking to do business with company XYZ. So when you analyze this, this contract, this is what you're, you know, this is what we're looking for, ⁓ or this what could affect us. I may even share additional information with it, like a copy of ⁓ cyber liability policy or other insurance policies that I have. And then last is ⁓ format or a structure. Do I want a a a summary?

⁓ some kind of widget or dashboard or something out of it. I know previously last week I talked about doing ask it a simple question like what can aid AI do for me ⁓ in my job role? What's the top 10 things it can do? I'm giving it no context. I'm just saying what's the 10 things that you can do for my business, right? It's gonna be very broad and it's gonna say AI can do X, Y, and Z for you. ⁓ so what we did is we came up following these frameworks with a with a prompt.

That says, here's your persona. I'm a wealth management operator. ⁓ And here's my daily tasks, my weekly tasks, my quarterly tasks. ⁓ this is what my day-to-day looks like. And here's my my entire job role. How can AI help me claw back hours in a week, hours in a day, hours in a month? And it will like fully list it out because I gave it context, I gave it a framework to work inside of.

And I I gave it ideas, I gave it a structure or persona and I just let it run. So I feel like that's kind of a good way to attack it.

Justin Shelley (11:06)
And and yet

to be fair, we've we talked about that when we talked about setting up the platform. You ⁓ I believe in Claude we mentioned it's called projects in chat GPT, it's called a custom GPT, where you go in and you give it context. And you know what you just described was just a structured way of doing that.

Joshua Holloway (11:25)
Yeah, structured prompt. You don't have to build a GPT. This is a structured prompt. You can copy and hold on to it because you might run it, you know, 15, 20 times or come up with a more fancier way of holding on to it. But it's just a structured prompt or how you prompt. I I think a lot of people who are are using ChatGPT as Google, they are Googling their prompts. They're saying, like, what's the capital of, you know, Illinois? Or you know, what's the population of India?

Justin Shelley (11:26)
Yeah.

Bryan Lachapelle (11:46)
Yeah.

Joshua Holloway (11:53)
Things like that, right? So they're they're not realizing that they can do a lot more with it by giving it more context, more framework, or more guidance.

Mario Zaki (12:02)
Yeah, I mean it comes back to what we were talking about last week. It's you know, it's what you put in is what you're also gonna get out. You know, you you ask it a question, you're gonna get an answer, you know, and that's it, and you're both gonna move on with your day. You know, if you interact with it, you know, treat it as like a ⁓ a person pretty much, and have a conversation either through text or through verbal, you'll get more interaction back.

Joshua Holloway (12:10)
Barbage in is garbage out.

Justin Shelley (12:31)
You got any thoughts on this?

Bryan Lachapelle (12:33)
it's pretty much the framework that I use ⁓ on a pretty re regular basis. I always tell the the GPT or or Claude, you know, this is who you are, this is what you do, this is how I I want you to do it, this is the ground rules and and it's almost exactly that framework. I I probably slightly modify it in some cases, but it it's usually spot on exactly what Josh was saying.

Justin Shelley (12:54)
Okay. ⁓ it and I know we've talked about you don't do anything live, right? With especially with AI demonstrations, but guys, I'm breaking rules today just because I'm feeling a little punchy. ⁓ Josh, I I want you to, if you're up for the task, and we can sit here and shoot the shit while you're, you know, prepping or whatever. ⁓ just raise your hand when you're ready to go. But I want you to just give a an actual live example. I know you described it, but I want you to do it. Pull up your favorite

Mario Zaki (13:05)
Mm-hmm.

Justin Shelley (13:23)
platform LLM put in a generic prompt read the prompt read the output put in your ⁓ frameworked prompt read the prompt read the output when you're ready to do that are you ready now okay that's what that's what I'm saying like I'll I'll give you time I can sit here and just talk about shit I don't know ⁓ what should we talk about guys listen we're we're gonna ⁓ what I want to talk to you Mario and Brian about is do you use this concept

Joshua Holloway (13:35)
Okay. Well you guys talk some to yourself.

Justin Shelley (13:52)
in your in setting up like your custom GPT or your project or whatever, or do you use it just in a in a a live prompting session, a chat session? And Brian, you're you're nodding, so yeah, go for it.

Bryan Lachapelle (14:02)
well yeah.

So ⁓ I it's funny enough, I just ⁓ brought up my custom GPT that I created I don't know, maybe six months ago for ⁓ building a quarterly key initiative. And in that prompt ⁓ or in that settings in the in the instructions, ⁓ I actually have you know AI personality definition. You are a strategic execution coach, planning and architecting and trained in performance operations, you combine empathy with implementation.

Anyway, it goes on and it talks about who it is. And then the next, very next prompt is tone and behavior. And then it jumps into ⁓ quarterly key initiative workflow and scope. So it tells it how it's going to do what it's going to do. So what that like essentially what I'm looking for. ⁓ my prompt is or this this instruction set is massive because I've got a lot of constraints that I was trying to make sure that people wouldn't try to bypass things. ⁓ I'm happy to bring it up or share the initial ⁓ instruction set.

with you ⁓ to post for the audience on on our unhacked ⁓ page. ⁓ but essentially I'm following that almost exactly that same framework on a custom ⁓ version of a prompt. Now the difference between, if you remember from last week, between what Josh was saying is is a prompt that you enter in in a new chat and a custom GPT is I don't have to tell the custom GPT each time what it is. I just open it up

And it already knows who it is, what its purpose is, what it's going to do. And I just have to give it the first set of instructions, like, hey, I'm ready. And then it'll go through whatever process I've told it. Right. ⁓ so that's the difference between having a custom GPT and having to remember to prompt it each and every time is that you you're essentially recycling that information, that that prompt over and over and over again.

Justin Shelley (15:43)
Right.

And I'll I'll say and Mario, you're you're up next, so start prepping. I I'm gonna tell myself a little bit because I've been doing I customize my I I hate in the prompt itself. I just want to talk. I don't want to give it a bunch of instructions and I set that all up ahead of time. But I'm building today's outline, and like I said, I've kind of been grappling with do we do this in two, do we do it in one? And now I'm looking at the clock, I'm like, God, maybe we're done already. I don't know. I'm ready to move on to integrations. Anyways, I I

I got into the argument with ⁓ I'm using Claude, you know, we duked it out and finally ended up with, yeah, we're just gonna combine this into one. And I'm like, great, give me the bullet point outlet. Out outline, not outlet. And it it comes up with this just like unformatted, whatever. And I'm like, guys, for 80 whatever episodes, or I don't know how many I've been preparing outlines with ⁓ AI, it's it's quite a f what the hell? Joshua Holloway's recording has stopped.

Josh, did you leave us? I I guess I pissed him off by giving him assignment. Anyways.

Mario Zaki (16:48)
It it something

s something with ⁓ with Riverside ⁓ is

Justin Shelley (16:52)
I know. They

they did an up I tell you they did an update and it caused problems. Anyways, I can hear you, but you got no video, but that's okay. ⁓ screw around with that for a little bit. I'm gonna keep telling this story. Okay. I don't know. That weirds that's weird. That threw me a little bit. Anyway, so I I I realized that for dozens and dozens of episodes, I've been having ChatGPT or Claude, depending on the day, create an outline

That's formatted however the hell it wants to format it that day, whatever mood AI is in. And then I take and I copy and paste it into my standardized outline. I'm like, this is the dumbest shit. I'm doing a a 12 slash 11 episode series on how to use AI, and I'm still using it wrong. So I grabbed my outline that I even have a template for, and I threw it in and I said, Hey, can you just format it like this? Which it did perfectly. So ⁓ and I I had everything else set up. I had all the context and everything else so that I don't have to.

really worry about how I prompt it, I just chat with it, but I had missed that one piece and it made I mean here again, we're talking like probably fifteen, twenty minutes worth of work that I've been putting in every single week that I just did away with by importing a template. It's stupid. Mario, what are your thoughts on this? What are your ⁓ lessons learned or whatever?

Mario Zaki (18:07)
I mean the one thing, you know, even for for me personally, I I'm gonna guess it's the same thing with you guys is and and it it gets frustrating. There's so much different things not only can you do with AI, but there's so many different types of AIs. You know, you have to you have to use the right tool ⁓ when you're trying to get a different result. If you're trying to generate, you know, images, there's a special tool for that or a special AI for that.

There's ⁓ when you're trying to get, you know, put together, you know, an outline, you know, you use, you know, one thing. even when you're trying to do like an automated ⁓ like phone system, you use a different prompt. You know, they're they specialize in different things. And it it gets frustrating keeping up with it. And you know, for me, I I every time every day I'm learning about a different tool and different ⁓ system, you know, and it's sometimes it's hard to keep up.

Justin Shelley (19:06)
Yeah. Well it yeah, 'cause it changes. Like you get it all set up and then ⁓ all of a sudden your your great friend from Canada comes up and says, Hey, there's a new tool you should be using and it just like every goddamn thing changed. This whole week I've been setting it we're that's why we're gonna move to integrations. It's like all week long I've been setting up new integrations because Brian couldn't just shut up and let me do it the hard way. He's gotta say, Hey, there's this new thing I'm like, God. So

Bryan Lachapelle (19:31)
I I don't even think it's new, it's just new to me. It was just like, Hey, I you know.

Justin Shelley (19:34)
⁓ right. New to you,

Mario Zaki (19:34)
Yeah, new to us, you

know.

Justin Shelley (19:36)
new to me, new to whatever. It it cost me a full week of reworking how I do vibe coding, which we're not to yet, so that's good. That's the good news.

Bryan Lachapelle (19:44)
Yeah, yeah.

Mario Zaki (19:44)
And I I

I mean Justin, I I'm not gonna lie to you. I signed up with with hats like you you you were talking about in previous episodes. It took me a few minutes just to realize where to go just get to get started. You know, it's it's it's a little it's a little ⁓ hidden with a couple yeah.

Justin Shelley (19:52)
Yeah.

There's a lot in there. There's a

lot in there. And and the everybody's used different ⁓ naming conventions. So, you know, what they call an agent, somebody else calls something completely different. And they've got ⁓ agents, workflows, and apps. And I spent like a day trying to figure out the difference between an app and an agent. I don't I don't know, whatever. ⁓ basically an a an app is kind of like a project.

Or or this framework we're using right now where you just go in and you tell it how to behave. I think I don't know. See see right now, I don't even know. God damn it. Josh, are you back yet? I

Mario Zaki (20:34)
Yeah.

Justin Shelley (20:41)
Josh, could you even hear us anymore? He's muted. He's got no camera. I'm gonna unmute him. No, I'm not. He can't be unmuted. Well, I guess we're gonna have to go on without Josh. That sucks. yeah, hats. So what do you think? You you're you're a fan of hats or not so much, Mario?

Mario Zaki (20:41)
He's muted. He's muted.

Bryan Lachapelle (20:46)
I think we broke

Mario Zaki (20:59)
I haven't really gotten a chance to use that. I mean maybe maybe one day you I mean ⁓ yeah I I can figure it out and I could ask it how to use it, but I I do wanna see s how y how

Justin Shelley (21:09)
It not one of its strong

points, I'll tell you that. I keep asking hats how to use hats. It is not one of its strong points. It's like, I don't know.

Mario Zaki (21:15)
Yeah.

Yeah. So so so maybe maybe, you know, I'll have you because the other day when you shared your screen, I was like, wow, this is pretty cool the way you have this set up. So I I I may want you to kind of, you know, be the teacher and I'll be the student. Yeah.

Bryan Lachapelle (21:17)
I don't know.

Justin Shelley (21:27)
Yeah.

I'd happy to do so. You know what?

We've got a ⁓ unplanned break here. Why don't we just do it? I'm gonna pull up hats and I'm I'm gonna kinda walk Phoenix sorry. I can't I cannot multitask to save my life. ⁓ all right, now let's go back in here.

Mario Zaki (21:51)
I mean I have to admit

I just asked it are you able to spin up sub agents and it's been generating for like three four minutes. I'm actually pretty surprised how slow it is.

Justin Shelley (22:01)
Has been.

Mario Zaki (22:04)
I mean right now it's ⁓ it's just says generating. ⁓ I don't know what it's doing.

Bryan Lachapelle (22:10)
Generating Hello

Justin Shelley (22:11)
Yeah, like give it a give give it a minute.

Mario Zaki (22:12)
Ha ha ha.

Justin Shelley (22:14)
It's gonna it's gonna let you know here in a bit. ⁓ I'm gonna show you real quick, I'm just gonna show you a couple of my favorite things on hats. Let me get my screen shared. And guys, if you're not watching on YouTube, you probably should be. ⁓ because I don't really know how to

Bryan Lachapelle (22:17)
Mm-hmm.

Mario Zaki (22:31)
And for for our our viewing audience or listening audience, ⁓ hats hats is just, you know, uses the same agents that, you know, Claude and other places use, but they do it in a secure manner. You're able to manage your your employees, so that way you you're aware of what they're working on or I don't know if you're you're aware of what they're working on, but you could at least control ⁓

you know, who has access, you you you have them use this because you know where it's been, you know it's not gonna leak any sensitive information. It does have a filter knowing, you know, that it's if you put in a social security number, you know it's not going to be, you know, r released in the wild.

Justin Shelley (23:14)
Right.

It's not going out. This is a contained environment. That is my favorite hands down, excuse me, hands down my favorite feature. ⁓ that said, let's just look. I'm gonna pull up a basic chat here. And first of all, you can you can select from all these tools. It's just like, you know, I I'm not gonna try to go over them, but everything Microsoft three sixty-five. Now, most of the platforms are gonna do some version of this. You have integrations all over the place. We're gonna come back to that.

Bryan Lachapelle (23:19)
That's right.

Mario Zaki (23:43)
Wait, I didn't even see where you got to that. Where did you click on that? tools, okay.

Bryan Lachapelle (23:47)
Tools.

Justin Shelley (23:47)
tools down here in the chat.

But really what I love is you can you can use all of the different LLMs. So you can set it to auto, which I really love. That's their newest feature that is just phenomenal because it'll choose which one is best for whatever you're doing. But if you go in here to all of their different ⁓ like there's there's just I wish I knew the number, but there is a ton of different LLMs you can choose from. So

anyways, what else was I gonna

Mario Zaki (24:18)
Yeah, the the auto

the auto is pretty good because if you're tryna I mean I on here you you there's a button that says generate ⁓ image generation. But if you if you have it automatically pick, I'm sure it's gonna pick the best one ⁓ to generate, right?

Justin Shelley (24:27)
Yeah.

Yeah, because you can't

do it in Claude. Like Claude doesn't do image generation. So yeah. Anyways. So you've got all the LLMs. That's that's one thing that I really love. ⁓ but what I also like is you can build

It's like flow charts. I'm trying to see if I've got a good one, a workflow. Yeah.

I don't know. I don't want to spend too much time on this. but it it's just this whole environment where you can ⁓ pull in all the LLMs, you can you can create your projects, your custom GPTs, as it were, and and then you can trigger things from, you know, no, not workspace workshops. You create workflows, and then you can use webhooks. So you can have some other thing trigger the this AI workflow. ⁓ you just do it like

Like I said, add new step, there we go. Prompt step, you can add a web scraper step.

I know. Document processing. ⁓ you know, you can just have it do one thing and then another thing and then another thing. And within each step, you can change the LLM. So I want it to use Claude here and then down here, I want it to go switch to something else. You have to go view all and let's just pick GPT. We switched to GPT. ⁓ I don't know. This thing and and the the frustrating, the good and bad about it is the more ⁓ like every day there's something new and different about it, and it's just like

⁓ integrations. That is the other thing I was blanking on what else I wanted to show.

their their integrations page, it just keeps getting longer and longer and longer. Whatever. ⁓ let's see. Josh. We're gonna have to like live, try to troubleshoot Josh. Maybe we'll cut this out. Maybe we won't. Josh wants to join. We're gonna let him in.

Mario Zaki (26:29)
There he is.

Justin Shelley (26:29)
Josh, are you

back? Holy hell, that was rough. Like I'm I'm trying to eat time up, but

Bryan Lachapelle (26:32)
Wow.

Joshua Holloway (26:34)
Huge.

Mario Zaki (26:35)
I actually think I see

you sweating. What happened?

Joshua Holloway (26:38)
Wow. I I don't think

Bryan Lachapelle (26:38)
He didn't.

Joshua Holloway (26:39)
I've

ever had like a hardware crash like that in a long time.

Bryan Lachapelle (26:44)
Hey you a hey Josh, do you need an IT guy?

Justin Shelley (26:45)
Well, the good news is

he needs an IT guy. The joke that's never been told. And and Josh, you y I I didn't you didn't miss anything. I did not do a great job. I tried to demonstrate hats and I I was all over the place. I'll probably go back and I don't know.

Joshua Holloway (26:48)
Hey, you know what? Thank God I am an IT guy 'cause I got it fixed.

Mario Zaki (27:03)
No, no, you did. It was it was pretty good. ⁓

the one thing I do wanna also add on to what you were mentioning, which is a very good point, is you can you said you can change the LLM. And for people that yeah, w in for newbies that you know may not be familiar with it, the reason you may want to do that is again, for a couple of reasons. Certain agents are better than others for certain things.

Justin Shelley (27:16)
Within the workflow. Yeah.

Mario Zaki (27:30)
But also certain agents are a lot more expensive than other agents. So sometimes if you're doing something, you know, trivial or something that is just ⁓ minimal thinking, you may want to use a cheaper LLM to save on tokens because none of these ⁓ AI services are giving anything unlimited. Even

And correct me if I'm wrong, even chat GPT and and even like those twenty dollar plans, like you you guys are gonna you you limit even if you're just having chats and stuff like that, right?

Justin Shelley (28:01)
I never

run out with 20 bucks a month on Chat GPT. Unless you're you're coding or something, then I will. But like if you're just chatting, I I chat all day, every day. It gets embarrassing how much I use it and I haven't run out. So but you're right. I mean, and and image if you're doing a lot of that. But I will say, Mario, with with the number of LLMs you can choose from, if you don't know which one does what, that doesn't really help much. So I love that they put in the auto feature.

Mario Zaki (28:04)
Yeah.

Joshua Holloway (28:06)
Cody needs it. Yeah.

Image gener generation can eat e tokens.

Yeah.

Justin Shelley (28:29)
Where it will automatically by the way, chat will do this. You know, you can set it up to auto-select its own, you know, with it its limited number of models. ⁓ you can have it automatically switch back and forth within hats, you can have it switch with all of the models of all of the platforms. So it's it's crazy. ⁓ Josh, let's go ahead and see if your technology is going to keep up. And do you remember your assignment after all that?

Joshua Holloway (28:49)
Yes.

yeah, 'cause it started going real crazy when I started prompting it. So yes, I remember my assignment. Yeah.

Justin Shelley (28:58)
Okay. it took over. That's the problem.

Okay. Good tell us what you got. And share your screen if you want. If you can just talk through it, read through it. That's fine too.

Joshua Holloway (29:07)
⁓ yeah, so I don't think I can necessarily share my screen because of where I've been prompting, but what I can go with is I went with a very simple prompt. I want to write an incident response playbook in the theme of cybersecurity for business email compromise, right? Kept it really simple. That's all I told it. I hit go. The first thing I ran into is it asked me three questions to define like, is it a reusable ⁓ client template?

What service are you using? Is it Google? Is it Microsoft 360? And then do you want it like a branded doc or anything like that? That was my its initial take. Then it went into like what colors do you want to use? What are the fonts you want to use? So it really started diving into a lot of questions on how to get this done. ⁓ it pretty much started asking me, do I want a cover page? Do I want this? Do I want that? And then after a while, yeah.

Mario Zaki (29:57)
Trasha, I wanna interrupt you. d you

I'm assuming in your settings you have it already set to say, you know, ask me any questions you may you may ha you know, need to to give me the best results. Do you already have that in place or 'cause a lot of times it will just

Joshua Holloway (30:11)
For for for Chat GPT

I do, for Claude I don't. Claude I kinda left it kind of I don't wanna say basic, but I I I turned that feature off. And I was using Claude in this one.

Justin Shelley (30:19)
And that's what you are using now?

Mario Zaki (30:22)
Okay. So

Justin Shelley (30:22)
Okay. I

I've got a key point. I want you to keep going, but I'm I'm yeah, Mario, that's a question I had and I've got something to say about it, but keep keep going, Josh.

Joshua Holloway (30:29)
Okay. ⁓

so it did give me ⁓ a a a playbook back, pretty basic, pretty pretty generic, right? I I didn't give it any context and I didn't follow any of the framework. ⁓ I then when my computer really started ⁓ going kind of crazy, this is how I I framed my prompt. You are an expert in cybersecurity with a a CISSP and twenty four twenty-five years of experience.

I want you to build an incident response playbook for a business email compromise. This will be used by IT and management staff to follow during this type of compromise. I need it to be in an outline format so that it can be followed with a lot of reference points because people will be using this during a high anxiety time. And the output output will be in Word so that we can update it in the future.

it did not ask me any questions. It just went to work. And ⁓ I ended up with yeah, with a crashed computer. Yay me.

Justin Shelley (31:26)
With a crash computer.

Bryan Lachapelle (31:31)
I don't think the two are correlated.

Joshua Holloway (31:32)
Yeah.

Yeah, no, it's that was a fun one. But here, if we want, ⁓ let me see if I can share this out for you guys to kind of see what the framework looked like.

⁓ and we'll we'll take it from there.

Justin Shelley (31:56)
Are you sharing it?

Joshua Holloway (31:58)
not yet. Hold on one second. Here we go. And share.

Justin Shelley (32:00)
Okay.

There we go. Okay. I didn't know if you froze up. I didn't know if I missed something. I was just like, Okay, go ahead. I'm sorry.

Joshua Holloway (32:03)
no.

And of course it's sharing the wrong thing. I love it. So that's on me.

Wow. That's a fun one.

Justin Shelley (32:22)
Just leaving dead air on purpose so that everybody gets really uncomfortable.

Mario Zaki (32:25)
Mm-hmm.

Bryan Lachapelle (32:26)
I'm just gonna fall asleep.

Justin Shelley (32:27)
Why is it this

is a this is a tangent. Why is it that humans are so uncomfortable with dead air, with quiet all right.

Joshua Holloway (32:33)
Okay, so here is our framed work based prompt. this is what it came out with. ⁓ gave its own title screen.

Bryan Lachapelle (32:33)
it's definitely a strategy to use in

Justin Shelley (32:42)
Just off the

top, Josh, how many pages was the first one?

Joshua Holloway (32:45)
⁓ the first one was 12 pages. This one is ⁓ surprisingly 24, doubled in size. it makes it more of like a con ⁓ a policy flow. So there's going to be acceptance, who's the owner, everything else, how to use this document, a table of contents. So it actually took the time to go through and you know, it gave us appendix processes, internal responses.

Justin Shelley (32:48)
So it's doubled it. Okay.

Joshua Holloway (33:12)
Everything that would typically be found in a incident response playbook.

A pretty thorough one. I can share it with you guys if you want to use it. You could thank you can thank Claude for it. Yeah.

Bryan Lachapelle (33:21)
Ha ha ha.

Justin Shelley (33:23)
Shit, shit, I'll just build my own bitch. Anyways. ⁓

okay, so I want to come back to the question Mario asked. And and the point I wanted to make about that is because Mario's like, wait a minute, did you already set some of this up? Because it seemed like it was following instructions by what you described. And that is the point that I want to make. When I first started getting into AI, it was super important to structure your prompts a certain way. And now it is still important to give it context.

But the structure of your prompting, the the language and even this framework that we're talking about, I feel you guys, tell me if I'm wrong, I feel it's becoming less important.

Mario Zaki (34:05)
It's it's because it's getting ⁓ it's getting smarter and I think it it it like if you're using the paid version i it it's learning you, you know, like yeah.

Justin Shelley (34:16)
Absolutely. Yeah.

Joshua Holloway (34:16)
Yes.

Justin Shelley (34:18)
Yeah. But but that that is my point though, is as it becomes smarter, as it learns more about us, it auto-customizes. And so it is still important to give it the right context. And the like I, you know, the example as I kind of threw myself under the bus, like it was doing a great job of giving me the information I wanted, but I'm like, no, I want it formatted like I'm normally seeing it every week. I want it to look the same. ⁓ in including my typical sign of, you know, like

Bryan Lachapelle (34:18)
That that I would agree with.

Justin Shelley (34:44)
I'm Justin. Remember, listen and take action and keep your business unhacked. The dumb shit that I say. So I I give it the template and it throws it in. So now I won't forget what I said. That's that's the good news. ⁓

Joshua Holloway (34:49)
Unhacked.

Bryan Lachapelle (34:55)
Yeah.

Now you gotta update

your template to include all of our sign offs. ⁓ Yeah, we do.

Justin Shelley (35:03)
It well you guys don't have a standard sign off.

You do? Well then put it in there 'cause

Joshua Holloway (35:08)
Well we just all

say unhacked at the wrong time.

Bryan Lachapelle (35:11)
Yeah.

Mario Zaki (35:11)
Ha ha

Justin Shelley (35:12)
I know we

need like a like we need like a timer or a bell. ⁓ anyways, okay, so ⁓ listen, we have already made I feel like this is kind of a rambling episode, and I'm apologizing in advance. We're gonna get way more ⁓ engaged next week, and maybe it'll get too engaged. I don't know. But guys, do you have any final thoughts on this? Anything else that we need to cover on prompting and just basic pre integration AI use.

Bryan Lachapelle (35:15)
⁓

Justin Shelley (35:42)
Anybody, anybody.

Bryan Lachapelle (35:43)
Well,

what what I would w the my biggest my takeaway would be this. If you're going to be asking AI for anything, even if you're not sure, let it know what you're doing and ask it for the best prompt that you you can use that will give those four steps and then it will generate a prompt and then you can correct places where it went wrong and then reprompt it with that same prompt. ⁓ that's been my go to if I'm you know, like how what would be the best way of asking this question? Because if you ask it, it won't it won't try to optimize itself.

Justin Shelley (35:55)
Yeah.

Bryan Lachapelle (36:12)
with the instructions you gave it. But if you say, I don't know how to ask this properly, can you help me draft the best prompt I can to accomplish this? It will then give you a prompt that if you feed it back, you'll get a totally different result than if you just fed the initial question. So it's ⁓ it's pretty fascinating. It knows itself enough to give you a better prompt to get a better result.

Justin Shelley (36:33)
Use AI to use AI. Yeah. I mean tr it's it's it's the truth.

Bryan Lachapelle (36:34)
Yeah.

Joshua Holloway (36:37)
Well,

I I think the other thing too, and Justin, if I'm not mistaken, we're gonna share a couple not just a PTCF framework, but all the other ones what for so people have access to them. You can even say, I wanna use a framework. I maybe I don't quite understand it, ⁓ but I I know I wanna get my point across a certain way. Here's the framework I wanna use.

Justin Shelley (36:59)
You go ahead and and dive into that. I I wasn't really gonna spend a lot of time today on that. I can include it as a download. ⁓ yeah, but but talk through any of those that you want to bring up.

Joshua Holloway (37:06)
I would. Yeah.

Yeah, so couple of the frameworks you have like tag, which is task, action, and goal, right? So define the job or activity, describe the goal, and then discuss its purpose, right? So you can say I want to use the tag ⁓ framework to build my prompt. Help me do that. Here's my action, but I'm not certain of my goal, or here's my goal. I don't know where to start with my action. And then like help me build this out. So that would be one of them. Another one would be like Bab, which is gonna be before, after and bridge. ⁓

I definitely would think this would be a good consumable download so that people can see. I mean, obviously you can Google all the different prompting frameworks and you'll end up with like a million of them. But we came across, I think it was nine in total, ⁓ that we kind of hit on. you have care, which is gonna be contact, context, action, result, and expectation, or grow, goal, reason, and output. So you're you're giving it this framework or this expectation that you want to work through, and either you can.

work with it yourself and build out your prompt that way or just say like this is what I want to follow because I feel like this is going to get me a better answer and this is the this is that framework I wanna ⁓ I wanna work inside.

Justin Shelley (38:18)
Yeah. Okay.

Guys, any thoughts on that? Mario, you got any takeaways?

Mario Zaki (38:25)
one thing too that I was gonna say is that you you know, with AI it it it does very well when you give it examples. So like if you have something ⁓ and you're able to upload it and say, I wanna improve on this or I wanna change this or create my own, it does pretty well. You know, it it it knows it it's because you what you've essentially done is put the the structure that you guys were just talking about. It's given it

you know, the structure. Very similar to like Justin's example. Like, here's the template, stick to this. You know? and I and I was just looking at it, I think they got rid of it, but there used to be like ⁓ a thumbs up and a thumbs down on Chat GPT where, you know, y the results and stuff like that, but I I don't know, either mine just decided it didn't w really give a shit any more what I care about or they did away with it.

Justin Shelley (39:08)
Yeah.

Joshua Holloway (39:09)
Mm-hmm.

Justin Shelley (39:17)
No, I think it's like,

listen, we know you better than you know you, so I don't need your feedback.

Bryan Lachapelle (39:22)
Yeah.

Joshua Holloway (39:23)
Well, you know, it's perfect

Mario Zaki (39:23)
Exactly.

Joshua Holloway (39:23)
that you you say that, Justin, 'cause one exercise after using any one of these after a while, you should just ask it, like, what's your perception of me? No, it's it's a it's a good exercise. No, no, no, it's a good exercise.

Justin Shelley (39:31)
Hell no. Not doing it. I don't want to know. No, people people, there's

Mario Zaki (39:32)
Mm.

Justin Shelley (39:37)
a there's a Reddit, sub subreddit, whatever you call it, about ChatGPT. And people love to get on there and and say, hey, draw a picture or you know, make a photo of everything you think you know about me. And like s it it puts up some weird shit. I don't I don't I'm not doing it. I did it a while back and I'm just like, nope, I don't wanna know. It's getting too close to iRobot, guys. This is just weird.

Mario Zaki (39:51)
Yeah.

Joshua Holloway (39:57)
Yeah.

Justin Shelley (40:01)
⁓ okay guys, ⁓ we're we're gonna kinda wrap this up. It ⁓ unless I've missed something. So I'm just throwing it out again. this has been an episode of rambling. We're gonna come back with better structure next week. ⁓ but lies unless you guys stop me with something important, I'm gonna give Mike

Bryan Lachapelle (40:01)
Yeah.

Mario Zaki (40:04)
Mm-hmm.

Bryan Lachapelle (40:20)
The the only thing

Joshua Holloway (40:20)
Read all reboot

all your computers before you start recording.

Bryan Lachapelle (40:21)
that we

Justin Shelley (40:23)
That's important, maybe twice.

Mario Zaki (40:23)
Mm-hmm.

Bryan Lachapelle (40:25)
The other thing to

add was that we were gonna give assignments out every week. And so ⁓ if if I'm correct, the assignment for this week was supposed to be to create your own custom GPT ⁓ or project if you're in Claude, ⁓ upload some files, ⁓ give it a set of instructions, tell it what you want, upload some context files. I'll I'll give you a quick, quick, really brief example of one of the chat GPTs I made really early on is I took my entire entire employee manual, ⁓ all the

Justin Shelley (40:36)
Right.

Bryan Lachapelle (40:52)
policies and procedures for the entire company, uploaded it to a chat customer GPT as additional files, and then told that you're an expert in all the B4 Networks ⁓ employee manual and all our policies and answer any questions our employees have. And so instead of them having to go through every single policy, they can say like, hey, what's B4 Networks' address code? Boom, there would be the answer. and it was locked down limited to just that set of documents. So it couldn't go out on the internet. like I had configured it to do that. So

If if it's okay with you, Justin, that would be my challenge to the audience is go and create yourself a custom GPT, ⁓ that you can upload some files and some instructions to and ⁓ and see what what the cop capabilities are.

Justin Shelley (41:31)
Yeah, absolutely.

Mario Zaki (41:31)
Now, what do

you think would have happened if you don't tell it it's an expert in this? Do you think ⁓ you think the results would be different?

Justin Shelley (41:40)
Test it, Mario. Like this is supposed to be what we do today. Right now, I'm not kidding you. Go to, I don't know, like do it while we're on the air. You know, I don't see you moving. Okay, so challenge denied. Got it. ⁓ I but no, it's it's what I wonder too, because it it used to be, it feels to me like, I don't have evidence of this, but it feels to me like the the setup is becoming less important.

Bryan Lachapelle (41:41)
Yeah.

Joshua Holloway (41:43)
Yeah. See, that was your challenge.

Justin Shelley (42:06)
Maybe the more we use it, maybe the more AI develops, maybe a combination of the two. ⁓ but as I found out the hard way, that has led me down a path where I got complacent and and like I was costing myself time because I didn't take the effort, you know, ahead of time to give it a template, you know, to do a little bit of setup would have saved me, you know, hours of time over the course of prepping for the podcast. So ⁓ I'm gonna I'm gonna

Throw out a grievance that I have about AI lately because it makes it so that we can be more efficient. I mean, that's our big claim, 10 to 50x, more efficient. One of the things that drives me batshit crazy is that when I'm chatting with something relatively simple, I can ask it like, what color is the sky? And it doesn't say blue. It writes a goddamn novel about how and why and how sometimes it's not quite blue and it's a little bit of a green, and then the clouds come in and it's a

Mario Zaki (42:52)
Yeah.

Bryan Lachapelle (42:54)
Mm.

Justin Shelley (43:02)
And it it talks about clouds and why there are clouds and how the clouds form and it's just like, Jesus Christ, answer the question and shut up. And sometimes I'll have to tell it that. Like, just give me a one word answer. And then it like gets butthurt. It's like Blue. I

Mario Zaki (43:12)
Yeah.

Joshua Holloway (43:14)
Well, it's 'cause you didn't give it a framework to begin with. Say I would just

like a one sentence response.

Justin Shelley (43:19)
Jesus Christ, Josh. God damn it. You're right. You're right. So maybe this whole thing I've been saying, these frameworks aren't important. Yeah, I guess they are.

Mario Zaki (43:26)
No, but

the thing is the the thing is it does it does also you also have to take it with a grain of salt because I you know, I I think I've actually mentioned this ⁓ a few months back. You know, my wife ⁓ for her birthday, she wanted us to to just stay in the house. She wanted ⁓ fondue. She wanted instead of going to like melting pot, one of her favorite restaurants, she wanted to stay in, she wanted us to do, you know, me and the kids to do all the work.

Bryan Lachapelle (43:26)
They absolutely are.

Mario Zaki (43:54)
So I left work and I opened up my Chat GPT and I'm like and I gave it the problem. I'm like, you know, today's my wife's birthday. She wants me to cook dinner and she wants fondue and her favorite ⁓ you know thing from melting pot is the Wisconsin trio. So give me all the exact ingredients that are there. And I ended up

It it it's like, okay, great, you know, it's a great cheese, you know, no problem. I'll go ahead and create you a ⁓ you know, a list of all the ingredients you need to do and I'll also give you the exact things that you need to do to cook or whatever. And you know, I wasn't thinking about it, I was driving and I I I got the lists, I went inside, you know, the the supermarket, I got all the things that it was on those lists, and I went home and you know, I started prepping and my wife's like

Bryan Lachapelle (44:43)
Yeah.

Mario Zaki (44:47)
What the hell is this? I'm like, What do you mean? She's like, What you you got all these cheeses. I've never heard of these cheeses. We've done this, we've made this before. You know, usually you need this, this, and this. She already knew. She didn't even to look needed to look it up. And the c the Chat GPT completely gave me wrong recipe. And then when I went back and I said, Hey, you gave me the wrong, you know, ingredients for this and she's like, you're absolutely right. You're good, you know, you good

Justin Shelley (45:07)
Mm-hmm.

Mario Zaki (45:18)
It i it very good of you. Yeah, it good good catch. G you know, like this is the correct and I'm like, No, you're forgetting this. She's like, ⁓ another good catch and it kept and I'm like, Forget it. You know, I and

Joshua Holloway (45:18)
Good catch.

Bryan Lachapelle (45:28)
The irony is

you call ChatGPT she and I call ChatGPT he's so stubborn. It's gotta be a man.

Mario Zaki (45:31)
And it it drove it

drove me crazy. I ended up and I didn't want to ruin her birthday, so I had to go back and get, you know, three, four different cheeses and stuff like that. So y you please.

Bryan Lachapelle (45:40)
⁓

So can I can I add to that story?

I also have been cooking using ChatGPT and I I had it give me recipes. And so I get home and I'm like, okay, I want to cook this recipe. Give me step-by-step instructions. And it starts giving me the instructions. I'm like, dude, you had me get all these other ingredients. What like when when do we use those? And it says, ⁓ those don't belong in this recipe. And it's like, then why the heck did you have me buy them?

Mario Zaki (46:08)
Yeah.

Justin Shelley (46:10)
What? What ⁓

Mario Zaki (46:15)
Yeah,

I I don't know why there's times where it just goes on this weird tangent. It's like and it just makes you so angry.

Justin Shelley (46:16)
God.

Bryan Lachapelle (46:18)
Yeah.

Joshua Holloway (46:23)
I mean I'd classify

a hallucination, right? It hallucinated the the rest

Bryan Lachapelle (46:26)
So I have spinach

in my fridge and I I don't know why I've got spinach in my fridge.

Mario Zaki (46:31)
Yeah.

Justin Shelley (46:31)
Well I I had

it this morning tell me that Nevada is in the mountain time zone. And it's like, no, it's not. you're right. It's not. my bad. Like, what? Well, listen, and it's important too, because you know, I'm gonna bring this a little more serious tone. What we haven't talked about today is security. And, you know, this isn't directly security, but we have to be careful not only with what we put into it, for fear of that getting

Bryan Lachapelle (46:38)
I think we're we're giving our audience a lot of confidence in using AI.

Mario Zaki (46:42)
Yeah.

Justin Shelley (46:59)
⁓ stolen or misused in some way, but we have to be super careful about what it gives us back. We can't just accept it as gospel. And that's really hard because we're getting into vibe coding here in a little while. Guys, this thing spits out thousands and thousands and thousands of lines of code that we probably don't understand. And we're going to talk about how to handle that. But man, you can't just let this thing run amok because it will and it'll crash all our computers as we learned.

Mario Zaki (47:25)
Yeah.

Justin Shelley (47:26)
Okay, guys, we are gonna wrap. We've been we've been going for 47 minutes. That's long enough to wrap up this first segment. ⁓ I do need to ⁓ provide an update on the unhacked portal. I have been working feverishly on this thing. The full reveal, the full release will happen in you know, week, whatever it's it is now, I think nine slash eight. ⁓ but I have been making some major update updates to that. And in fact, Josh, if you'll give me you already have the downloadables that you were talking about.

the the new portal is gonna have every episode. And then when you go into it, you know, we can very easily add attachments that people can download. So I'm gonna I'm gonna use yours as a guinea pig to test the resources section of today's episode, which is 89, not 90. hopefully your documents don't have any malicious content to them. Did you check them for that? That's what I'm saying. ⁓ maybe we should upload your instant response plan as a resource as well. I don't know.

Joshua Holloway (48:09)
Wow.

It's not gonna make your computer crash like mine.

Mario Zaki (48:17)
Ha ha

Joshua Holloway (48:24)
Yeah, exactly.

Justin Shelley (48:26)
All right, guys, that is it for this week's episode of Unhacked. I think we have beat this horse clear to death. And next week, I mean, like, don't tune out because next week we're gonna start making this thing really sing and dance. ⁓ go to unhackmybusiness.com. That will soon be redirected to the new portal that I'm building. For now, it still goes to the my own company website that doesn't have as many resources as I wish it did. but still go there, check us out. All the episodes are there. And with that, guys, we're gonna sign off.

Final goodbyes. And before I let you guys say your goodbyes, I'm just gonna say again, thank you so much for being here, for bringing your time, your brilliance, your experience, because ⁓ the number one benefit for me being on this podcast is that I get to learn from the brilliant minds out there. You guys are part of that. Brian, go ahead and say your goodbye, then Josh, then Mario, then we're gonna wrap.

Bryan Lachapelle (49:17)
All right, thank you very much, Justin. likewise back to you. ⁓ my name is Brian Lashbro with B4 Networks. if you are ⁓ trying to improve your technology, your use of AI, your use of automation in your business, ⁓ give us a call, we'll help you in your journey.

Every day.

Joshua Holloway (49:34)
Hey, thank you. And I'm

Justin Shelley (49:35)
Ever

Joshua Holloway (49:36)
Joshua

Mario Zaki (49:36)
Yeah.

Justin Shelley (49:36)
ever day.

Joshua Holloway (49:36)
Holloway with Seventh EI Technologies. And we're here to help you with your compliance and technology needs and trying to get AI to fit in there too. So if you're looking for help there, Seventh AI Technologies, help you

Mario Zaki (49:48)
Yeah. Yeah. Mario Zacki, CO of Mastec. ⁓ you know, like I've said before, we we specialize in in helping business owners have peace of mind. ⁓ we will be there for you every step of the way, making sure that your company's there n in the morning when you wake up. And we'll protect them from those Russian hackers, Chinese hackers and all those other guys that you know, yeah, you know I I I had to

Joshua Holloway (50:10)
I wanted to make sure that was coming.

Justin Shelley (50:12)
I le I left out this week. They got a pass. I don't I'm I'm pissed at all of

Mario Zaki (50:17)
that they're too busy coding. ⁓ but you know, just like everybody said, we're we're here to help you guys with any any ⁓ stuff that you guys need, both technology, help desk, IT, anything you need, AI included, we're here.

Justin Shelley (50:19)
Yeah.

And I I just gonna pause to say, Mario, you've got the best name of everybody here because I'm like, Brian, your turn. Josh, your turn. now it's my turn. I'm Justin. But I can go Mario. I can't do that with anybody else's name. So ⁓ all right, guys, I am Justin. Remember listen and take action and keep your businesses unhacked.

Joshua Holloway (50:52)
Not a hack.

Mario Zaki (50:52)
Unhacked.

Bryan Lachapelle (50:52)
Yeah.

Mario Zaki (50:53)
Better, we're getting better.

Justin Shelley (50:56)
Yeah, kind of.