Building the Future: AI Coding, Agentic Advisors and Custom Tools

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Loren Horsager (00:00):
That's the piece that now we get to pull out of
the AI models, keep in ourcompanies, and that allows us to
actually build out much, much,much more sophisticated
processes, but we still needpeople to run that.
It's it's somebody's assistantthat they need to run.

Joshua Schmidt (00:18):
Welcome to the audit presented by IT Audit
Labs.
My name is Joshua Schmidt, yourco-host and producer.
In the studio today, we haveEric Brown and coming from
Texas, Nick Mellon, the usualsuspects.
Big Tex.
How are you gents doing today?
Also, our guest, LaurenHorsauger from Model Mind AI,
good friend of the IT Audit Labsfolks here.

(00:39):
We've spent quite a bit of timewith uh Lauren over the last
year or so learning a lot aboutAI, how to integrate it into our
business.
Lauren is a great teacher and afriend of the show.
So thanks so much for comingon, Lauren.
How are you doing today?
Fantastic.

Thanks everybody for having me and excited to uh
talk AI today and how thatapplies to your business because
uh yeah, we've got to spend alot of time.
We we focus at Model Mind on AIcoaching and training.
And I think that's the criticalaspect to actually make this
work.
So I'm excited to talk with youguys about that today and

(01:13):
excited to hear what you guysthink has worked or not worked
with with what you've done aswell.

Yeah, we'll get into all that.
We're gonna do an icebreaker,but before we get into the
icebreaker, why don't you giveus just a little more background
on yourself, Lauren?
Um, anything you else you wantto bring up or that you're
working on currently, what's topof mind in the AI field?
Um, what's animating you today?

You caught me on a crazy day.
So uh I uh my history goes along way back.
So I've been doing AI since1993.
So this is not a new thing forme.
Um I actually have an economicsdegree and was working at the
Minneapolis Grant Exchange andwould go over what you call a
meetup now and uh at PiperJaffery Tower on Tuesday nights
and talk AI with some people.

(01:54):
And it was a very interestingthing.
It was so interesting to methat I quit my job a few months
later and got into IT andsoftware development and
consulting and uh been able toapply AI many places over the
years.
So exciting to see where we'reat today.
And uh I think I can't think ofanother industry that is so
like exciting with the changes,and you have to want to have

(02:17):
changes because uh literally twonights ago I uh got access to
another tool and started playingwith it, and I was blown away
and I haven't stopped since.
And um anyway.
Oh boy, which tool is that,Lauren?
Well, you know what?
Uh I I wasn't really gonnashare.
I'm not I'm we're we're we'restill learning about it.
So there's a lot of things.
How was it, Lauren?
How was it?
The tool the tool is Grockbot,and uh I it's very interesting

(02:42):
for a few reasons, um, but wecan talk more about that.
But it's it's just it it's Ihad this um well, about four
years ago, I'm part of a CEOgroup, and my CEO team, like,
hey, you know something aboutAI, can you tell us something
about it?
And that's kind of how it gotme into doing AI coaching and
training.
And uh which group are you partof there, Lauren?
Uh CEO Nexus.

(03:03):
And uh so it's big in Florida,but the guy that started it was
from Minnesota, so there's aboutfour groups in Minnesota.
We'd love if anybody'sinterested, I'd love to have
you.
It's a really he's the bestguide for uh CEO groups I've
I've ever worked with.
So um very exciting to workwith.
Is it similar to like an alliedmodel?
Yeah, it used to be an allied,and allied's great as well.

(03:24):
Um I I think they all havetheir things that they're really
good at.
Um and it it's very similar tothat.
Uh so yeah, yeah.
But uh it just depends whoyou're with and how that works,
you know, to really beeffective.

Eric Brown (03:35):
It's all who's in the room, absolutely.

Uh Lauren, have you done any AI trading?
I know uh did you put a fewjelly beans on the uh Robin Hood
AI, the uh the agentic trader?
And how is that how's thatgoing?
I I have not messed with it.
Okay.
No.
I did one.

I have not, but that that's that's partly what
got me into this way back.
Uh that's what we were tryingto do.
So I'm very fascinated.
I I haven't done that in awhile, but I I love the process
of how there's so many thingsthat were so challenging that
now AI solves, and we canactually think about the real
problem and and how we want tostructure a solution around that

(04:11):
versus like like it's the mathproblem, right?
It's like, but but I don't haveto understand all of the
variables in that.
I can think about how I want toapproach the how I want to
approach trading or I want tolike, and this applies to any
problem, which I think is justso cool.
It's like you don't have to youcan come in with a perspective
and then have the AI work withyou to craft like how you
structure that.

Eric Brown (04:32):
It's great.
Because and and I I want to getinto it a little bit more.
We're gonna get into it.
Oh, good, because I I'm anoptions guy, I'm an options
trader guy.
Are you?
Oh, cool, yeah.
So remember back in the daywhere you'd have like an iron
condor and you know you you knowyou're trying to figure out,
okay, what's the high going tobe in 30 days, what's the low,
and we're gonna try to you knowtrade in range.

(04:53):
Yep.
And you know, that's how you'regonna you know make your money
off of a spread or whatever.
Uh, but now you could say,okay, I I I'm interested in
whatever these particular eitherstocks are, or this uh index or
this industry is, and then youcould you could just take so
much data and AI could tell youwhat is the appropriate trade,

(05:16):
you know, in this sector rightnow, for example, and it's gonna
be things that you haven't eventhought of yet.
Because it used to be it'slike, all right, I got to study
this, I gotta figure out whatwhat its movement's gonna be.
I'm looking at candlesticks,all this stuff that now I don't
even have to think about.
I just put some money in and Ipress a button.

I want to learn more from you about that because
I'm very interested and I'vebeen I dabble.
But it sounds like Nick has gotsome agentic uh trading.
Two weeks ago, and it's almostdoubled my money.

Eric Brown (05:42):
No way.

Is it on Robert's?

Eric Brown (05:44):
Did you put a dollar in?

Nick Mellem (05:45):
No, I put 150 bucks in.
And it's almost doubled it.
Almost.

Eric Brown (05:50):
All right, what's what did you tell it?

What it what I asked for lower I'll I could
actually pull it up, but I said,you know, I want to buy as many
shares as I can with this $150to play with, and it came back.
I think it was called S SBetts, I think is the I'll I'll
look and put it in the in thechat here.
But this is this is onRobinhood, just to be clear.
It's on Robinhood, yeah.

Eric Brown (06:09):
So you told it I want to trade this 150.
I want to trade $150, so I wantto maximize my profits.
Good.

Yep, I think it's at 280 or whatever it is.
It's uh it's called 80%.
And did you narrow down whatwhat you allowed it to do or
anything?
So the guard it it won't placethe bet for me.
It's coming back to me saying,hey, let's do let's look at
buying this or selling this now,and or what industry you want
to look in.
And then you it narrows down,but it can't actually place the

(06:38):
bet for you, but it will comeback within my Claude agent, and
I can hit buy and it will goback and place place bets.

Eric Brown (06:45):
I'm going all in Vikings Super Bowl.
This is their year, guys.

I don't know about that.
Um we could do have a uh uh uha Viking, not a Vikings.
We do have a uh fantasyfootball league happening at
Ice.
Touchdown Browns alreadygetting heated and we haven't
even drafted yet.

Yeah.
Oh man, we could take that somany different ways.

Here we go with the icebreaker, right?
Today's icebreaker.
What is one random fact thatyou know?
Yeah, Nick.
Obscure.

Eric Brown (07:18):
You hear Nick clicking on the keyboard.
He's like looking at the randomfact that he knows.

I'm trading the Claude right now.
He's trade.
He's frantically.
He's tracking.
Okay, so I was actuallythinking about this, the obscure
fact.
So I had this ramble around inmy head, and I was like,
finally, I get to use this.
So you know the tall chef'swhite chef's hat that they wear,
and they've got all the pleatsin it.
Sure.
Apparently, there's a hundredpleats, and what that represents

(07:43):
is a hundred ways to prepare anegg.
Really?

Eric Brown (07:47):
Chef cordon blue.
All right.

So I'm hitting y'all with some truth today,
some knowledge.
Noodle on that.

Uh I'm a wordsmith, you know.
I love poetry, I love reading,I write songs.
So mine is a palindrome foryou, and it's a long one, and
here it is.
If you folks don't know what apalindrome is, it's something
that you can read the same wayforward says it reads the same
way backward, much like NickMellum's last name.
So my palindrome is a big word.

(08:15):
A man, a plan, a canal.
Panama.
There it is.
A man, a plan, a canal.
Panama.
Wow.
It's a palindrome.
It's the same forward as it isbackwards.
Wow.
Okay.
And the crowd goes mild.

Eric Brown (08:33):
Good one.
Good one.
So I my my mom lives with us,right?
She's 93, mild cognitiveimpairment.
I don't even know if we can usethe word mild anymore.
But um, last night, so um I shewas talking about how like
sunflowers weren't gettingwatered or something, and the

(08:55):
the the sprinkler was clearly onthem.
And I was like, well, obviouslythey're getting watered.
She looks at me and she goes,obviously, that's a big word.
So anyway, so mine, Lauren, andthen and then you can go here.
Um I'm I'm chatting with mybrother the other day.
He's he lives in Maryland.
His name's Rax.
He's uh uh an artist and taughtat Maryland Institute College

(09:20):
of Art.
So, you know, not a technologybackground, but man, does this
dude send me a lot of TikToksand other stuff I can't even get
to?
Um friendship homework, by theway.
The dude retired, right?
He retired, then like day two,he retired.
I'm just getting flooded withall this TikTok stuff.
But it it's good.

(09:41):
Some of it, there's some somegems in there.
Anyway, he goes, um Did youknow that if you count the
chirps from a cricket, you count14 or 15 seconds, and then add
40, it's going to tell you thedegrees in Fahrenheit that it is

(10:02):
outside.
So I'm like, okay, all right,whatever.
So I hang up and then I there'sI'm sitting outside and I
listen to a cricket chirp.
Sure enough, I count 29 chirpsin 14.63 seconds, right?
I got my phone out on time.
Wow.
And it was 69 degrees out.
Like I looked, you know, wow,boom, you know, uh temperature

(10:22):
69.
And then so I call him back,I'm like, all right, that's
crazy.
It worked.
And then he's like, Yeah, ifyou want to do the centigrade,
he's like, you count the chirpsfor 25 seconds, you divide it by
three, and then you add four,and I'm like, okay, all right,
I'm fine.

But wow, I can't wait to be retired.

All right, Lauren, you've had a little
you've had a little bit of time.

I don't feel like I've had enough time to
think about this, but here's mydad's had too much time to think
about exactly exactly.
My uh my thing, so I lovegardening and I think I'm fairly
good at gardening.
And I built uh, you know, I Imoved to Rochester, Minnesota a
few years back, and my wife andI we made built a big garden.
Uh well, not big, it's 10 by12, but it's all raised beds and

(11:08):
everything.
And I I'm questioning mygardening skills now, but uh I I
have been working on fightingwhite flies all summer.
Oh I I feel like we may havesolved the problem.
And I I went out and I bought asprayer, and you have to spray
the bottom of the leaves with uhuh like we're just using an

(11:29):
organic soap and water mix.
And I did it to my potatoesfirst because they were getting
eaten up with them, and itknocked all the leaves off the
potatoes.
I thought I killed them.
And a few days later they allcame back, and the white flies
weren't there, but then theywent on to everything else.
So I may have solved theproblem, and I it's not
something that other peopledon't necessarily know, but for

(11:50):
me in the gardening, well, it'sI never had to deal with this
before, and I have learned, so Ifeel like it's more like life
is a state of learning.
So what is a white fly?
It's just a tiny little bug.
And if you had seen me a fewdays ago, my neck was covered in
bites because I was outspraying and I just got eaten up
by them in like 10 minutes ofthe white flies?
Of big bites.
All of our tomatoes and stuffare like up six feet high, and

(12:17):
everything, everything goes upthe wall, up we have uh fences,
so you have to kind of likestick your head in and find
these things like when you're asyou're picking and spraying.
And I just got covered and Iwas like, wow, I didn't know
they bit, but I just got allthese bites from going out and
and working.
You're really getting up inthere, huh?

That's what you gotta do.
Okay.
Well, let's get into some AItalk here.
We got a lot to cover, and uh II know we these things go fast.
So I'm gonna do I'll just getus into the AI, and then Eric
and Nick, you guys can take itbecause we've done a lot of work
together.
But I want to just rewind theclock a little bit, um, Lauren,
to you know, ask you thequestion what was the moment you

(12:54):
knew AI had moved fromtheoretical to like actually
useful for clients?
And and you kind of hinted atit before, but when this was
something you wanted to moveinto full time.

I I think the one story, so uh obviously I've
been doing AI for a long time,so I I've known it's productive
for a long time.
I will tell you back in 1993when I was trying to do this in
finance, we never got it to workbecause we had 286 computers,
right?
So well, I tried that over thein the 90s, and since then I've
applied it with a lot ofclients.
But I think the big thing withgenerative AI specifically, uh,

(13:23):
I was uh I was doing softwaredevelopment for a company and uh
I was working on a project.
I was sitting at home and Itried uh I can't even remember
what it was called, but it was alibrary, basically an agent.
And you know, I had done alittle with chat and and
co-pilot and these differenttools that were out at the time,
but um, I I hadn't really usedit for software development.

(13:44):
And I set up this agent, andwhat I did, I created two
different agents.
So I had they were exactly thesame instructions, and these
were scarily weak instructions,um, but maybe a paragraph or two
of each.
But I made one change.
One was an optimisticdeveloper, and one was a
pessimistic developer.
And I'm an optimistic person.
I have a good friend that we'vedone a lot of projects

(14:05):
together.
He's very pessimistic and hedrives me crazy.
But at the same time, I haveone of those friends.
Yep.
It makes everything betterbecause he finds all the reasons
why something doesn't work andthen you gotta fix it before you
go live, right?
So I did this, and it was coolbecause at the time, this tool,
I can't remember what it'scalled, you could um use uh hook
it up to 11 labs and turn onvoice.
And so uh there was no like itdidn't know when to end or

(14:28):
anything, but I put in myproblem and I cranked up the
volume and I was running aroundthe house doing laundry and
other things while it wasrunning, and I was just
listening to it, and it washilarious because I was
listening to these twodevelopers go back and forth on
we need to do this.
No, it's good enough.
This is like this argument backand forth constantly.
In I let it run for four hours,which was way more than I had

(14:49):
to run ultimately, but I waskind of testing this process.
In four hours, I got 52 filesof code that all I had to do was
in two places go edit the theAPI key for that for the client
specifically, and that has beenin production since then.
This is like four years ago.
Oh, neat.
So it was a real reallypositive first experience, but
the dialogue of like how itthinks was so fascinating to

(15:13):
learn as you hear it negotiatingback and forth, and that's
actually what I'm doing todaywith with uh Grokbot is the same
process, right?
It's like when you get the AIsto talk to each other, that's
where it gets really powerful.
Because uh, you know, if I if Iask one person to do something,
it's not like they no no oneperson can solve all problems,

(15:33):
right?
And we all get better when wework in a team, and it's it's
the same with AI, and sometimesthat means we're teamed with
with AI and humans.
Sometimes it can be all AIteamed, but we need the team
aspect where we're approachingthings from a different angle to
get a better result.

Eric Brown (15:48):
I I was just talking to a to a developer today,
Lauren, and and he was talkingabout a skill that he he wrote.
I think he he borrowed it fromthe internet and modified it,
but it was the council approachwhere you have a council, you
have a chairman of the councilor a chairperson or a chair bot
of the council, if you will, andthen you have different members

(16:09):
of the council.
So you've got a contrarian,you've got a newbie, you've got
a first principles thinker.
Um there's two other roles, Iforget what they are, but you
form this council, and then youask it a problem that, you know,
not like a yes or no question,but something that you're really
trying to wrestle with, send itto the council, AI council, and
then it'll give you back, ithands it back to the to the

(16:32):
chairman, and uh, and then thethe the chair bot consolidates
it and then sends it to you.
And I'm I'm sure you know whatI'm talking about here and
probably have done somethinglike it.

Yeah, we actually teach a class.
We used to teach a class, and Ido do a lot of public
demonstrations using that as anexample.
Um, because I talk a lot toCEOs, and so it's something that
they all want is like, how do Iget strategic feedback on a
thought or or something that alot of times they maybe feel
that their team can technicallyexecute, but can can you get
strategic feedback on things?

(17:03):
So having a team AI approachlike this is really powerful.
There's actually a company thatum supposedly has built a board
of directors um from AI agents,and that's how they run their
whole company by having a boardof directors that is AI running
that.
And in class, what we would dois we would say, Hey, go pick
five people who you'd want onyour board and add them to the

(17:26):
set of agents, because the nicething about that is I can say,
Hey, I want you know Tim Cookfrom Apple, and I want um Geno
Wickman, who who's you know theEOS guy.
And so like you can tip pickthe different perspectives you
want and then say, now use them,propose a question, and give me
strategy, and they're all gonnaapproach it from their own
perspective.
And you can then have like a uhwhat we usually do is have one

(17:50):
kind of um like a boardcoordinator that's kind of who
you're working through, and thenthey'll go off and pass that to
everybody else.
But um, what we're seeing withGrokbot right now, and this is
the big thing I'm testing, ishow do we do a team?
Um, like I set up 46 differentagents in Grok Bot yesterday,
and about 10 of them aremarketing different marketing
roles.
And I was able to tell it, hey,use the marketing team and

(18:14):
before you actually give backanything, have the whole
marketing team give theirperspective on this, right?
So it's that same concept wherewe can use a board of advisors
to solve every problem we'redoing with AI, which the
interesting thing is they alluse their perspective and
question things differently, andthe results go up and your
inaccuracies go way down becausethey're looking at it

(18:35):
repeatedly and differently.

We've been doing loop prompting here and we've
been doing some uh AI projectsaround.
Um so I'm just wondering whatuh if we could explore that
intersection uh that we'velearned together by working with
you, Lauren, and maybe we canbat some some ideas around
between you and Eric and Nick oflike what that intersection is
between AI and then security.
Uh we're security-focusedcompany, right?

(18:58):
You're an AI security-focusedcompany, but we have a lot of
crossover there.
So um I just thought it wouldbe cool for our guests to kind
of get into some where some ofthat intersection.
I've got some contrarian views,so I'll go last.

Eric Brown (19:10):
I think you should go first.
Maybe you had you gotta get inthere with that.
Well, uh yeah, yeah, we we haveuh where where we are today
with middleware software, I meanwe've just got just kind of
expanded with bloat all over theplace, right?
Like you you just can't uheverywhere you turn around,

(19:31):
you've got maybe 15 differentproducts that'll solve the same
thing.
And they're all trying to get amonthly recurring revenue from
you.
Um one of the the passion areasthere is the CRM, which is just
you know essentially a databasewith a a front end on it.
And i I I don't know, like Idon't know that anybody says

(19:54):
that they really love their CRM,right?
It's and and you're you'repaying lots of money, right?
In in some cases, easily tensof thousands a month over a
hundred thousand a year for thisfront-end database.
And Lauren, I I know you took aunique approach on this and and
built your own in a matter ofwhat days you you got the thing

(20:18):
built?

Days, and now we continue to customize it.
Yep.
I I I think what's interestingabout that though, Eric, is like
the CRM is like we're I don'tfeel like we're unique anymore.
We have these for a lot ofcustomers, and we have lots of
other customers that we've givensome advice and they've gone
off and built their own.
But what's interesting aboutsoftware is I don't feel like
the interface matters anymore.

(20:40):
So nobody likes our CRM, notbecause it's not useful to track
in information and have theinformation in there, but the
interface is terrible for howeach individual works
differently and it's notdesigned for them, right?
And so what if you don't everneed to touch the interface, but
you still get to use all thelogic and the security and the
process that's involved in a CRMor any other solution through

(21:03):
your AI.
So we have built this system,and it turns out we hardly ever
use the interface to it, butwe're all using it constantly
through our through our MCP, ourAI interface to that, right?
So we're all using it inmultiple different AI tools to
do things.
I just wrote a bunch of contentfor next month that we're gonna

(21:24):
be publishing, and it goes outthere first.
And I it just the AI posted itout there, and I did bring it up
just to make sure everybodycould see because we're tweaking
a couple of settings that we'readding.
But other than that, it's likeI don't even go out there and
use the tool, I just go throughthe AI to talk to the tool, and
then it also has access to allthe other tools that I'm using.
So it's like there's no silosbetween anything anymore, uh,

(21:45):
and you don't really need aninterface.
So what that means then is ifyou're thinking about solutions,
uh build a really good databasefor whatever your process is,
and then work with the AI todesign the process around that
and the security around that.
You have a working solutionthen.
Which I think is reallyinteresting because we just
don't need these complex UIslike we used to in the past.

(22:08):
And honestly, that's a hugepart of the development usually.

Eric Brown (22:13):
Well, the UI could even be on the fly, right?
So a UI that I might like mightbe different from Josh or Nick
or you.
And if that if you have thatchat interface to the data, and
then we're going to have ameeting and we're going to look
at CRM data, for example, it inalmost near real time, it could
pull up the data and present itin a view that it knows that I

(22:35):
like, which is different fromeveryone else, but we're still
talking about the same thing.
So I think that's really cool.
And the companies that are inthis space now, in that
middleware space, where they'regoing after that monthly
recurring revenue like aSalesforce, it might be time to
get that AI trader out and shortsome of that stock and or sell
those, sell those puts.

(22:57):
Because I can't imagine whyanybody would be dumping money
into Salesforce now wherethey're coming out with like
this agent force, but how how isthat differentiated from
anything you could get for $20 amonth?
Right.

And do you do you want to use the AI in that
tool or do you want to use thethat connects to all of your
tools, which you can Nick, thisis a big opportunity, I think,
for the creative team because umor creative minds, right?

Because we've we've seen a lot of like graphic
designers and stuff displaced.
But now that everybody candesign things, I think what ends
up happening so far, and thisis gonna change, but um making a
case from some of my own workhere, you know, the AI makes a
web page or it makes uh you knowwhatever, uh an advertisement
or something.
It's a lot like how we in a zoowe make a you know a landscape

(23:49):
for the tigers to live in or thepenguins.
And at first glance, it lookskind of impressive.
And it looks like theatmosphere and the temperature,
and they got the water and youknow, they got all the elements
that would be in their habitat.
But the more you look at it,the more it's like, oh, that's
kind of it, it's a you know,it's it's kind of just our human
take on what their habitat is.
Whereas the AI is giving uskind of their AI take on what

(24:10):
the website should look like.
So I think that's a greatopportunity to come in and
customize things with that kindof graphic design eye or that
creative team eye to really dialthings in.
And I think that will be a bighurdle for for people that are
just vibe coding things orwhatever.
That's like that they don'thave that team coming in to make
sure that it's running smoothlyor the UX.
It's like when you go up to adoor and it has a bar on it and

(24:33):
you expect to push it and it andit doesn't do, it doesn't move,
right?
And you're all of a suddensupposed to pull.
It's just not intuitive.

Eric Brown (24:38):
So I don't go to zoos, Josh.
That's like a animal jail.
It's disgusting.

Yeah, I know.
You don't go to the barn thingat the state fair.
Yeah, we learned I don't dothat either.
The birthing barn, the miracleof life.
Some people go in there to eatlunch, you know.

Well, this is like this is the same kind of
conversation I think I've hadwith about five clients over the
past two weeks, three weeks.
Everybody's subscribed to thisthinking that they have to go
buy something versus build it.
And I one specific organizationI was talking to over the past
couple of weeks was like, well,you know, talking to them about
building their own tool, butthey go back, well, you know,

(25:17):
that's that's not on our list.
Like we have a list that's beenapproved of approved vendors
and this whole grandiose changemanagement things that they
they've got put themselves in inlike a jail, like where they
they can't do anything.
They've they've restrictedthemselves and then they've
restricted their thinking.
So now everybody, especiallythese big organizations, they've

(25:37):
subscribed to that only, thatthey can only do that, they can
only work in these confines, butthere's all these
organizations, IT hot labs,Lauren, like we we work together
and you have great partnershipsthat can build these tools and
and and host these tools.
The other thing that I've beenlistening to, and I was on a
user group, um, AI user groupover the past couple months.

(25:58):
And um a lot of people,everybody's just like a
developer now, right?
Like, oh, I can go in here andI can build these things.
And that's that's really cool.
Like go in there and do that,learn these things, because it's
probably a big part of thefuture.
But then what?
What are you doing?
You get the 80%, but what'sthat last 20%?
What are you doing when you getthere?
And I've been noticing in thisuser group specifically, and I'm

(26:20):
kind of going down two rabbitholes, but follow me here if you
will, that what do you do?
What's that last 20%, right?
Because if everybody says theyget there, but then they come
back with the question is howare we hosting it?
How are we securing it?
How are we connecting it backto our company memory?
Where's the data going?
What data's going there?
Right.
And then how is it beingaudited, compliance, et cetera?

(26:40):
How are people, you know,there's just a whole suit of
questions.
But I was just noticing a lotof trends of a the last 20%.
And then the first thing I wastalking about was the thinking
that everybody's around changemanagement and actually building
a tool or just buyingsomething.
Because for Eric Sorry talkedabout it with the CRM, if you

(27:01):
buy HubSpot, maybe use 60% ofit.
But if you build it, you use100% of it, and it's exactly the
flavor you want that might bedifferent every week.

Or you build it and you realize that, oh, we
built it because we used HubSpotand we built it like HubSpot
and realize that's not what wewant.
We changed it.
And oh, like I could change itevery day or every week because
it's so easy.
And one thing you mentioned,Nick, is like, hey, you know,
work with us to build this.
We we don't actually want tohelp, we don't want to build it
for you.
This is so easy that you can doit yourself.

(27:33):
And what I mean by that, thisthis shocks people.
And the whole term about vibecoding, I I am it just makes my
skin cringe.
But I I the concept is real,and I can give you a whole list
of people, including CEOs ofcompanies, that have sat down
for a weekend with they gothroughout class and then

(27:53):
they'll they will go sit downand they will do build
something.
And we have a client that umdrove home four hours uh from a
meeting uh that he went throughour class, he then built
something and it went live onMonday.
He brought it in on Monday andsaid, Here's what we're gonna
do.
This is live, and it's beenlive since January.
Right.
So it's like you just have tohave the business knowledge.

(28:13):
You do not need the technicalknowledge that the AI can help
us through that process.
Now, we strongly recommend youtalk to someone that understands
security about how you'rerolling something out before you
actually publish this.
But the process for building itis so easy.
It's it's really just can youdescribe what you want?

Eric Brown (28:31):
What's the coolest thing you've seen someone build,
Lauren, or yourself?

I I think right now our solution is is uh it's
really complex.
It's not just a CRM.
So we built out our whole CRMand then we've added all of our
training classes on there, allthe scheduling.
And now we just rolled out anew website a week and a half
ago.
And what's great is thatwebsite is completely hooked up
to this.
So our my operations person,when she schedules a new class,

(28:56):
she just goes and puts it inwhere it's all the scheduling
about what trainer can we put onthis and when is it going to
be?
Where is it?
Is it scheduled?
Is do we have the roomsreserved, like everything like
that?
And it just shows up on thewebsite.
And is it open class or is itat private to some company that
we're doing it?
Like it's just all automatic.
Um, and then we do coaching,right?
So what I love about this is Ican go out and meet, like when

(29:19):
we were meeting, it's like wesit down, we have a
conversation, we try to recordthose as much as possible.
And then I just talk to my AIafterwards.
I'm like, hey, you know, logthis meeting, create it, create,
create any to-dos that come outof that.
Um, and I do all that throughmy AI.
So I can be in my car and I cansay, hey, I'm driving up to IT
Audit Labs.
What do I need to know fortoday's meeting?
What got done?

(29:40):
What are we waiting on?
What are the questions thatneed to get asked this time,
right?
And so it's just a database,right?
It's really just a databasethat holds us, but we have to
think about the structure.
And to be honest, I don't evenknow what the structure looks
like.
And I created it, but the AIcreated it and I just described
all my problems.
And then I had to do some umreviews on that to say, hey, how
would you design this better?

(30:01):
How would you improve it sothat it and it'll do that?
And then I worked on thesecurity side of it, say, hey,
we want to make sure it's secureso this person can't get to
other, you know, you have tohave access to the right things
to be able to do it.
We ran tests to verify allthat, but uh, it will do it for
you if you can describe theproblem.

So this goes into scope, right?
You know, you can kind of wetalked about a lot of fun
projects here and we've tackleda lot of them.
And um, you were kind ofmentioning how you can just
build these things in a weekend,right?
And they can go live.
But I think Nick's also right,you know, there's a lot of big
ideas and they they get 90% ofthe way done and they just don't
get implemented.
And and what do you see whenyou're teaching these classes as

(30:39):
that thing that keeps peoplefrom finishing that project or
hooking that up to that becausethat's the part that's hard for
me too, right?
Like I'm an idea guy, I'm acreative, I I can get I can get
90% of the way there, but then Iget stuck on how do I hook this
up to my notion or how do I logall this stuff so that it's in
a rag model.
You know, that's that's thedeveloper side of the brain I

(31:00):
don't have.
So what do you see being thething that really gets people
unstuck or moves this from beinga theoretical, you know, money
uh problem solver, moneymakerinto like reality?

Uh uh the number one thing we see is that people
bite off too much.
So if you're trying to do thisscope and it's just that there
ends up being too manyquestions, it's not that
technically you can't do it.
You just can't get everybody toagree on that.
If you start by one or twopieces of that, say, hey, let's
do this, everybody agrees, thenyou add on the next thing and
the next thing and the nextthing.

(31:32):
It is way, way easier becausewhat I've learned in all my
years of doing softwaredevelopment and working with
clients is that very few peoplecan conceptualize a problem in a
way that it's like, hey, here'show the UI is going to work or
here's how it should work.
They have to touch it, feel itin order to actually understand
how they should use it.
And so we spend way too muchtime working on UIs and stuff.

(31:57):
And instead, it's like, whatare we trying to solve?
What is the data we need tocollect?
What are the processes we needto run?
And let's start with a smallset of those and then add on to
them.
And if you do that and you takethat first one all the way to
production, so it's inproduction, now you have a
system that people will continueto add to.
If you don't take that, yousay, hey, we did we did these
first two pieces, but now we'regonna add on the next thing and

(32:17):
the next thing before we everproductionize this.
That's that's gonna kill it.
So get it all the way toproduction, then you start to
add to it and add to it.
And it's not that hard to addto it, but people come along
with you then because they cansee, I want it to work like that
piece, but it has to do thisfunction, right?
And they start to see whereyou're going with that, and it's
way, way easier to bringeverybody along with you.

That's good advice.
Yeah, Lauren, actually, I thinkyou might have been on my, I
think you might have been on thecall I had with another
organization where it soundslike very similar to the one I
had.
And uh, they were trying tosolve four or five different
problems.
It was a finance organizationspecifically, and they were
trying to solve, like I said,four or five problems, and
things kept breaking, right?
They weren't working right.

(32:59):
And, you know, I think myadvice was, well, what if you
just solved one of thoseproblems first?
Just solve one of the problemsfirst, and then continue to add
on to it.
And then you get the you solvewhat's breaking, right?
You see what's being introducedto the environment.
We can roll things back, makesthings easier.
Instead of throwing, I I thinkwhat happens is people catch the
bug, right?
Like they get into Cloud andthen you can give it a million

(33:22):
things and it will chew on it,and it'll chew up tokens, right?
And it'll get through there andyou'll get this cool looking
thing, but it might not worklike you thought.
But if you just scaled it backto maybe one or two things, then
it works, and then keep goingfrom there.

Eric Brown (33:35):
That's why Nick's rate's a thousand dollars an
hour, right?
Just that consultant.
Just work on one problem, andthen he hands them an hour.

And it will say it's not a save view.
It's easy.
Like they'll they'll they'lltell you all five of those
problems are easy, but you putthem all together and they're no
longer easy, right?
Mm-hmm.
Right.
So I I think that's uh verygood advice and uh number one
thing that we see as as anissue.

Eric Brown (33:57):
So Lauren, we uh recently won an award for uh uh
best places to work in the TwinCities.
So I saw that.
And uh congratulations.
Yeah, yeah, thanks.
Josh mentioned um Sam was overat Top Golf, right?
So um the other day the awardceremony was last night, and

(34:17):
then we we went out to um to todinner afterwards with uh with a
small group from the company.
And so they're like, Eric, weshould go to Top Golf tomorrow.
And I'm like, should we?
And yeah, why don't we whydon't we get a tea time at two?
I'm like, well, I got apodcast.
Um and they're like, you know,we did win best places to work,

(34:38):
and there is this work-lifebalance.
So I feel like I'm getting likethis whole thing held over me.
So I don't know how long it'severy week it's gonna be
something.
Like we we need to have icecream party, right?
You know, we want to win againnext year, don't we?
No, we don't.

Well, I feel if if if others are like Sam eyes
on glass, I feel like theydeserve it because uh I I get
the emails from Sam at two inthe morning.

Eric Brown (35:00):
So Well, I I mean, you know, can we can Sam be
replaced with an AI agent?

No, no, no, no, no, no, no.
We're deleting that.
We gotta delete Sam.
You didn't hear anything.

Eric Brown (35:10):
The AI is not an asking for ice cream socials.

Okay.
I wasn't good to bring this up,but I'm not now I gotta bring
this up.
But you already have the golfgloves in the car ready to go.
We've learned that when you'rekayaking.

Oh boy, you gotta keep waiting for teaming
up.
This is worth talking about alittle bit.
So the gloves?
No, not the gloves.
I I think in this AI world,like everybody expects that we
won't have um we won't havedevelopers.
And that's just it's be it'svery much proven to not be true.
The role of a developer changesa lot, and this is the same

(35:48):
with every other role that'susing AI.
The role changes a lot, andwhat we need is more critical
thinkers.
And anybody that works withschools and like we love, we
have an internship program thatwe love, and it's like how we do
that really matters how wetrain people up and bring them
on because we need smart peopleto participate in these

(36:08):
processes.
The AI does nothing by itself.
The only way it does it byitself is when we're automating
something.
And but someone had to tell ithow we're gonna automate, what
we're gonna do, how we're gonnado that.
Once we do that, the AI can runthat.
But if it ever needs to change,we needed somebody to help.
Yeah, you know, how do we wantit to change?
What are we gonna integrate itwith?
How are we gonna do this?
Somebody has to think aboutthat and understand the process.

Eric Brown (36:30):
Lauren, you're so right on that.
And it's disappointing now.
Some of the things I'm hearingfrom developers where they're
anti-AI.
So I was in a uh a CEO meetupyesterday, and and one of the
people was talking about how thedeveloper in their company is
anti-AI and resistant to AI.
And in the back of my mind, I'mthinking, like, okay, that

(36:51):
dude's gonna be washing dishesat Culver's in, you know, about
five minutes.
Because you can't be allowed tobe a lot of people.
Right.
I mean, if you're a developer,like this is the time to lean
all the way in and embrace itand be the voice in your company
of how are we going to takesome of the routine things that
we're doing, and not to displacea workforce, but how do we make

(37:13):
our our jobs better?
How do how do we get you knowto a solution faster or whatever
it is the company is iteratingon so that you know we can take
people's workload and maybe makeit easier for them so they can
do other things.
But it it just really saddenedme that someone who's a
developer and is already acritical thinker is so resistant
to the technology that is notgoing anywhere.

It's an odd thing because tech people are
almost always ahead of every newtechnology that comes out.
In this case, they're about thehardest people we have to train
are tech people.
They they just they'reresistant, they don't want to
come along.
And once they see it and theystart to realize that no, I'm
part of the process, um, it theytheir head, their their mind
changes, but until they see thatthey are very resistant because

(37:59):
they're afraid of their job.
And they they they rightfullyshould be.
Like they're it's changingdrastically.
But if you're the one thatunderstands and understands the
system, you're critical to thatprocess.
The AI is just going toaccelerate what you can do.

I was gonna say maybe it's because the CEOs are
making fun of uh the fact thatthey might be replaced.
That has might that might havesomething to do with it.
Yeah, the the general uh uhsense of of of angst.

But if you go to top golf, we're building an AI
agent to replace you.

Um but I I wanted to I want to touch on
that because this is I know thisis an emotionally charged, but
it's it's a great topic.
And I know we all have a littledifferent angles on this.
It's just that that pushbackthat is in the zeitgeist right
now against AI.
That's a big word.
Thank you.
And then, you know, I'm always,you know, you've got a couple
in there too, so I got to keepup with you.
But you know, we've seen it onsocial media and and then we've

(38:53):
seen it in some companies likeFord, Walmart.
You know, we they went reallybullish with the AI, replaced a
bunch of people and ended uphiring people back.
And it was a really expensivelesson for them to learn.
So I just wanted to get, youknow, Lauren's take on that, the
FUD in the market, if you will.
And um, where do you see thatgoing?
And how do you see that kind ofit sounds like we're all could

(39:13):
agree that's gonna be abalancing act, right?
Because we need people, we dodefinitely want to lean into AI
and we love AI here.
So how do you how do you seethat all panning out?

Well, let me tell you here, here's our our um
most important class we teachright now is uh our you guys
have gone through this, but wecall it now uh build your AI
assistant.
And it was our processengineering class.
So we get into a lot ofprocess.
But what what's interestingabout this, and I actually have
a couple questions about it foryou guys as well, but the the

(39:43):
process is basically taking allthe knowledge and the decisions
that we're making as we'rehaving chat conversations and
building an assistant that is nolonger tied to any one AI
model.
Um, instead, so right now whathappens?
Let's say I go out to to Claudeand I have a conversation and I
make some decisions along theway of no, I want you to do

(40:04):
this, I don't want you to dothat, um, in whatever in area
role we're we're talking about.
It doesn't really matter.
All of that logic and the thecriticalness of like what
decisions were made, whatknowledge was captured about
that, all lives in our chathistory.
The goal is to take all thatout of chat and actually put it

(40:24):
into files that live somewherethat we own and can manage.
And it doesn't matter what youknow, am I using quad code?
Am I using you know chat GPT?
Am I using cursor?
What it doesn't matter whattool I'm using anymore.
The same knowledge is mycompany knowledge that I have.
Now, what's cool about thisprocess somebody has to own that

(40:45):
and run that.
This is where the people partof it comes in.
So we need people that can runthese assistants.
The goal is the day-to-daytasks that you repeat over and
over again, you have to be lessinvolved in from a tactical
perspective.
You get to be more involvedfrom a strategic perspective.
So we want you thinking aboutwhy do we do this process this

(41:08):
way?
Is this the right way to thinkabout it?
But when I'm gonna run it,maybe I don't have to do most of
the steps.
The AI can do most of thesesteps for me automatically,
right?
You can take that and you canactually pass it to someone
else.
So we're working with ahealthcare company in Minnesota
right now that um they ran acertain group of people through
our training initially, and theybuilt all these different

(41:29):
processes.
And now they're starting totrain up kind of the next level
of managers and and executors.
And what they've been able todo is say, hey, I already built
out the process.
This is you need to orderproduct every week, or you need
to review um all of ourinsurance payments every week.
Well, here's the process.
I can actually take the skilland all the knowledge and hand
that off to you because it's adocument in our company that we

(41:52):
control and we own.
So uh this goes hand in handwith people talk about the
security around AI models andnot, you know, don't train on my
models and stuff.
It's not really the data thatyou're putting in that usually
is actually the biggest thing.
It's the decisions you make andthe rules you're creating as
you're making those decisions.
But that's the piece that nowwe get to pull out of the AI
models, keep in our companies,and that allows us to actually

(42:16):
build out much, much, much moresophisticated processes, but we
still need people to run that.
It's it's somebody's assistantthat they need to run.

Eric Brown (42:24):
I I like the concept of the second brain, but
sometimes we some people need afirst brain.
I agree.
Uh the and and that concept,Lauren, is pretty cool.
And while you were explainingthat, I I was just thinking, you
know, as humans, we we've justkind of been programmed to to
show up, perform a functionsometimes almost robotically,

(42:49):
and then you know, we we we gohome at the end of the day.
And AI is really freeing us up.
And I'm, you know, I'm notgonna jump on the AI bandwagon,
so to speak.
You know, I think it's got atime and a place, but it it is
freeing people up in certainroles to do more of that
critical thinking and be freethinkers and be more strategic.
Where you come in to to youknow, perform your function that

(43:10):
you were hired to do.
And like, well, how can I dothis better?
How can I challenge myself tothink differently?
And here's a suite of toolsthat I can use to take some of
the repetitive stuff that, youknow, I don't think humans are
particularly skilled at doingthat repetitive, brainless work,
right?
I mean, we're not built forthat.
So it it is a way to kind ofraise the bar for all of us to

(43:32):
do things in a in a differentway, but we just have to embrace
that.
And and in some cases, it'scompanies making that okay for
people to embrace and change theway in in which they work.
And it it probably starts muchlower down in the education
system, you know, like where thethe Montessori method uh of
education, where I think thatthat educational model does

(43:57):
teach people more of thatcritical and free thinking than
just.
Kind of the, okay, I'm going todo math for 50 minutes.
Now I'm going to go do Englishfor 50 minutes, and I have to
know these certain things, whichif you think about all of the
math and things like that thatyou were in, it's like, well,
why did I need to spend so manyyears memorizing formulas that

(44:19):
is really not serving meanything these days, right?
Like I don't I went throughCalc 3.
I don't still remember any ofthat stuff, right?
That that stuff was out thewindow the day I took the final.
So I don't need to know how tocalculate the surface of a 3D
object, right?

Are you sure it didn't help you think about
things in a different way,though, or give you a new a new
chamber in your mind to kind oftap into for other problems?

Eric Brown (44:47):
I think it it may have, but I don't know that I
needed the forced memorizationof formulas to do that.
I remember in in one of thelike uh physics classes I went
to, we went to an amusement parkand we took some tools and we
were meant measuring like the Gforce on different rides in the
amusement park.
And I think I got more out ofthe year long of physics from

(45:08):
that one afternoon than I didfrom anything else I did in that
class.
Because what was I reallylearning until it could be
applied?

Yeah, I couldn't agree more.
I was just on a podcast and Ijust had a clip that went out
today that says, don't pay formusic college because you could
take, you know, uh an eighth ofwhat you spent on your music
college and go make an album ina studio or go have an
experience and learn just asmuch.
And so same concept.
But but you know, I think a lotof that just is antiquated

(45:37):
educational things, and thenthey gotta stretch it out so
they can get the money out ofyou.
But I think really what we'retalking drilling into here is
just uh just how it can free ourtime up, right?
And um, how we don't have to dothose repetitive tasks anymore.
And it's gonna hopefully freeus up to be a lot more creative
and a lot more forward thinking.
And and then to Lauren's pointabout you know, freeing up our

(45:58):
workforce to be a little morestrategic and a little more
top-level thinking instead ofjust going through the motions,
right?
So when we're doing that, we'regoing fast, we're making new
stuff, how do we make sure we'reintegrating good cybersecurity,
good uh security measures?
And and where do you see uhwhat you know when to bring in
security people, Lauren, whenyou're integrating these new

(46:20):
concepts into businesses early?

Uh I I I think I think that what's interesting
about security, so we we talkabout governance some and we
don't do like hardcoregovernance and stuff around
this, but uh uh the there's acouple things that I always tell
businesses, and we work with alot of well, we work with large
businesses and a lot of smallbusinesses, and um people are
are scared, but they also knowthey need to do something.

(46:45):
And so the question is how howdo you move forward?
So there's a few differentthings to think about um from
our perspective.
I think you guys will add awhole bunch more on top of this.
But the the first critical one,I think right now people are
very concerned about the modelsand having access to your
information.
I think this is way overblown.

(47:07):
And because they're soconcerned about this, they're
missing a whole bunch of otherthings that they really ought to
be concerned about.
The way we connect our AImodels to different tools that
we want to connect to is throughMCPs now.
It's like an API for AI, right?
And there was a guy, I thinkhis name is John, and he
published one to QuickBooks.
And this was like, I mean, MCPshave only been around maybe

(47:30):
eight to ten months or somethingnow.
And when they came out, um, ittook a little bit for Intuit,
who owns QuickBooks, to kind ofput theirs out there.
And so if you did a Googlesearch, and I think this has
actually changed just recently,but if you did a Google search,
the first one that came up toQuick to connect to QuickBooks
was John's server.
And the scary part about thatis it worked, right?

(47:52):
It would work.
I could use his tool and Icould connect to QuickBooks.
And the scary part was all ofmy QuickBooks data was going
through his server.
I don't know if he was doinganything uh nefarious or not.
I don't think so, but that'sthe risk, right?
I don't know where my data ismoving in a lot of cases.
And this is why what we do whenwe work with a team, we
generally try to work in teamsof five with an IT person on

(48:13):
every team.
And the reason is anythingsomeone wants to build, like we
we have these people that havebrilliant ideas and they want to
build this.
And then the IT person says,hey, hold on, what do you want?
It's like, okay, we can figurethat out.
Or no, don't do that, andhere's why, and explain that,
hey, this violates our companypolicy or this this shouldn't

(48:34):
happen because of these securityconcerns.
So we want someone along theprocess thinking about security.
Um, if someone doesn't havethat, we recommend they go hire
someone to do that, and it canbe a consultant or someone like
you guys or whatever, but theyneed to have someone thinking
about the security implicationsof the decisions they're making.
Um, because there's so manythings that are so easy to do

(48:55):
now, which means we can doreally dumb things really
easily.

Well, I know Eric and uh Nick have been doing
a lot of thinking on this andhow we can help folks with their
AI integration and not only howwe can help them integrate it,
but how so we can make it betterand improve upon that.
So it's your time to shine,boys.

Eric Brown (49:12):
Yeah, Nick.

Well, I think it's yeah, I mean, I think the
conversation starts is what doyou want?
What what do you how do youmake your data work for you?
You know, but I think a lot oforganizations probably haven't
even started that thoughtprocess of what data, what what
do they have already that canwork for them?
Right.
Like I I again, I was in thatuser group.
And I think the I think if Iremember right, it was a garage

(49:34):
door organization.
They repair and put in garagedoors at new homes, old homes,
whatever it is.
And then I posed the questionthat he he is he's building an
app for his salespeople to getquotes.
And then, you know, my questionback to him was well, does the
does your app have the abilitythen to text the client for
payment?
Can it text them to say, hey,put a Google review in?

(49:56):
Does it connect back to theCRM?
Like then, so your marketingteam then can do like a DRIMP
campaign, you know, out to theseareas that maybe are more
focused, or to an organizationthat might have 150 part
numbers, but maybe only 20 ofthem make the revenue and
they're just wasting a bunch ofmoney on the other parts.
Well, maybe just get rid ofthose, delete, and then focus on

(50:16):
these.
And these are this is justabout the data that the company
already has that they're notusing.
But then the conversation goesto also, you know, the
governance.
Do you have an acceptable usepolicy, right?
Those are the more not as funconversations to have.
But I think it's veryinteresting to see organizations
sitting on this pile, stockpileof data, and and they're not

(50:37):
making it work for them.

Eric Brown (50:40):
Yeah.
And one of the things that thatwe're doing, Lauren, along the
lines of what you're talkingabout there is, and I think what
you're doing out, you know,educating people of, you know,
how do we, it's it's not scary,how do we approach this?
How do we start thinking aboutit?
We've gone in and and helpedorganizations take these
disparate sources of data.
Maybe you've got some stuff ina CRM, you've got some stuff in

(51:02):
an ERP, you've got some stuff inspreadsheets, but then how do
we pull that data together foryou to visualize it in a way
that you can act on the data andrun your business like you
know, we're we're you're you'rethe experts in running your
business, but we can bring someexpertise into helping you think
through new and novel ways tolook at that data and then do

(51:26):
something actionable with it.
A lot of the times theconversations that I'll get
brought into are initially it'saround maybe a problem that
they're trying to have of how dowe streamline this quoting
process?
And then we start to peel backthe layers of the onion of like,
okay, well, yeah, you want tostreamline the quoting process,

(51:46):
but what is it that you'rereally trying to do?
Where where is your data?
How are you protecting thatdata?
And oh, wow, you know, yourorganization has a lot of
different formulas in it, forexample, right?
And that is your IP.
Well, how are you protectingthat IP in your organization?
And some of the time thebusiness leaders haven't really

(52:06):
thought through the biggerpicture of what their
organization's doing andpotentially things, other pieces
that they could be solving.
We've built a number of theseportals, and and I remember one
from a from a person that owneda variety of personal fitness
gyms.

(52:27):
And they were when you own apersonal fitness gym, you're
really focused on the number ofsubscribers that you have and
making sure that the new numberof subscribers outweighs the
number of levers in theorganization, right?
So people leaving theorganization versus people
joining the organization, andyou always want the more people

(52:48):
joining than leaving.
But the amount of data setsaround those numbers are really
mind-boggling because you canget into things like, well, how
many, how many people arewalking through the door?
What time of day are theywalking through the door?
The ad the ads that we have insocial media or on the radio,

(53:09):
the time of day of the ads.
Like we now we can start todirectly measure and influence
the content that we're puttingout, the promotions that we do
or the the in um in gym thingsthat we can do, right?
We can in this gym, we couldpilot a buddy group of we've got
a 6 a.m.
workout group here.

(53:29):
Does that does that lead toretention?
What are some of the thingsthat we can do in the gym to
keep people sticky?
So what are the things earlieron in the cycle, if somebody is
gonna quit in November, whatthings are we starting to see in
September or August that nowall we want to do is keep them

(53:51):
one more month and then one moremonth and then one more month?
So I I really love digging inand sitting down with business
leaders and not just talkingabout like, okay, yeah, you want
to make this particular processbetter, but let's think about
your business as a whole and howwe can leverage technology to
either get you data that is thatyou can then take action on or

(54:15):
present data in a way that ismeaningful to you.

I I think it's interesting in in marketing.
There's the concept of A-Btesting, right?
And what I think is interestingwith AI in so many areas, it
now opens up the ability for aA-B testing in life, basically,
on every role, every process,every customer engagement.
Like we get to A-B test that inways that the AI can actually
help us understand.

(54:39):
Um, which I I think is prettyinteresting around the
opportunities.
Like we we talk about this withclients all the time.
It's like, hey, we're puttingwe're putting out this financial
um analysis that we're using AIto build.
And, you know, should we gothis way or should we go this
way?
I'm like, well, why don't youtry them both and and see what
makes the most sense for how youwant to manage this?
And it's it's like before weare always constrained because

(55:02):
to do a project was so hard, solong to get there.
Where now it's like, in anhour, I might have this figured
out and and and actually have anidea on where where I need to
think next or what, you know, Ihave a solution potentially.
It's like, well, let's try boththough, right?
And we have an opportunity totry both of them in a way that
just hasn't been possiblebefore.
And that applies across theboard.

(55:22):
We're we just built um helped aclient build a routing engine.
They were looking at putting inlike a quarter million dollar
uh routing tool.
So they they have about, Idon't know, 45 trucks or so.
They have uh two differentwarehouses and they have clients
all over, uh customers all overthat they got to deliver, and
their on-time rate was prettylow.
And so they wanted to get thatup.
They're like, if we can get to80% on time, that would be

(55:45):
fantastic.
Well, we ran the process and wegot to 100% on time.
Um and how we did that wasbasically we used A-B testing,
right?
So we said um we use the AI,but we didn't use the AI.
The AI does not run every time,this runs every hour to
reprocess.
Instead of doing that, we saiduse the AI to write a Python
script to actually recalculateconstantly.

(56:06):
So while the first time we didthis, we used hundreds of
dollars in tokens to have the AIcalculate this.
And it basically tried thisalgorithm and this algorithm, it
tried all these differentalgorithms based on all the
rules we added, and then weturns out we had to add a couple
more rules.
So we did a re-ran this processand said optimize it again.
And now it's just a Pythonscript that runs every hour.
It says, here's orders,calculate outer trucking.

(56:28):
And and it's at 100% um on-timedelivery, which is crazy.
That's cool.

Well, I think we're kind of getting towards
the end today, unless you guyshave anything else to do.
I do have a slight pivot hereto maybe maybe current, maybe
future state.

Eric Brown (56:44):
You had in the in the beginning, Lauren talked
about 11 labs, and I'm reallyexcited about the intersection
of voice and AI and some of thethings that, you know, maybe,
maybe you are future thinking.
Um, on the gaming side, um,Nick does a little bit of
gaming.
Uh on the gaming side, like Ithink it's gonna be awesome five

(57:07):
years from now, where theinteractions that we have in the
games are agentic, right?
So every player is gonna have aunique experience in-game,
which will be really cool.
And you know, some of theMMORPGs are gonna be really
dynamic with um that sort ofcontent.
But then I also wonder, right,with some of the tabletop games,

(57:30):
right?
We do a game night here, firstWednesday of every month, but
some of the tabletop games,maybe more of the traditional
tabletops like uh Pathfinder,Nick, you play a little Dungeons
and Dragons too, that's thatsort of stuff, right?
Like where if you have a, youhave all the content, you have
all the rule books loaded in.
Um soon I would think you couldhave an agentic game master

(57:53):
that is doing the storytellingand interacting with voice with
the players.
And I don't know how far offyou think that is, but I think
that's gonna be some reallyexciting stuff.

Maybe we could go around and talk about what
we're most excited about.
That's that's that's some coolstuff.

I want to hear I think I need to rethink my uh
best places to work survey.
And and and here's the otherlast thing here we have in the
chat here.
I think Faye9915 put in bouncecastle Fridays, and yeah, he
nailed it.
We'll get that going.
We'll get that.
He wants to come work here.

Eric Brown (58:30):
Yeah.
He's been on the podcastbefore, I think.
Uh but Lauren, what do youthink about the voice and what
are you doing with voice thesedays?

Yeah, so uh I'm I'm I'm still a little skeptical
on voice in some ways, but Ifind there's there's I I think
the question is voice with AI isnot the same as voice with a
human.
So what that means is we haveto be cautious of where we use
it and where we don't use it.
So when we first started doingtraining as an example, we
started building out umAI-generated videos to do

(59:03):
training.
And those were a big no.
Um, everybody that watched themsaid, I don't ever want to
watch this, and I I feeldisgusting at having watched
what we put up.
We've come a long way sincethen and all of our trainings in
person, basically, right?
But uh I I think it'sinteresting.
I feel a little bit the sameway with voice in the sense that
there are cases, like I talk tomy AI all the time.

(59:26):
I'm in my car a lot, and I'malways on.
If you see me driving down theroad talking, it's more than
likely that I'm talking to myAI.
And I am dialoguing, I'mstrategically thinking about
problems, I am giving it actionsto go code things, to do
things, to verify issues, likere-review things all while I'm
driving.
You might not want to bedriving near me now if if if you

(59:47):
know that, but uh that I'musing the voice all the time,
right?
However, I don't have anexpectation.
Like people ask me, can we usevoice to do outreach sales
calls?
And there's lots, lots ofpeople are telling you this
works great.
And I'll be honest, it's close,but I don't think it's there.
Um, but we we have a guy thatuh is on our team, Darren, and

(01:00:10):
he was on a call recently withuh a guy um for help desk
support and or help support forsomething I can't remember what
it was about.
And uh he was on the call forlike 45 minutes and he's like, I
think this is an AI I'm talkingto.
And so he started asking himquestions, right?
And he asked him question afterquestion, like random, random
questions.
And finally at the end, theguy's like, Why do you ask me

(01:00:32):
all these crazy questions?
And he's like, Well, I thoughtyou were an AI, but I'm not
really sure now.
He's like, Well, I'm definitelynot an AI, but uh that's
interesting.
Nobody's ever accused me ofbeing AI before.
And he's like, after he got offthe call, he's like, I'm not
quite convinced he was an AI,but so it's getting there, but
it's not quite the the speed anddepth um that that we need, but

(01:00:54):
we're getting really close tothat.
AI uh voice is a fantasticmeans of communicating with AI
right now today, though.
The communicating back to ahuman when you don't expect it,
I think one, there's a trustissue.
Like, am I actually talking toa person that actually cares
about what I'm doing?
Or are they just trying to getto a an outcome?
Right.
Um, so there's more you'rewilling to open up to people if

(01:01:14):
it's an actual person.
Um, and so to build that trustis important, but also then is
it high enough quality so peoplewant to interface with it?

Nick Mellem (01:01:24):
Lauren, what do you think the disconnect was with
the AI training?

Loren Horsager (01:01:28):
Do you think people are still wanting that
connection with person-to-persontraining or what they want
someone to see that that theirface looked like they just they
didn't feel right when they sawthis?
And they want someone to noticeand ask them a question about
that and say, hey, did did thatmake sense?
What don't you understand?
And you engage with them.
It is it is incredible what welearn from facial expressions

(01:01:51):
and reactions um in a classroomthat you would never ever pick
up.
We we do one class that'sonline, our co-pilot class,
because it's usually people allover the country.
Um, and it's just it's way,way, way harder to teach.
It's it's less um interestingto people because it's just so
much harder to teach versus inperson.
You get a much betterexperience, you walk away with

(01:02:13):
real results because yourquestions have been heard and
people are attentively asking,hey, did that make sense?
Can I see?
Like, did you do like all ofour classes are hands-on?
You're doing things like, whatdid you do?
Can you can you produce that?
Can you show me?
Um, and where did you getstuck?
And what are you gonna do next?
And like it's just seeing thebrain going, seeing that, hey,
they sat through a break andwe're can you do any work on

(01:02:34):
this?
What are you working on?
What's so excited about?
What are you so excited about,right?
It's understanding those kindof things you never are going to
get from uh an online.

Joshua Schmidt (01:02:42):
I think there's a psychological component there
too.
They it's the man behind thecurtain effect.
They want to know that theycare enough to, you know, put
something out there that's gonnabe real.
There's an authenticity thing.
I mean, it's the same reason.
Like, I I mean, I'm not a fanof um AI music.
Is there a time and a place forit?
You know, probably.
Uh, but you know, I'm not goingout seeking AI music.

(01:03:06):
There's a human element tomusic that uh just contains a
soul, right?
And then and then that's becomemore valuable now with the uh
ushering of AI into that space.
So I think that's where you'regetting some pushback.
And I think there's alwaysgoing to be the mainstream and
then the counterculture, right?
Though if if AI becomesmainstream, there'll just be
people that want to push backjust to push back or just to be

(01:03:27):
cool.
I mean, you know, we still havepeople like myself and Eric
listening to vinyl records.
You know, do we need to dothat?
No, but there's a there's atactile.
Oh, you do too, Nick?
I love vinyl, vinyl records.
It's the same thing likesmoking, smoking a pipe or
whatever, right?
Like you get to hold it, youknow, there's something to
touch.
There's you get to spend timewith it, right?
It's it's not just somethingyou're putting on in the

(01:03:48):
background, like passivelylistening to.
There's an activity there.

I think it's fun that you brought up the uh
vinyls, Josh.
And I actually just this justpopped into my head right now.
I was cleaning out my I had anold bag in my closet that I
found and I was cleaning out uhlast week, and I found an iPod
in there.
And it's so crazy to hear youtalking about vinyls, but then I
but when I looked, found this ithis iPod, I was like, this is

(01:04:14):
like archaic technology.
It's probably worth a bunch ofmoney.
Yeah, exactly.
And it's been all over theworld with me when I was in the
military.
But then to think about thevinyl records, that that seems
like a technology we're comingback to using, but you couldn't
even imagine using an iPod rightnow.
Like that disconnect, thatmiddle ground is gone until
maybe another 50 years we'llcome back, wow, an iPod.

(01:04:35):
Like I can just put some songson here.
It's it's it's kind offascinating to think about that.

I think they're already I think they're already,
you know, I think we're gonnastart stop looking at those
things as regressive or oldtechnologies, but there's a time
and place for them, right?
Like not everything needs to bethe newest.
Like the iPod is a perfectexample of something you want to
give your kids.
Like I'd be more than uhwilling to load up an iPod and
give it to my five-year-oldrather than to hand them the

(01:05:01):
phone, you know, with a YouTubeconnection or a Spotify
connection, right?
Because I know what's on it andand uh and I know you know what
they're listening to, and thenthere's no way they can connect
to the internet or get introuble there.
So I think time and place.

Eric Brown (01:05:13):
I I wrote a uh a newsletter article on LinkedIn
back in June called The NumbersBehind the Curtain.
And it the premise of it waswhere you had said the person
behind the curtain, kind of likethe Wizard of Oz pulling the
strings and whatnot.
But the the premise of thearticle was really around
understanding how the AI islooking at the data that we're

(01:05:35):
giving it.
And it's really just looking atit from a numbers perspective
and vectors, and it reallydoesn't know what we're
thinking.
Um, it has no memory and uh orany of that.
And the the article kind oftalks about like from a society
perspective, we're putting allthis weight and faith in this
tool that we think it is actingin a certain way, but when you

(01:05:58):
start to pull back the Thelayers and you look behind the
curtain, it's really justalgorithms and really
complicated math that is supercool that humans came up in in
an ingenious way of doing this.
But yeah, it was an interestingdeep dive.
I was on it for a couple ofweeks just thinking through how
impactful it is that it'schanging society, and probably

(01:06:22):
only 1% of the population, maybeeven less, really understands
how it works.

Is that different than our human brains?

Um I was just gonna say that, yeah, the neural
network, right?

Right.
So when I got into this in '93,what I first started to do, I I
started reading medical, notnot like but brain books,
basically, right?
I was just fascinated by thefact that these neural networks
were essentially trying toreplicate all the brain work.
So my question was, how doesthe brain work?
And I think it's veryinteresting that uh uh there's
so a ton we don't know aboutthis, but I'm like, I I don't

(01:06:55):
know.
I think it might actually thismight be this a similar process,
right?
Well, one thing I'll add tothat is what you're talking
about here, it's an interestingthing.
People want a human connection.
What I will tell you, liketoday I worked and yesterday I
worked all day with my AI, andit I just I'm happy and it's a
fun thing, right?

(01:07:16):
Um when I have to spend timelike coding, it it's not nearly
the same pleasure for me.
And some people really do findpleasure out of that.
But they did a study, it was umsoftware engineers using AI.
Actually, I think it was about70% happier um once they started
using the AI because they'rehaving a human-like experience.
So it's not a human experience,but it's a human-like

(01:07:37):
experience.
So I think there's like we kindof need this window to evaluate
a lot of things through abouthow human is this, and are we
are we uh violating any of ourhuman experience by providing
this capability?
Um, because I think that goesback to like our original
training videos.
Like it violated all of it,like this is not real, and it

(01:07:58):
was obviously not real, andeverybody rejected it.
So we have to be very cautiousabout that perspective um as we
present a human-like likeoption, how human-like is it?
And are we not violating any ofour human like a human
perspective on that?

Nick, I know you got some thoughts here too, and and
Lauren, I just want to add onto that human piece.
And I'm in the process of uh uhof writing a book on the
differentiator of uh managedservice providers.
Book's gonna be out, I think,second week of uh November.
Um there's uh around authorshipbecause there's so much being

(01:08:35):
written by AI and and and AIgenerated that uh books that
were published in the pre-AI erathat you know are more
guaranteed to be written by ahuman are uh are are are
considered more valuable, Iguess, than than books written
in in this era.
And I saw an article um theother day where they were taking

(01:08:59):
these really old books.
This might have been anotherthing my brother sent me, um,
where they were taking books andcutting the spines off of them
so it was easier to scan um thebooks.
And there was a lot of backlashon that.
I'm thinking to myself, like,okay, yeah, I I understand that
you know people maybe are like,yeah, that's wrong that they're

(01:09:22):
destroying the books.
And then I thought, well, is itthough, from that author's
perspective, who's probably beendead over 50 years, now if
their content it makes it intothe the the general knowledge
base of humanity, is that betterthan the book remaining intact?

I saw that.
I think it's an interestingconversation because you can
kind of see both sides of that.

Well, especially if it's gonna get into more
people's hands, right?
If you have this book with allthis packed with knowledge that
might be relevant to more peopleif it's in AI or you can find
it on the on a website orwhatever it is.
But if it just lives at aBarnes and Noble or at a library
where maybe you wouldn't findthis knowledge, I think that's
the power where I would want mykids to have access to anything

(01:10:13):
and everything that they can.
Right.
We don't want to take it forfree, right?
That whole thing, right?
Play by the rules.
But if we can push theknowledge to more people, I
think that's the power in it.
And we want to do that.

I mean, the most excited, uh, we'll just, you
know, wrap it up here, maybe,um, unless you guys have
anything else you want to chatchat about.
I'm I'm most excited about umhow we've democratized music by
having Apple and Max ineverybody's home, and you can
make a studio into your bedroom.
Um, the democratization ofmovies and film.

(01:10:45):
You know, it takes a big budgetto make a you know kind of
Hollywood blockbuster type film.
And I've been seeing these umshort films on YouTube.
One's called Pomegranate.
It's a sci-fi short film, and acouple of other ones that are
AI generated.
They're basically businesscards for these AI, you know,
generative uh film studios.
But they're getting reallygood.

(01:11:06):
And I'm I'm excited to see whatwill happen to film and video
and movies when when we all canstart making them, or it gets a
little more uh easy because it'sbeen so gate kept for so long,
like Hollywood, and um, youknow, it can get political at
times, and you know,everything's become so PC.
It's like there's not a lot ofgreat stuff anymore, in my

(01:11:29):
opinion.
Like I'm often find myself justscrolling through Netflix or
like we hardly go to the movietheater anymore because there's
not a lot of great stuff outthere.
So I'm really excited to seelike when some really creative
people get a hold of this and itcan kind of come back down to a
cottage industry again and andhopefully revitalize it and and
start making good movies again,you know.

I read an article recently in Esquire magazine um
that that uh they interviewedBrad Pitt, and you know, they
said Brad Pitt's like you knowthe top 1% of the top 1% of
actors that's ever been.
And they asked they asked himabout AI and the impacts of AI,
and and one of the things thathe said was he had always
destroyed the digital likenessesof himself, because in in a in

(01:12:12):
his contract, like you know, hehe writes that they don't get to
keep them.
And then he was looking backand he's like, Man, I wish I
didn't destroy those becausethat was him, you know, all
those years ago, and he couldkind of see himself come up and
you know, I think he's in his60s now.
But um it it it was interestingin that that mindset change.
And he even in the article saidthat it would be interesting

(01:12:33):
that if a director could workwith a digital Brad Pitt, could
it get that emotion in thatscene that he as a human is
bringing to the film at thatparticular point in time when it
and it's that partnership andworking with the director of
like, no, I need a little bitmore of this for this scene or a
little less of that.
And can you give that directionto the AI?

(01:12:56):
And you know, maybe you can'ttoday, maybe you can tomorrow.
But I I thought it was a goodarticle.
It's a good read Esquiremagazine interview with Brad
Pitt.

What were you gonna say, Nick?
Did you have one more thing?

I was gonna say if you're looking for a good movie,
you better get out there nowand go see the Odyssey.
I heard that was good.
Reading the book right now.
We saw it opening weekend, mywife and I.
Incredible.
Probably my favorite movie I'veever seen.
And I think because uh thephotography on the side, but it
was shot in 70 millimeter film,right?
So it's like kind of what Joshis talking about.

(01:13:30):
And it was like no CGI, right?
Like everything was done.
Like if you go see the movie,you'll notice, yeah, practical
effects, thank you.
Like using a large puppet orsomething, right?
Like driving boats in a circleto create a whirlpool, like how
it should be done.
And and I think, and I sawsomebody else post about I think
it's this weekend that Fast andthe Furious is 25 years old and

(01:13:51):
it's going back to the movietheaters, and everybody's
talking about the movie theatersare packed to see it.
Right.
And and that was again justlike the Christopher Nolan
movie, The Odyssey.
It's practical effects.
There's no CGI.
It's how it wasn't they'dfilmed it, how they intended you
to watch it and perceive itright the whole thing.
So that's cool.
That's cool.
So go see the Odyssey if youhaven't.
It's worth the see it in thetheater.

Well, it's fun to leave this on a on a positive
note.
Yeah.
And um once again, thank you somuch, Lauren, for joining us
today.
It's been fun working with you.
We'll have to have you back on.
I know we have some thingscooking behind the scenes.
So the bouncy house,apparently.
The bouncy house is on Friday.
Bouncy Fridays.
Now it's every Friday.
Every Friday.
I'll be there.

(01:14:34):
Thanks for everybody forjoining.
You've been listening to theaudit presented by IT Lab Audit
Labs.
I'm your co-host, producer,Joshua Schmidt.
We've been joined by Eric Brownand Nick Mellum from IT Audit
Labs, and our guest today wasLauren Orsauger from Model Mind
AI.
Please check out Lauren online,his website, and uh you can
find us wherever you stream yourpodcast, Apple Spotify video on

(01:14:54):
YouTube.
Thanks, and we'll see you inthe next one.

You have been listening to the audit presented
by IT Audit Labs.
We are experts at assessingrisk and compliance while
providing administrative andtechnical controls to improve
our clients' data security.
Our threat assessments find thesoft spots before the bad guys
do, identifying likelihood andimpact, or all our security
control assessments rank thelevel of maturity relative to

(01:15:20):
the size of your organization.
Thanks to our devoted listenersand followers, as well as our
producer, Joshua J.
Schmidt, and our audio videoeditor, Cameron Hill.
You can stay up to date on thelatest cybersecurity topics by
giving us a like and a follow onour socials, and subscribing to
this podcast on Apple, Spotify,or wherever you sourced your

(01:15:42):
security content.

All Episodes

Episode Transcript

Popular Podcasts

Stuff You Should Know

Crime Junkie

The Clay Travis and Buck Sexton Show

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}Building the Future: AI Coding, Agentic Advisors and Custom Tools

Episode Transcript

Popular Podcasts

.css-r6mb8g{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:1;overflow:hidden;}Stuff You Should Know

Crime Junkie

The Clay Travis and Buck Sexton Show

All Episodes

Building the Future: AI Coding, Agentic Advisors and Custom Tools

Stuff You Should Know