Be honest.

What's one task that you keep pushing to the bottom of your to-do list?

The admin, the proposals, the blog posts that you meant to write last month.

Well, what if AI could do it for you in your tone of voice

in a fraction of the time.

In this episode, we're exploring how to teach AI to do your least favorite tasks.

You'll learn how to spot the tasks AI can do better than you.

A simple framework to decide what to automate and what to keep human.

Plus, how to train AI to sound like you not like a robot, and stick

around to the end to learn how our guests built an entire workshop.

Using AI slides, notes, handouts, everything in just a few hours.

Welcome to Architecture Business Club, the show that helps you build

a better business in architecture so you can enjoy more freedom,

flexibility, and fulfillment.

I'm your host, John Clayton, and if you're joining us for the

first time, don't forget to hit.

Follow or button so you never miss another episode.

We're joined by Brian McCartney, the CEO of Arc Mark, a leading business coaching,

branding, and marketing firm working exclusively with architecture firms.

He has founded and led six creative service firms since 2002, and has

helped more than 5,000 architects transform their practices.

To connect with Brian on LinkedIn, click the link in the show notes.

We are gonna talk about how to teach AI to do your.

Least favorite tasks so you have more time to do more of what you love.

And wouldn't we all like to do more of what we love in our week?

For sure.

So, um, I think, yeah, I think everyone's gonna be really interested in this one.

Brian, you've, you've said before that ai, you don't think it's gonna replace

architects, but it will make the firms that use it far more competitive.

Can you explain what you mean by that?

Why, why you believe that to be the case?

From just a practical standpoint, if, if you can use a tool

like AI to do a lot of the, what I call the busy work, you know, the routine

things, the things that don't excite you and the things that, uh, are.

Typically those, those, those tasks that are kind of in, in the, you know, the

things that you have to do to get to the things you want to really wanna do.

Uh, I think there's a huge amount of value in that, not just in

terms of productivity, but in terms of enjoying your work.

Right.

Um, I, I am focused so much more now.

I, I'm, you know, with ai I am able to, I'm able to automate things that I don't,

I never liked doing in the first place.

And so now it just gives me a lot more time to think about strategy or

think about innovation and creativity.

Um, I think as a designer, I, I went to the graphic design school and

I'm not an architect, but you know.

I went to school to be creative.

I went, I, I wanted to learn to be a creative, and that's what I focus on.

My creativity these days isn't as focused as it, it used to be on design,

but it's now it's more on how can we be more creative in our business?

How can we, how can we find new ways of, of thinking and using tools?

I've, I've, I mean, I think I shared this with you before.

I, I am a early adopter with, of a lot of tools.

I, I'm a, you know, most people who know me will probably agree

that I have a tool for everything.

Uh, if, you know, if somebody asks like, Hey, how do you do this?

I'm like, there's a tool for that.

I've always been looking for ways to do things better, faster, easier, cheaper.

Um, and I think AI, of all the tools I've ever used is probably one of the

most impactful, um, tools that you, I've ever been able to, to utilize.

I mean, it, it reminds me of when, you know, I started my career

around the time that the max, uh, came out and desktop publishing

started flourishing.

You know, we had, previous to that we were cutting, you know, cutting

paper and, and, and film and, using markers and brushes and, uh, then

along comes this, this device that can, you know, produce, artwork.

Um, I think we're at that kind of inflection point again with ai.

Um, so yeah, I, I believe that.

AI isn't necessarily gonna replace designers.

maybe someday, I, I don't wanna say never,

but, uh, for the moment I, I see that as farther off.

And I think where architects and other creatives should be obsessing right

now is how do I use these tools so that I can make more time for, for, for

my creativity, for my design work, or whatever it is you want to focus on.

Yeah.

I love that.

That's, that's sort of part reassuring and also part exciting as well, that actually

at the moment we're, this is a really exciting time to be alive, you know?

Um, such a, a unique time.

Um, and absolutely this is something we should be learning

about and exploring more.

you talk about identifying some of the biggest time wasters, so some of those

tasks that the, the stuff that kind of drains you, drains you weak, the

things that you don't enjoy working on.

What do you think are some of the most, um, common tasks that

architects, you know, what are some of the things that they, they've told

you they'd like to hand over to ai?

Yeah.

I mean, some of the things that, um, that we see as obvious, obvious

opportunities are, you know, and I know, I know a lot of architects are already

doing this, but, um, uh, meeting notes

is, is one thing I, I see a lot of people using meeting recorders.

What I don't see is a lot of them taking that and then taking it to the next step.

Uh, now when, in our system, when we have a meeting.

It's recorded, A transcript is created like most meeting recorders, and a summary

is produced like most meeting recorders.

But, but in our system, we can then take that and that will create tasks

in our, uh, project management system so that, that, that AI will look at

that transcript and say, oh, here was some deadlines that they talked about.

Here were some, um, uh, some tasks that they talked about.

And it will organize that into clickup, our, our project management system.

And then, yeah, we have to go in and, and verify those.

But you know, that, that cuts out a huge step of somebody

having to sit there and type that

out.

Uh, another place that I, I see is with email responses, right?

Um, a lot of us, uh, are writing our emails from scratch every

time.

I mean, I, I, I got in the habit years ago of saving some templates

that I, I reuse over and over again.

Like when people ask me for references or when people ask me, you know?

Uh, do you have a list of websites that you've designed that, that I can look at?

Yeah, I do.

It's, it's a template and I just call it up and I personalize it.

Well, you know, now I can have AI actually find that email that's

requesting those references.

It will, you know, personalize that, that email and send that response for

me without me having to do anything.

That's so cool.

so.

You know, so that's, that's a huge time saver for me.

Um, uh, and I, I'm sure it would be for a lot of architects too,

because a lot architects get contacted all the time, Hey, can

you show me some of your projects?

Can you show me some?

Uh, do you have references, uh, or people that you've worked with?

These things are like, those are like the things that like, you know, you

get them, you get them regularly, but not like so often where it's

like, oh yeah, I'm just going to, I'm just gonna grab this and send this.

Uh, so having AI to be able to do stuff like

that, or, um, you know, RFIs even, you know, like, I mean, you know, there's

so many repeatable tasks that I think architects just do every day that,

you know, if you really sat down and just took an inventory, you know, for

maybe a couple of weeks and started listing out these things, I think

we'll talk about the framework as

well, but, um.

You know, really understanding like what, where your time is going is so important.

And as we, as these tools become more used, you know, your competitors

are gonna be using them too.

So the sooner you get ahead of this and sooner you start adopting these things,

uh, the better off you're gonna be.

You know, that's, that's what it's gonna come down to is that it

won't be the AI that replaces you.

It'll be the people using it, so,

Yeah.

Yeah.

Um, I think for the vast majority of, um, architects and service-based

business owners, generally, I guess that there's, there's a huge proportion

of their time each week that.

It's not spent doing, they, what, what's traditionally thought of

as the work, like the, the kind of thing that they were originally

trained to do and that they did.

Maybe when they worked for another practice, like designing buildings.

There's a huge proportion of time, particularly for the sole practitioners

where they're doing all this other stuff.

You say like, there's all this admin and emails and, and a lot of

it is repeating processes, so there

even res, you know, the, like, there's things like research, you

know, like, oh, I need to do code research, or I need to,

uh, I need to review the code,

right?

You can upload that code into a tool like Notebook lm, which

doesn't, you know, search the web.

It only works with the information you give it.

Uh, you can cut that down time down to nothing.

You know, it's, it's simple.

It's a simple request in, in chat, uh, to get the information you need.

I, I think there's just ways of thinking be, and I, and I think too, it's, you

know, what I see is a lot of people getting stuck with Chap GPT, like thinking

that's the end all be all.

It's not, uh, there's so many other tools that you can employ

that are, um, many times free.

Um, or just better tools to

Yeah, for sure.

I think chat, GPT is probably the one that most people have heard of and

it's a good kind of like generalist AI to maybe test out and try.

But as you say, there's far better tools for, for specific tasks.

Absolutely.

For sure.

Um, you've developed a framework.

Let's have a chat about that.

This is called the, the teach framework, and this is to help

people figure out what to automate.

Could you briefly break that down for us and just sort of walk us through

each of those steps in that framework?

I did a workshop for, uh, cran, which is the Custom

Residential Architects Network from the a I A here in the us.

So they had their national symposium in, in Virginia, um, uh,

I guess it was about a month ago.

And I did, I led a workshop on how to, how to teach AI to do the least favorite task.

That was the exact name of it.

And so this, uh, so there was a framework that introduced there that we've

kind of moved forward a little bit.

And then the framework is called teach and it, it stands for task clarity.

So first you have to understand like, Hey, can I describe exactly

what done looks like for this task?

Right?

What's the end goal of it?

Uh, then you have, uh, so then the next checkbox is expertise required.

So does it truly need my unique skills or expertise?

If the answer is no, it might be a good candidate for ai.

Uh, then you have articulation.

Can you explain the steps or the rules or the process for the task?

In about 10 minutes.

Right.

Uh, we don't wanna, you know, if it, it's gonna take more than 10 minutes.

It's probably too

complicated.

Right.

Uh, consistency.

Does it follow a clear, repeatable pattern or process?

Right.

Is it repeatable?

If it's something where you're constantly having to make judgments and, you know,

uh, different decision making, it's probably gonna be too complicated.

Uh, and then high frequency, is it something you do weekly

or even more often than that?

If it is, it's probably, you know, if it meets all those criteria,

then it's probably a good, uh, a good task to think about automating

or using AI to help you with

that's really helpful.

Um, and makes it really simple then that if we do a task audit, we, or we have a, a

task that we have to do in front of us on our to-do list, we can use that framework

to figure out, very quickly, we can figure out, is it worth me using AI for this?

Am I gonna, is it gonna save me time if I use ai?

Because if it doesn't meet all those criteria, like if it takes so long

to explain to the AI what to do, or if it's something that, um, you also

mentioned about it being high frequency.

So if it's not something that actually, you know, once you've put

the effort in to create the, the prompt and instructions for the ai, um.

If you're hardly ever gonna use it, it's not really gonna save you much time.

Um, so yeah, really, really important and, um, really simple

framework that people can use that.

So that's really good.

Um, you mentioned about your workshop now.

In your workshop you talked about the idea of treating AI a little bit like

a summer intern, so it's not like a sort of magic bullet or magic solution.

Um, can you explain a little bit more about your, your thoughts on that?

Yeah.

The reason I like this analogy in, in comparison is because I, I find

that a lot of people, when they use ai, they are going into it.

Like they're, they have this expectation, like it's gonna just

magically solve everything for them.

And it's like, you know, poof, it's, it's like, you know, their wish is come true.

Like it's a genie in a bottle.

It isn't, it is a tool.

Uh, and when you think of it like an intern, imagine you just hired

an intern and you brought them into your office and you're like, Hey, I,

I need to give you some work, right?

What are you gonna do?

Are you just gonna say, here, do this?

No, you're not, you're, you're gonna say, okay.

Here's, here's, here's the, what I need you to do.

I'm gonna give you a set of instructions on how to do this.

Uh, you're gonna go away, you're gonna do it.

Uh, I'll probably give you some additional information, maybe some

background information, things like that.

Explain, you know, what we call the context.

And then, uh, and then what we need to do.

You need to go attempt to do this task.

Then you're gonna come back, I'm gonna give you some feedback.

I'm gonna tell you what you did right, what you did wrong, and then

you're gonna go back, do it again.

And you're probably gonna still have some mistakes.

But we're gonna go back and forth.

We're gonna iterate through this.

And that's, that's how we need to look at ai.

It is not.

It's not this magic bullet, it's not something that's just gonna solve

all your problems in one prompt.

Um, I think I, I see a lot of people get frustrated with AI because they,

you know, they tell me, oh, well I tried using AI to do X, Y, Z, right?

And I'm like, okay, great.

So what did, what did you do?

Well, I just, I asked it to do that.

Okay.

Did you give it, explain, did you explain what you wanted it to do?

Did you explain like what role it's supposed to take on, or

did you give it any background?

Oh, no, I didn't do that.

Okay.

Well that's why you're frustrated because you didn't take the time to actually

give it the information that it needed.

It doesn't have all your context, all your experience, all your knowledge.

Um.

And I think the other point about this too is that what I see a lot of people

trying to do is to use AI for things that they don't know how to do themselves.

How you explain it if you don't know how to

Yeah.

How do you know what good looks like with the output you get

back if you don't understand what you are reading or receiving?

Yeah, exactly.

I mean, you're not gonna solve nuclear fusion using ai.

I'm just, you know, if you're not a, if you're not a nuclear scientist, I'm sorry.

It's just not gonna work.

So,

yeah, I, yeah, I, just think it's, you know, I think, uh, and

I also think slowing down, right?

Just, just being a little bit more, uh.

Accepting of the fact that, you know, there is a, a, there's a back

and forth between you and the tool.

It, I know, it, it like, it, it talks, it sounds like it's almost human and, you

know, but it is just a language model.

It's not, it's predicting things.

That's how it works.

Uh, so, uh, you know, you have, sometimes you just have to slow

down and, you know, look at your own thought process behind it too.

Right.

It's one of those things I, I have, I have fell, I have fallen into this

trap myself where I'm like, you know, why isn't it doing what I wanted to do?

And often I realize it's because, well, I didn't do the

things that I was supposed to

Yeah, you

garbage garbage out, know?

well that's it.

You can then look back at your original prompt and go, ah, okay.

I didn't say specifically to do it this way, or I didn't

say not to do it that way.

Um, I think there's a lot of parallels, um, with using AI and outsourcing as well.

If you think about where, um, architecture firms might outsource some of their work,

and I think if you've not done outsourcing before, there is this expectation that

you're just gonna send it off to somebody.

Here you go.

There's some drawings.

This is the building.

Like, just do all of the detailing for us.

Just like draw all this stuff and there's an expectation that it's

gonna come back often and it's gonna be like perfect or near perfect.

And often it isn't.

And it's like, well.

How, how solid was the brief?

Like, you know, if you are not giving a really robust set of instructions

and a really solid brief for exactly the context of the projects, like

the expectations, the deliverables, guidance on the, the cloud or BIM

protocols that the, uh, person should be using you, you are not gonna get

it back looking like what you want.

And, um, even in, remember when I worked in architecture practices and we, it

wasn't outsourcing but delegating, so delegating a task to, uh, a colleague

or a junior, um, in the office, that the instructions needed to be crystal clear

because what, otherwise what happens?

You, you come back with the drawings and you'd be like, oh, they've not

done this, they've not done that.

Why have they done that that way?

And then I look at my instructions and think, ah, hang on.

It's my fault.

Yeah.

Now, now there's another side to that, right?

With AI is that if you give it too much information, it's also

maybe not gonna perform as well.

Uh, so we gotta find a balance in there.

What I like to, to kind of, you know, I, it, it would be the same with an

intern or, or, uh, you know, delegating something to an employee, um, instruct

on the result that you want rather than the method that you want it done.

Right?

Look for, look for, uh, giving the advice, instructions on, on what the

expectations are and what, what the, um, what the parameters are for the Enderol

rather than, oh, do this step, do this.

Because a lot of times you'll find it, it finds a different

way to think about things and.

Uh, you end up with much better results if you don't dictate all of the,

all of the steps in between.

That's a great tip.

You also have, um, a simple five step formula that we can use for

writing a really great AI prompt.

Um, can you run us through that and share how that works?

Yeah.

It's really simple.

I mean, and, and I don't, it's not anything that I created,

uh, it's something I've just learned, uh, over the time.

First you have to define a clear role.

So what, what role is the AI going to perform?

Uh, next you have to be clear about the task that you want it to do.

What is the, what's the task, uh, that you're asking it to do?

What's the expectation that you have?

Um, next is the context.

What information will it need to complete that task?

So if there's additional, um, background information or, um, maybe, maybe a

specific, um, document that it needs to reference, you need to provide that.

you, you then need to kind of help it understand what structure

you want for the output.

So it's this gonna be formatted text in markdown.

Is it gonna be code?

What, you know, what are your expectations there?

And then finally, you want to give it some.

Uh, what I, what I call like, the things to avoid or the things that

you don't, you know, like the mistakes that you don't want it to, to make.

And so, uh, helping to understand like, you know, just, I think we said it at

the beginning, or you said it at the beginning, you know, what does, what

does, uh, what does done look like, What's gonna make you happy as an end result?

Yeah, could we maybe try together and do like

a little worked example of that?

I'm, I'm trying to think of a simple idea that we could share here, um, for, you

know, let's say there's an architect, maybe a, um, let's say it's a residential

architect and the task is to do something, let's say something fairly simple like

write a blog post or I dunno, like a LinkedIn post or something like that,

that would appeal to its customers.

How would we then use that, that formula that you've provided to instruct the ai?

If we were gonna say, use something like, let's say chat, GPT.

uh, Let's say it's for a blog post.

You're an expert, expert, uh, content creator, uh, for, uh, architecture

firms, you know, that, uh, write, helping them write blogs, right?

So you might want to define something like that as the role, the task.

The task is you gonna, you're gonna help us write an article about, Choosing an

architect in our, uh, local area, right?

So let's say Cleveland,

Ohio.

How to, how to choose an architect in Cleveland, Ohio.

Uh, the context, uh, you'll probably want to give it some background about yourself.

Maybe in this, in this particular case, what I would recommend is that

you ask the AI to come up with a series of questions about this topic.

Then you answer those, let's say in a recording or, or

whatever, have that transcribed.

That would be the context.

So you're actually giving it your expertise on this ob

uh, on this topic as well.

That's by the way, how we do our articles

that's a good

our clients.

Um.

So, uh, then you wanna give it structure.

So you want to, you want to define like, um, how many words the, uh, article should

be, uh, should it be a SEO optimized?

Should it be, uh, formatted with headers and subheads and

bolded texts, uh, for emphasis?

Uh, probably, uh, also want to have some bulleted lists and,

uh, a call to action at the end.

Like, Hey, uh, go to our website and, uh, you know, get a,

uh, get a free consultation.

Or in most cases, apply for a free con consultation.

Um, finally, you would want to have some don'ts, right?

So, uh, in this case for a blog article, I'd probably say, uh, don't use obvious

ai, uh, uh, response, uh, you know, kind of, uh, formatting of text.

You know, don't use words like ensure and delve

and, you know, uh, yeah.

yeah.

Yeah, exactly.

Or, or in x, y, Z world.

Uh, that's one, that's my biggest pet peeve.

It's like today's busy architecture work.

You know, it's like what?

Why is saying that?

Um, uh, and avoid things like, you know, M dash is, uh, you know, all

over the place, that kind of thing.

So, yeah, that, that would be the kind of, i, I, you know,

the skeleton of the prompt.

You'd probably wanna refine that for your own

needs, but, uh, yeah.

No, that's great.

That's a really good example.

Thanks for that, Brian.

Really good.

So you mentioned earlier about how AI has helped in significantly

increase the profitability of your own agency in the past year.

Um, I'd love to hear about that and about what's kind of changed day to day in the

business and, and how this has happened.

So I've been talking about AI within our, our,

our firm, uh, I think since 2021.

Just as a side note, John, uh, I actually, my first experience

with AI was back in 1998.

So I actually ran a, I was, I was part of a, uh, a marketing campaign

for an advertising campaign for Calvin Klein, where we used, uh, we

used ai, uh, an AI system from MIT,

uh, to, to create a, uh, a storyline.

Now, back then it was probably called machine learning or something.

It wasn't real AI like we have today, but it was pretty cool.

Um, but I, when, when AI came, kind of came on the scene, um,

uh, prior to chat GPT, we had a tool called, uh, conversion ai.

And it was one of the first AI tools that came out, and I jumped on it.

I bought a lifetime, you know, deal with that.

And, uh, I started using it right away to write articles and things like that.

And so I've been promoting AI within our team for, well, almost four years now.

And last year, um, you know, we've been using AI to write articles for

the last couple years, but last year I told my team, I said, Hey, I want

everybody to, you know, start thinking about how we can use AI more deeply

in our, in our day-to-day business.

You know, what, what, what other tasks can we, can we, um, can we use AI for?

Uh, in my own case, just to give you some examples of how I'm using it, um, it used

to be that when we onboarded a client.

We would go through a pretty, a pretty intense process.

Like it, it was like, it was exhausting for me because I'd be spending like

three weeks, you know, going through like, uh, transcripts from, uh, all

these calls and, and trying to synthesize a lot of information all at once,

uh, into like some sort of report.

Right now what I do is I record those calls.

I'm very engaged on the calls.

I ask better questions on the calls because I'm not writing notes.

I'm actually talking and having conversations like we are right now.

But when those calls are transcribed, what we do is we take those and then we have,

we have, I have projects that I've created that, uh, basically know what I want

to do with those particular transcripts and it will synthesize that information,

draw out the information I need.

And it will create, uh, basically an outline of what we call that.

We, we go through two major workshops in our onboarding.

One is called our Brand Build and the other is called our Clear Story.

Brand Build is about understanding the firm.

Clear story is about understanding their ideal client.

And so I'll use these projects to basically take the, the, the, the

information I've gathered from the client.

And now, uh, that project will, will then synthesize that for me.

It'll extract all their great answers, um, and create an outline for me.

And then what I do is I take those outlines and I have a,

I use a tool called Gamma.

Gamma is a, uh, AI powered, uh, presentation tool.

And I will basically copy that, that outline into Gamma and it will

create the presentation for me.

And so what used to take me three weeks now takes me about four hours.

So it's, it's very, I, and sometimes I'm getting quicker at it.

So, um, I, I'm getting that time down quite a, you know, there's a lot of,

I gotta read it all and, and whatever.

So

that's probably gonna, there's always gonna be a review stage, but

just to be, just to be able to do all that in that amount of time,

it, it's so much more efficient.

And quite honestly, it's better because a lot of times the language model

will pick up on things that I'm, I, I might miss, or, you know, I, I might

go through a transcript and just like.

I'm only human, right?

Uh, so it might be something subtle that it picks up on that ends up

being really helpful later on.

And so once we have those documents, then we can create any

kind of content for that client.

That's the, that's the basis of their messaging.

So websites, um, you know, uh, taglines, any, anything we want,

uh, can now be created, uh, social media, blog posts, uh, all that stuff.

So it, it's a huge, huge time saver.

And yeah, it's, it's enabled us to increase our profitability

from, we were at 21% a year ago.

We're at 50% right now,

Wow,

so,

that's

amazing.

Yeah.

Very, very cool.

Yeah.

Yeah.

It's, it's pretty awesome.

I'm, I'm really happy

about it.

I bet.

I bet.

Absolutely.

Yeah, I'd be happy to.

Um, as an aside, um, I do use some AI tools in the

production of, of this podcast.

use the script, I think, right?

yeah, so we use, um, descrip for editing, and there's a variety of

kind of AI tools that are built into that.

Um, and also in sort of other parts of the process, I'll use chat GPT as well.

Um, I have used, what else have I use Claude.

Use Claude

Mm-hmm.

I highly recommend Claude.

Yes.

Yeah.

Seems to be better for more creative writing stuff.

The output seems to be,

It's better for a lot of

things, to be honest.

I I, I have used it to write code.

I've used it to, um, that we, I write all of my articles now using Claude.

It's, It's, amazing.

Um, I, yeah, I mentioned this before.

We, we, we start chatting, I've actually built an AI advisory board in my, for my

business, and that's powered by Claude.

So it's a, it's a project where I have custom instructions for all the roles

that I want my board to, to fulfill.

And so like I have, I have, uh, I have a marketing and a

marketing strategy director.

I have a, uh, architect, client advisor.

there's several, I think there's like 10 roles that are on my board.

And so whenever I have a question about my business, like, um, you know, or an idea,

Hey, you know what, if we tried this or, what, if we wanted to launch a new course

or something like that, I can go in and ask my board and it will, I can basically

dialogue with it and have a conversation with it, and it will give me advice.

And it's, it's, it's amazing.

It's absolutely amazing.

I, I highly recommend anybody that wants, you know, that runs a business.

You should have an AI advisory

I love that.

That is definitely going on my to-do list.

Yeah, it's, and I had AI help me build it, you

know, I was like, I, I don't know.

I don't even know what I, what roles I want on it,

you know?

So I said, I went into Claude and I said, Hey, I want, I want

to create an AI advisory board.

And it was like, oh, great, you know, and I, it gave me a whole list

of, you know, possible, uh, roles.

And then I was going through them and I was like, well, wait a minute.

You know, I, I don't think I need this one.

But it gave me other id, you know, I started thinking about it and

I was like, oh, wait a minute.

There's a couple missing here.

And I, I, and it was like, oh, that was great.

So it's like, so we ended up settling in on a few different roles.

And then, uh, what I did was I said, okay, this is great.

Here's my list of the roles that I want on this board.

Now I want you to create custom instructions for a Claude project that

will be able to power this, this board.

And it did, it gave me the custom instructions, and then I just created

a new project and put that in there.

And now I'm, you know, I'm putting other things in there, like more

information about our business

yeah,

so cool.

pretty cool.

a great idea.

Yeah.

for someone that's listening though.

Or watching this.

Um, 'cause hopefully by the point this one goes out, I will have actually sorted

out our YouTube channel and actually start posting these on YouTube as well.

That's a conversation for later, I guess.

But, um, for anyone that's, uh, that, that's listening to us today, that

the, that's feeling perhaps a bit overwhelmed, a bit skeptical about ai.

Maybe they've tried chat gt Yeah.

Can't even say it.

Maybe they've, maybe they've tried chat GPT once or twice and they didn't

really get great results from it.

Where should they start?

What would your recommendations be for them?

I think that anybody who, uh, is, is curious

about this stuff, you, you need to.

Start small, right?

Start.

And, and, and where I often recommend people start is with the tasks that

they know how to do really well, The tasks that you can describe really

well, the tasks that you know, what done looks like, that you, uh,

you have a clear process for that.

You have, you know, all the things that we talked about with teach if, if, if a task

kind of fits with within those parameters.

Um, you can clarify the role.

You can clarify, uh, you know, the what, what the task is.

The you have context that you can give the AI structure.

you can tell it what mistakes avoid to avoid and, and if it fits within that.

Teach framework of, you know, just being, being clear, uh, not requiring

your unique expertise necessarily, you know, being easy to explain, uh, have a

consistent, uh, or repeatable pattern.

Something you do often, these are the best tasks to start with, right?

And you don't have to like, you know, there's a lot of stuff out there in

YouTube about how you can, you know, hook up all these different tools and

automate every, you don't need to do that.

All you need to do is, is, is be able to use AI effectively at the start.

Like you need to gain confidence.

And so if you start with things you really know well and gain that confidence,

you're going to get more curious about it.

You're going to wanna know like, okay, well I can do this now.

Maybe I can do this.

It just, you know, allow yourself the time and the grace to make mistakes and,

you know, screw up sometimes and realize you're not gonna get results a hundred

percent all the time that you like.

What you're really aiming for is something that's in that 80 to 90% range, right?

If you can get AI to give you 80 to 90% of what you want, uh, and then

build on that consistently over time, make it a routine, develop a habit,

then you're gonna, you're gonna feel a lot more confident and a lot more

comfortable using AI on a regular basis, and you're gonna eventually see results.

I mean.

You know, we've been able to increase our profitability, but it took us,

it took us a while to get there.

You know, it didn't happen.

Like, oh, we discovered ai.

Now we're 174% more profitable.

It, you know, we, we've been using AI for years now, and we, you know, this, this

is something that developed over time.

So don't go into it with, you know, oh my God, this is gonna change my life.

Go into it with, you know, how can I, how can I use this

effectively in, in one task?

Master that, move on to another one.

You know, build incrementally, and that's how you're gonna fight

Yeah, so don't feel, don't feel any pressure to think like, oh my

goodness, I, I need to like completely reinvent everything in my business.

It's all gonna be AIed overnight.

You know, like you don't have to do that.

You said start small, pick one task and just build your confidence and

then iterate and improve as you go.

Um, I think that's really sensible advice for everybody.

The, the other thing I would also say is that if you're

using the free version of any tool, really consider upgrading to at least

the, the first level of paid, right?

Because, um, especially for a business, uh, when you, like, uh, for example

with chat GPT, um, if you're using chat GPT, uh, on a teams account, they will

not use your information for training.

Okay.

At least that's what they say.

Um, but you know, you have the expectation of privacy.

Same thing with like, if, if you have Google, if you have a Google Workspace

account, you have access to Gemini.

Gemini is amazing.

It's a great tool.

Um, it's not as much talked about as some of the other

tools, but you have Gemini, uh, in workspace, you have Notebook, lm you

have the paid version of Notebook, lm if you have a workspace account, which

is Notebook lm, we have not talked about it much, but it is an amazing tool.

Uh, it is just, just, uh, there's so much you can do with that.

Um, if you wanna learn a topic or if you're studying

for an exam, go just use LLM.

You'll be amazed what you can do.

But yeah, I mean, it, it, it, you'll find that you have more, um, capabilities with

the paid versions of the tools, uh, than you will with the, with the free version,

you know, and, and for $20 a month, I mean.

You know, I mean, here in here in Cleveland, that's like a couple of

cocktails on the weekend, you know, or that's, you know, that's maybe like

four Starbucks this week, you know?

Or it's not that bad.

Uh, you can afford 20 bucks.

That's great.

It's a great tip.

So you used ai, you, you've used AI to build out the workshop you talked about

recently, which was what we've based this conversation on today is based around the,

the content from that recent workshop.

Now that was about ai, but you practiced what you preach, like you

used AI to build out that workshop.

Um, that was slides, you know, worksheets, even like you, you

know, help with the speaker notes.

Could you tell us about how you pulled that off?

You know,

I had this idea about this workshop.

I was approached by America McKeel.

She's a, uh, uh, an architect, uh, in New York.

Uh, she's the chair of Cran this year.

And so she approached me at a, i a national and she says, Hey,

I want you, I want you to do a workshop at CRAN about ai.

She had been to, uh, a presentation I did a year ago, uh, at, at

Mark LePage's event in Nashville.

And so, um, I I, we had done a workshop on how to use she GPTs and

so, uh, or custom GPTs, excuse me.

And so she wanted, she was like, I want you to do a, a workshop on ai, but it

needs to be like really simple, you know?

She was like, she's like, most of the people in cran are not like.

Technically sophisticated.

I was like, that's kind of how she put it.

I was like, I was like, okay.

I was like, uh, so, so we just need something that everybody

can kind of follow along and do.

And so I, I recorded our call.

Okay.

Or, or, well, actually we met in person and I, I asked her, do you mind

if I record our, our session here?

And she's like, sure.

And so we were brainstorming.

We had, you know, there was some ideas that I came up with on the spot

and there were some things that she kind of wanted me to be aware of.

I took that recording and immediately put it into Claude.

And I said, I want you to, I want you to review this discussion and help me come

up with a concept for this workshop.

And it did, it gave me an outline.

I went back and forth with it.

We iterated, developed the outline, got the outline to where

I wanted it, sent it to America.

She, um.

She wrote me back.

She's like, this looks amazing.

I was like, great.

That's cool.

So then I took the outline and brought it back into the Claude, and I said, okay,

now I want to create a slide presentation and I want to think about exercises.

I want to think about worksheets.

I want to think about, you know, what are some examples that

we can showcase, et cetera.

Right.

And so iterated with AI to develop that presentation outline

basically the presentation content.

Once we got all of that figured out, and I was happy with it, then I took

that, that document, brought it into Gamma, which I, I mentioned earlier.

Uh, gamma is a AI powered presentation tool.

Uh, it blows away anything I've seen in at, in terms of like.

ease of use.

Um, so much easier than PowerPoint or Google Slides.

Uh, and it, it basically designs your presentation.

Like I have a, I I now have a, um, I have a theme that I've designed in Gamma.

So it has our fonts, it has, uh, our styles and colors.

It has our logo.

And so I put this, uh, text outline into Gamma and I said, I want

you to create this presentation.

It did, granted there was a lot of editing I had to do, but I would estimate that

that process alone normally would've taken me probably a week and a half

of, you know, if I were just trying to create all those slides from scratch

and, you know, write everything.

Once I had the presentation where I wanted it, it.

I know it was a lot less than if I had anything used it.

I, but once I had the presentation really refined and where I wanted it, then I went

back to Claude and I said, based on this presentation, I want you to help me to

create some speaker notes, which it did.

And then I just, I brought those into Gamma and, and that was

my presentation and I presented that live in Cran from Gamma.

I had Gamma run the presentation for me.

Um, yeah.

And it was, it was just so much more efficient.

I mean, I can tell you I've done a lot of workshops like this and a lot of

presentations, webinars, things like that, uh, is so much easier with ai.

That sounds like a huge time

saver.

Yeah,

absolutely.

So, um.

We need to start wrapping things up now, Brian.

So we've covered quite a lot here, shared some really great examples for everyone.

But what would be the main thing that you'd like everyone to

take away from our conversation?

Yeah.

I think it just goes back to what we were talking about.

You know, start simple.

Don't, don't look at AI like this.

I, you know, I think a lot of people have different opinions about

what AI is and what it's going to do and how it's gonna affect us.

I don't think those fears are unwarranted in any way.

I'm sure, you know, someday, uh, AI may be, you know, uh, you know, we,

we may all be running for our lives.

I don't know.

I don't think we're gonna see that in our lifetime.

But, uh, uh.

What I would say is that these tools are here, they're meant to be used,

uh, to improve our lives, right?

And so if you look at it that way, I think there's, you know, there's more of a glass

half, uh, you know, half full mentality of how do I, how do I actually employ this?

How do I get ahead of the curve?

You know, I've been to several events in the last month where

people have been asked rooms full of people, 300 people in a room.

How many of you are integrating AI in your business and maybe

20 people outta 300 raise their

hand?

Now you ask them if, if they're using GPT or chat, GPT or some other chat.

Yeah, most of the room raises their hand.

Are they actually integrating and implementing AI in their business?

Very, very few.

I've seen that at, I, I was just at May con, the marketing AI conference

last week, which is, you know, basically a lot of brands, uh, you

know, big brand companies, enterprise companies, and then, uh, a lot of

agencies, a lot of big agencies.

And to see that few people say we're integrating ai, I was shocked.

I was shocked.

Yeah.

get ahead of this stuff.

Don't, don't just wait for it to suddenly settle down because

it is not going to settle down.

I'm gonna tell you that right now.

It hasn't settled down since Jet GBT was launched, and I don't

see anything slowing this down.

It's only gonna get more, uh, widely used and pervasive and it.

And if you're not, if you're not figuring out how to integrate it

into your business, into your life, my wife uses it for everything.

She uses it to plan our meals.

She uses it to plan our trips.

She uses it to, you know, figure out our social schedule and

all the shows that we're going.

You know, she's, yeah.

So figure out how to use it somewhere in your life to improve your life.

Right.

That's great.

Thanks for sharing that.

Some great suggestions.

Um, was there anything else you wanted to add that we haven't

already covered about ai?

We've

I don't think so.

I mean, I think we covered everything we, we wanted to cover today.

This is a really good conversation.

I know we went a bit over time here, but, um, uh, I think, uh, I, I hope

people will get something out of this.

And, and if you're, what I call an AI skeptic, um, or, uh, uh, you know,

or, or you're just AI curious, use that curiosity, use that skepticism

to see if, you know, you can find something, uh, useful to do with ai,

right?

Uh, prove yourself wrong if you're a skeptic and, uh, fulfill your dream

if you're, uh, if you're curious of, you know, finding something that you

can replace, you know, with ai, right?

Um, not replace, but you know, improve,

Brian, thank you so much for returning to the show, uh, and

being a guest, sharing your expertise.

This has been really, really interesting, really helpful.

Um, just to remind everybody, where's the best place to connect with you online?

Yeah.

You can either go to our website, arc mark.co.co.

Um, so arc mark.co.

Or connect with me on LinkedIn.

Uh, Brian McCartney, B-R-Y-O-N.

I like to say Brian with a yo.

Uh, so you can connect with me on LinkedIn.

Uh, those are the best places to reach me.