Transcript
00:00
Play from here
Copy citation
Hey there, Eli is out of town this week, so I've invited Aaron Wilt, who's a very good
00:20
Play from here
Copy citation
friend of mine and a small business owner to join us today on AI at Work.
00:26
Play from here
Copy citation
Aaron, can you introduce yourself a little bit?
00:29
Play from here
Copy citation
Sure, yeah.
00:30
Play from here
Copy citation
Thanks, Kevin.
00:31
Play from here
Copy citation
It's kind of a random thing to jump on and do together, and I love being here, and thanks
00:35
Play from here
Copy citation
for the opportunity.
00:36
Play from here
Copy citation
So a little bit about me, classically trained, I guess you could say, as a programmer way
00:41
Play from here
Copy citation
back in the stone ages of the early 2000s, used to do Java web development for enterprises
00:48
Play from here
Copy citation
all around in the evenings when I was doing consulting, picked up online poker, started
00:53
Play from here
Copy citation
doing better at poker than I was at programs, did that for 10 years professionally kind
00:58
Play from here
Copy citation
of toward the end of that, I kind of saw AI coming, not necessarily AI in the way
01:03
Play from here
Copy citation
that we know it now, but saw it in the poker circles.
01:06
Play from here
Copy citation
I used to be able to beat the bots at poker, and then it was getting to the point where
01:09
Play from here
Copy citation
I no longer could.
01:10
Play from here
Copy citation
So I kind of saw the writing on the wall there, shifted to the next thing, which was e-commerce,
01:15
Play from here
Copy citation
which of course is where we met when we met in China, Canton Fair, and online, some e-commerce
01:23
Play from here
Copy citation
groups then have been doing that now since 2015.
01:27
Play from here
Copy citation
Well, I'm coming up on 11 years, actually 11 years, June 1st.
01:30
Play from here
Copy citation
So Aaron isn't giving himself enough credit.
01:32
Play from here
Copy citation
One of the reasons that I wanted to have Aaron on, in part, it's good to have a gut check
01:38
Play from here
Copy citation
of actual operators and what actual operators are doing on the ground.
01:43
Play from here
Copy citation
And Aaron is really an interesting creator because he has a way of thinking about things
01:49
Play from here
Copy citation
from a systems perspective that goes far and beyond the sophistication of most organizations
01:55
Play from here
Copy citation
that are out there.
01:56
Play from here
Copy citation
So I think it's a pretty good opportunity to calibrate on where a highly skilled SMB
02:03
Play from here
Copy citation
type operator sits as far as the AI evolution is right now, and all of the frustrations
02:09
Play from here
Copy citation
that come with that, because it's certainly not a perfect point in time, but yet the media
02:14
Play from here
Copy citation
is telling us, oh, you know, everybody is just like killing it in all of these different
02:19
Play from here
Copy citation
directions.
02:20
Play from here
Copy citation
And the reality is, is far, far more jagged.
02:23
Play from here
Copy citation
And Aaron's company is actually 40 people strong.
02:27
Play from here
Copy citation
So he's a pretty good representation of larger, small businesses, I would say, kind of even
02:33
Play from here
Copy citation
getting close to mid-market type businesses and where they sit, because I suspect that
02:39
Play from here
Copy citation
if Aaron could snap his fingers, he would have AI built into every corner of his business.
02:45
Play from here
Copy citation
Right, Aaron?
02:46
Play from here
Copy citation
Yeah.
02:47
Play from here
Copy citation
It's a real struggle because you see all this power, right?
02:48
Play from here
Copy citation
You see the power and what it can do, and especially you have some of these.
02:52
Play from here
Copy citation
I think each of us in our own way has kind of had some of these like aha moments where,
02:56
Play from here
Copy citation
you know, Chachi or Claude or whatever came back with something that was like, oh man,
02:59
Play from here
Copy citation
I didn't think about that, or, oh yeah, that's exactly what I wanted.
03:03
Play from here
Copy citation
And it came back like that.
03:04
Play from here
Copy citation
And I think once you kind of get that moment, it's hard to let go of that, and especially
03:09
Play from here
Copy citation
tech guys, you know, kind of like we are.
03:11
Play from here
Copy citation
And so, but getting everyone to have that moment, you know, it's hard.
03:16
Play from here
Copy citation
And I know some folks, like more creative minded folks on my team have been resistant
03:21
Play from here
Copy citation
to that because they'll be like, oh, I asked it to do this, you know, I asked it to put
03:25
Play from here
Copy citation
our supplement bottle on this woman's hand and she had six fingers, like this is garbage,
03:31
Play from here
Copy citation
I'm out, you know?
03:32
Play from here
Copy citation
And it's like, that's kind of where they stop and getting over that hump and getting them
03:36
Play from here
Copy citation
to the point where they can really buy in, I think is one of the challenges.
03:39
Play from here
Copy citation
And that analogy is across all different parts of business, not just creative, of course.
03:43
Play from here
Copy citation
So I like to think about AI in terms of sort of the eras that we're going through, and
03:48
Play from here
Copy citation
I'd love to get your reaction on this.
03:51
Play from here
Copy citation
You know, you dove in right away, like the very beginning of 2023, we were talking about
03:57
Play from here
Copy citation
this, I was in in 2022.
03:59
Play from here
Copy citation
That first bit was this chatbot era, where it was using the chatbot to do kind of interesting
04:07
Play from here
Copy citation
things.
04:08
Play from here
Copy citation
And in retrospect, it wasn't as cool as it felt at the time.
04:12
Play from here
Copy citation
But chatting with this like super intelligence that helps to give you some feedback or to
04:19
Play from here
Copy citation
help you polish an email or maybe look at a little bit of data or maybe make a six-fingered
04:23
Play from here
Copy citation
otter or like whatever it is.
04:26
Play from here
Copy citation
And then about later 2024, 2025, I like to think that we got into the context era, where
04:34
Play from here
Copy citation
it was more about making sure that the LLM, the chatbot had information from different
04:43
Play from here
Copy citation
corners of the business, such that it could understand more about you and about your business
04:49
Play from here
Copy citation
and to be more effective.
04:52
Play from here
Copy citation
And we kind of started calling that context engineering.
04:55
Play from here
Copy citation
And now in 2026, there are two different things that are happening, well, there are a lot
05:00
Play from here
Copy citation
of things that are happening.
05:01
Play from here
Copy citation
But one is that we're entering this era of orchestration, where those individual pieces
05:10
Play from here
Copy citation
of context and the chats are now being drawn together via agentic systems like OpenCloud
05:18
Play from here
Copy citation
or CloudCowork, CloudCode, Codex.
05:22
Play from here
Copy citation
You were talking about Hermes at the beginning of the call as well.
05:25
Play from here
Copy citation
And that's kind of changing things as well.
05:28
Play from here
Copy citation
So maybe just react to that as far as the arc in what you've seen in your own business
05:33
Play from here
Copy citation
and how that's played out for you as a highly technical operator with a team who is less
05:40
Play from here
Copy citation
technical.
05:41
Play from here
Copy citation
Yeah.
05:42
Play from here
Copy citation
I think the interesting part of that whole arc is I think there's kind of like, the way
05:45
Play from here
Copy citation
I think of it is there's kind of one thing that ties all that together.
05:48
Play from here
Copy citation
And I think that it's the, I don't know the right term, maybe you have a term for it,
05:52
Play from here
Copy citation
but it's basically like the percentage chance that the AI is right.
05:56
Play from here
Copy citation
And so at the very beginning of the chat, the percentage chance that AI is right is
06:00
Play from here
Copy citation
maybe, I don't know, let's say 80% back then, like 2023 or something, I was going to say
06:05
Play from here
Copy citation
some stuff and maybe you can keep 80% of it, you got dumped 20% of it.
06:10
Play from here
Copy citation
And then as that gets up to like 85%, we start trusting it more, maybe a little bit, 9%.
06:14
Play from here
Copy citation
And then now all of a sudden you can start abstracting that out, right?
06:16
Play from here
Copy citation
So it's like, okay, now it's getting right often enough.
06:21
Play from here
Copy citation
I just need to make sure we give it the right stuff and then that's going to make it right
06:24
Play from here
Copy citation
more often.
06:25
Play from here
Copy citation
And then I give it the right stuff and then I give it a little bit of like guardrails
06:28
Play from here
Copy citation
to orchestrate it.
06:30
Play from here
Copy citation
And then now all of a sudden it can take that 80, 90% of being right and apply that to another
06:37
Play from here
Copy citation
thing that can be 80, 90%, right?
06:38
Play from here
Copy citation
It can apply to another thing that's 80, 90%, right?
06:41
Play from here
Copy citation
And so then I think that's how it kind of grows because it's all, in a way, probabilistics.
06:47
Play from here
Copy citation
I think that it does allow you to do more and the same problem still exists where the
06:55
Play from here
Copy citation
human needs to understand which part it can use and which part it needs to dump.
07:00
Play from here
Copy citation
And I think where a lot of this ends up going off the rails is when the calibration of how
07:05
Play from here
Copy citation
much you trust it is way off.
07:07
Play from here
Copy citation
So if you trust it too little, then you're not getting the value.
07:09
Play from here
Copy citation
If you trust it too much, you're going to get burned.
07:11
Play from here
Copy citation
And so you got to have that right amount of where you put that dial and that's where it
07:15
Play from here
Copy citation
takes human expertise to get the maximum out of the AI.
07:19
Play from here
Copy citation
Okay.
07:20
Play from here
Copy citation
So there's a bunch there that really resonates with me.
07:23
Play from here
Copy citation
Number one is we do a lot of AI training.
07:27
Play from here
Copy citation
So we go into organizations and we work with them and I'm still almost obligated to do
07:34
Play from here
Copy citation
what I call AI 101.
07:36
Play from here
Copy citation
And AI 101 is doing at least a basic dive into what a neural net is and what a large
07:41
Play from here
Copy citation
language model is because that probabilistic nature of it that you just cited is really
07:47
Play from here
Copy citation
important for people to understand.
07:49
Play from here
Copy citation
They're used to going into a deterministic model like Excel where two plus two equals
07:54
Play from here
Copy citation
four every single time and it's coded that way.
07:57
Play from here
Copy citation
And it gets them into trouble because they don't actually realize that the drift, the
08:02
Play from here
Copy citation
hallucination is actually part of the magic that makes all of this work, but it's both
08:09
Play from here
Copy citation
a feature and a bug.
08:11
Play from here
Copy citation
And if you don't, if you don't enculturate people into that understanding, they get into
08:16
Play from here
Copy citation
trouble in a hurry.
08:17
Play from here
Copy citation
And the other place where they get into trouble in a hurry is by operating outside of their
08:22
Play from here
Copy citation
sphere of knowledge.
08:24
Play from here
Copy citation
So now you and I know a ton about digital marketing and products and like all of these
08:29
Play from here
Copy citation
other things, but we probably don't know a ton about particle physics.
08:33
Play from here
Copy citation
Well, you might. I don't know.
08:36
Play from here
Copy citation
And if I could probably ask an LLM to explain how particle physics works at a 10th grade
08:46
Play from here
Copy citation
level, but I have no way of validating whether or not that information is accurate because
08:51
Play from here
Copy citation
I don't have the fundamental knowledge to gut check it.
08:55
Play from here
Copy citation
And it's another place where people get into trouble.
08:57
Play from here
Copy citation
And you think about a new employee, you wouldn't assign them something really complicated and
09:02
Play from here
Copy citation
just set them loose on it.
09:04
Play from here
Copy citation
You'd want a manager or somebody to be gut checking them along the lines.
09:08
Play from here
Copy citation
Right. Yeah. And I did.
09:11
Play from here
Copy citation
Well, yes. And I would like to hire people who kind of have that natural skepticism in
09:16
Play from here
Copy citation
a way. Right. So again, it's like that dial.
09:19
Play from here
Copy citation
It's like you got to be both of us are kind of skeptical guys.
09:22
Play from here
Copy citation
And I think we use that power for good.
09:24
Play from here
Copy citation
Some people are skeptical, use that power for bad.
09:26
Play from here
Copy citation
But some people are so optimistic and non-skeptical that they just trust what anybody ever says.
09:30
Play from here
Copy citation
And so it's like you can't trust the LLM just like you can't trust the person who you're
09:36
Play from here
Copy citation
talking to, the consultants that you hire blindly.
09:38
Play from here
Copy citation
You have to make sure that you take what works for you and your business and what works
09:42
Play from here
Copy citation
based on your starting point and firm reference and then make the best choice you can.
09:46
Play from here
Copy citation
And I think it's the same way with LLM.
09:48
Play from here
Copy citation
So it's teaching that young person or that new employee, what you know, what questions
09:53
Play from here
Copy citation
did you ask? What do you get back from the LLM?
09:56
Play from here
Copy citation
Does that gut check make sense?
09:59
Play from here
Copy citation
If you don't know, go find the person who knows and run that by them.
10:02
Play from here
Copy citation
Right. It's like, hey, is this is this am I on the right track here?
10:05
Play from here
Copy citation
So I want to talk a little bit about your team and and how you've enculturated AI.
10:10
Play from here
Copy citation
Forty people, a lot of them I know are sort of back office and whatnot.
10:14
Play from here
Copy citation
What does AI usage look like on your team?
10:18
Play from here
Copy citation
What sort of licenses are you using?
10:20
Play from here
Copy citation
What's your tech stack?
10:21
Play from here
Copy citation
What is this? What is a day in May 2026 at Pulse V look like?
10:27
Play from here
Copy citation
Yeah. So I would say that usually I am on more of the bleeding edge of the stuff.
10:33
Play from here
Copy citation
I want to be the one trying things, breaking things, figuring out what what can be done.
10:37
Play from here
Copy citation
I hired a pretty good programmer from Eastern Europe who is also like very AI pilled.
10:42
Play from here
Copy citation
He loves it and really good guy, smart guy.
10:46
Play from here
Copy citation
I'm I'm for him.
10:48
Play from here
Copy citation
I'm trying to dial him back to like, OK, the AI tools are good.
10:52
Play from here
Copy citation
They're like they're the tool.
10:53
Play from here
Copy citation
They're not the answer.
10:54
Play from here
Copy citation
Whereas other people on the team, I'm trying to dial them up being like,
10:58
Play from here
Copy citation
hey, you didn't even ask Claude.
10:59
Play from here
Copy citation
You know, like at least run this by Claude before you bring it to me.
11:01
Play from here
Copy citation
Right. So I think we're kind of all on the spectrum.
11:05
Play from here
Copy citation
Literally and figuratively, we could say a typical day for me would look like I have
11:11
Play from here
Copy citation
right now before we reported this.
11:12
Play from here
Copy citation
I've got five cloud code windows open to Codex CLI windows open.
11:18
Play from here
Copy citation
I am right now myself pulling in a lot of our e-commerce data into Google BigQuery,
11:21
Play from here
Copy citation
setting up those connections to stream and automate that.
11:24
Play from here
Copy citation
We're using some server side tracking tools for our Facebook ads and Google ads and things like that.
11:30
Play from here
Copy citation
So I'm trying to make sure that all that data is getting in the right place.
11:33
Play from here
Copy citation
We can do good reporting.
11:34
Play from here
Copy citation
My goal for that is to then be able to give my marketing team like really good reports
11:39
Play from here
Copy citation
that they can trust because it's hard to trust Facebook portal.
11:42
Play from here
Copy citation
It's hard to trust any Google analytics, etc.
11:45
Play from here
Copy citation
So my dial for AI there is I want to be all in on the tech.
11:49
Play from here
Copy citation
I want to be able to just hand this thing to somebody else.
11:52
Play from here
Copy citation
Let them use it.
11:53
Play from here
Copy citation
AI helped me get there faster.
11:54
Play from here
Copy citation
For others, like maybe the graphics team and the content team, I want them.
11:59
Play from here
Copy citation
We have a couple of gals that are really bought in.
12:02
Play from here
Copy citation
A couple of them, not quite yet.
12:03
Play from here
Copy citation
So I'm trying to have the girls that are really bought in show what is possible.
12:07
Play from here
Copy citation
I'll make some videos about here's what I did.
12:09
Play from here
Copy citation
Here's how I got there.
12:10
Play from here
Copy citation
Again, try to like show them the path that way.
12:13
Play from here
Copy citation
So it's more like content, imagery, things like that.
12:15
Play from here
Copy citation
They're using, they jump around.
12:17
Play from here
Copy citation
They use Gemini some.
12:18
Play from here
Copy citation
We use the Chattanooga TV image generator.
12:21
Play from here
Copy citation
The one that came out recently that was not as much on cloud.
12:24
Play from here
Copy citation
And then last but not least, we got some folks that work in product.
12:27
Play from here
Copy citation
Yes, you'd say like product design, compliance.
12:29
Play from here
Copy citation
We do some things with medical devices.
12:31
Play from here
Copy citation
So we have a lot of compliance work.
12:32
Play from here
Copy citation
And some of those folks are not classically trained in any of this.
12:36
Play from here
Copy citation
So they are very much put this into cloud, put this in Chattanooga TV.
12:40
Play from here
Copy citation
Where does the Venn diagram overlap?
12:42
Play from here
Copy citation
Where are they separated?
12:44
Play from here
Copy citation
Get all the questions I can get.
12:45
Play from here
Copy citation
So that when we go meet with our consultant on Wednesdays,
12:48
Play from here
Copy citation
we have good questions to ask.
12:49
Play from here
Copy citation
They're like double checked.
12:50
Play from here
Copy citation
Is this right?
12:51
Play from here
Copy citation
Is this not right?
12:51
Play from here
Copy citation
And then we're also building a, redoing some of our quality management system
12:55
Play from here
Copy citation
with some of these tools as well.
12:56
Play from here
Copy citation
So we're kind of, that's, I don't know if that answered your question,
12:59
Play from here
Copy citation
but we're all over the place on all that.
13:00
Play from here
Copy citation
So I'm going to push a little bit.
13:01
Play from here
Copy citation
So, so we meant you, you mentioned basically every platform and my,
13:05
Play from here
Copy citation
my listeners will know that I have to use basically every platform too,
13:08
Play from here
Copy citation
because although they are sort of converging,
13:12
Play from here
Copy citation
they have different strengths and weaknesses for different things.
13:15
Play from here
Copy citation
But do you, are you using like an anthropic teams set up and shared projects,
13:22
Play from here
Copy citation
shared chat, shared context, that sort of thing?
13:25
Play from here
Copy citation
Or is it, is it kind of all over the map?
13:27
Play from here
Copy citation
It's kind of piecemeal for us.
13:29
Play from here
Copy citation
You know, I'm not against doing the, you know, the teams kind of aspect.
13:33
Play from here
Copy citation
It's just, you know, like for me and the one programmer I was mentioning,
13:36
Play from here
Copy citation
like we want to be on the latest thing, regardless.
13:38
Play from here
Copy citation
So I just, I got, we both got like a $200 a month plan.
13:42
Play from here
Copy citation
Like there's just ours.
13:42
Play from here
Copy citation
I got one for him.
13:43
Play from here
Copy citation
That's just his.
13:44
Play from here
Copy citation
And we go fast.
13:45
Play from here
Copy citation
The others, we let them lag behind a little bit because most of what,
13:49
Play from here
Copy citation
what they're doing it doesn't need to be on the bleeding edge necessarily.
13:54
Play from here
Copy citation
It just needs to get them down the path faster.
13:56
Play from here
Copy citation
So that's kind of where, where we've drawn that line.
13:58
Play from here
Copy citation
Okay.
13:58
Play from here
Copy citation
Have you baked it into performance?
14:01
Play from here
Copy citation
Anything like that?
14:02
Play from here
Copy citation
As far as organizational, like adoption and putting carrots and sticks alongside of it?
14:09
Play from here
Copy citation
Not as much as I wish we would have or could have.
14:12
Play from here
Copy citation
I think I've been, I've been, you know, Kevin kind of knows me as like a
14:15
Play from here
Copy citation
doer more so than like a manager.
14:18
Play from here
Copy citation
And I think the AI kind of feeds into that trap a little bit.
14:21
Play from here
Copy citation
Sometimes I have to pull myself back out and be like, okay, let's, let's make sure
14:25
Play from here
Copy citation
we're looking across the Oregon and doing things that will drive the whole thing forward.
14:28
Play from here
Copy citation
But for us, we're not, I haven't done a whole lot with
14:31
Play from here
Copy citation
using it to help manage other than my side project is trying to get,
14:35
Play from here
Copy citation
I had Opal Club for a while, trying to get Hermes going to ingest a lot of our.
14:39
Play from here
Copy citation
like Asana data, Slack data, and then have me ask it questions
14:44
Play from here
Copy citation
about what different folks are working on at a given time.
14:49
Play from here
Copy citation
That hasn't, it's not good enough for me yet
14:52
Play from here
Copy citation
to like let it loose.
14:53
Play from here
Copy citation
So that's still kind of a passion project, I guess.
14:56
Play from here
Copy citation
So I like to say that AI can generally
14:58
Play from here
Copy citation
do three different things.
15:00
Play from here
Copy citation
It can replace tasks, it can augment tasks,
15:03
Play from here
Copy citation
or it can create new capabilities.
15:06
Play from here
Copy citation
When you and your dev are leaning into these projects,
15:09
Play from here
Copy citation
which categories do you think you're focused on the most
15:12
Play from here
Copy citation
as far as those three?
15:13
Play from here
Copy citation
Okay, so run it by me again.
15:15
Play from here
Copy citation
So it can replace tasks,
15:16
Play from here
Copy citation
so that's like workflow replacement.
15:18
Play from here
Copy citation
It can augment, as in now that individual
15:22
Play from here
Copy citation
can do a heck of a lot more than they did previously,
15:25
Play from here
Copy citation
or it can create new organizational capabilities.
15:28
Play from here
Copy citation
So what you were saying earlier,
15:29
Play from here
Copy citation
as far as the rabbit hole that you're going down right now,
15:32
Play from here
Copy citation
as far as day doing, you know, marketing data absorption
15:37
Play from here
Copy citation
and things like that,
15:38
Play from here
Copy citation
you would have had to pay for a SaaS
15:40
Play from here
Copy citation
for that a few years ago.
15:41
Play from here
Copy citation
So now you're developing a new capability
15:44
Play from here
Copy citation
that you didn't have before.
15:46
Play from here
Copy citation
And in that case, that's obviously internal operations,
15:49
Play from here
Copy citation
but does that also hit a point where, you know,
15:53
Play from here
Copy citation
you're a direct to consumer business,
15:55
Play from here
Copy citation
so reinventing what it means
15:57
Play from here
Copy citation
to be a direct consumer business
15:58
Play from here
Copy citation
might be a little bit challenging.
16:00
Play from here
Copy citation
But of those three,
16:02
Play from here
Copy citation
where do you think the most effort is being spent?
16:05
Play from here
Copy citation
Yeah, for us, definitely augmenting.
16:07
Play from here
Copy citation
I mean, I think that's the real win, honestly.
16:09
Play from here
Copy citation
Like, I think that's, I don't, for me at this 2026,
16:13
Play from here
Copy citation
I'm not in the camp of like AI is replacing all of our jobs,
16:16
Play from here
Copy citation
at least not today, okay?
16:18
Play from here
Copy citation
Today it is AI is taking somebody who's a winner
16:21
Play from here
Copy citation
and making them win harder, right?
16:23
Play from here
Copy citation
And it's taking the loser and probably gonna lose harder
16:25
Play from here
Copy citation
because they're just gonna copy and paste and be done, right?
16:28
Play from here
Copy citation
And then they're gonna get in trouble
16:29
Play from here
Copy citation
and it's gonna be all fine until it's not.
16:31
Play from here
Copy citation
So for me, getting the AI in the hands of your winners
16:34
Play from here
Copy citation
and letting the winners run really fast,
16:37
Play from here
Copy citation
that is the way.
16:38
Play from here
Copy citation
And then to your point about new capabilities,
16:41
Play from here
Copy citation
yes, I would keep that for me
16:43
Play from here
Copy citation
and some of the highest,
16:45
Play from here
Copy citation
like my programmer I mentioned, the one particular guy.
16:48
Play from here
Copy citation
We have, I have a couple of programs on staff
16:49
Play from here
Copy citation
and I'm letting them do some of this too.
16:51
Play from here
Copy citation
But primarily me and this one particular guy
16:53
Play from here
Copy citation
are really trying to push the envelope on capability
16:56
Play from here
Copy citation
because more because of a control thing,
16:58
Play from here
Copy citation
like I wanna make sure I see it
16:59
Play from here
Copy citation
and I don't trust it enough to just blindly let it go.
17:02
Play from here
Copy citation
So your organization is small enough
17:04
Play from here
Copy citation
that you can have that sort of central control
17:06
Play from here
Copy citation
in that you see, okay, there's an opportunity over here.
17:10
Play from here
Copy citation
I can lean in with one of my people
17:12
Play from here
Copy citation
that have competency or yourself to lean in,
17:15
Play from here
Copy citation
solve the problem and move on.
17:17
Play from here
Copy citation
But in larger organizations, I think it gets pretty hard.
17:20
Play from here
Copy citation
If you were four times your size
17:23
Play from here
Copy citation
and you had 40 people in four different verticals,
17:26
Play from here
Copy citation
how do you think you might think
17:28
Play from here
Copy citation
about how do you amplify like a sales vertical
17:31
Play from here
Copy citation
or something like that?
17:34
Play from here
Copy citation
What's the right approach organizationally?
17:36
Play from here
Copy citation
Do you teach everybody on that team to fish
17:39
Play from here
Copy citation
such that they're like running around doing things
17:42
Play from here
Copy citation
or do you teach them to at least identify where the fish are
17:46
Play from here
Copy citation
and then bring the fishermen team in to do it?
17:48
Play from here
Copy citation
Am I killing my metaphor there?
17:51
Play from here
Copy citation
But, you know, it's that idea of,
17:54
Play from here
Copy citation
do you build that capacity and allow people to run
17:57
Play from here
Copy citation
or does that actually work?
18:00
Play from here
Copy citation
I have opinions about it,
18:02
Play from here
Copy citation
but I'm curious about what your opinion is.
18:03
Play from here
Copy citation
I think right now, well, I think in the future,
18:09
Play from here
Copy citation
I think letting people run
18:10
Play from here
Copy citation
and like really going from kind of ground up
18:13
Play from here
Copy citation
and letting the experts like really flourish
18:15
Play from here
Copy citation
with the AI tool in their pocket,
18:17
Play from here
Copy citation
I think in the long run, probably that's where it will be
18:20
Play from here
Copy citation
and that will be okay.
18:21
Play from here
Copy citation
I think right now it's not quite there yet for me.
18:25
Play from here
Copy citation
If my organization was four times bigger,
18:28
Play from here
Copy citation
what I would want to do is I would want to,
18:30
Play from here
Copy citation
and I'm still trying to do this even at my size,
18:31
Play from here
Copy citation
but I'm trying to like get people involved,
18:34
Play from here
Copy citation
see what they do.
18:35
Play from here
Copy citation
Some of them are going to take to it right away.
18:37
Play from here
Copy citation
Some of them are going to go kicking and screaming.
18:39
Play from here
Copy citation
Find the people that take to it right away.
18:41
Play from here
Copy citation
Make sure that, I generally find
18:43
Play from here
Copy citation
that there's two different kinds of people, right?
18:46
Play from here
Copy citation
More of the like skeptical, technical,
18:48
Play from here
Copy citation
like I want this to multiply me kind of people
18:50
Play from here
Copy citation
and then the dreamer kind of people.
18:52
Play from here
Copy citation
They're like, oh my God, the world is changing.
18:54
Play from here
Copy citation
This is going to be so amazing.
18:55
Play from here
Copy citation
I think you kind of, right now,
18:56
Play from here
Copy citation
I think you kind of need both
18:58
Play from here
Copy citation
because you need the guy, you need the dreamer to like push
19:01
Play from here
Copy citation
and see what can we do.
19:02
Play from here
Copy citation
And then you need the like more skeptical,
19:04
Play from here
Copy citation
like logical kind of person to be like,
19:06
Play from here
Copy citation
okay, here's the best we can do right now.
19:08
Play from here
Copy citation
Here's the part where it's like,
19:10
Play from here
Copy citation
we still need you to make sure
19:11
Play from here
Copy citation
this thing doesn't go off the rails.
19:12
Play from here
Copy citation
And I think if you can build your organization
19:14
Play from here
Copy citation
in a way where you get those two personalities
19:16
Play from here
Copy citation
really working together, I think the magic can happen.
19:19
Play from here
Copy citation
But I think if you constantly let the like skeptical guy,
19:22
Play from here
Copy citation
like always be the naysayer
19:23
Play from here
Copy citation
and you constantly let the dreamer constantly dream,
19:26
Play from here
Copy citation
I think either way you're going to not like maximizing
19:28
Play from here
Copy citation
and you're going to get into trouble.
19:29
Play from here
Copy citation
So let me give you a specific hypothetical.
19:31
Play from here
Copy citation
You do high volume on Amazon
19:34
Play from here
Copy citation
and Amazon is super fun to deal with.
19:37
Play from here
Copy citation
And there are a lot of, you know,
19:39
Play from here
Copy citation
different types of file types that float around
19:41
Play from here
Copy citation
for different products and things like that,
19:42
Play from here
Copy citation
highly administrative.
19:44
Play from here
Copy citation
And you have somebody on your team who's like in that,
19:47
Play from here
Copy citation
not to say, I'm not going to call them dreamer,
19:48
Play from here
Copy citation
but they're excited about the technology.
19:50
Play from here
Copy citation
They're excited about amplifying themselves.
19:52
Play from here
Copy citation
Are you just going to let them reinvent
19:55
Play from here
Copy citation
some core process in the business?
19:58
Play from here
Copy citation
Or what are the checks that are in place
20:01
Play from here
Copy citation
so that they know kind of the limits of the things
20:05
Play from here
Copy citation
that they can mess with and the things that they don't?
20:07
Play from here
Copy citation
Because I'll take an enthusiastic,
20:10
Play from here
Copy citation
like tech adopter every day,
20:12
Play from here
Copy citation
but boy, you could really break stuff in a hurry
20:15
Play from here
Copy citation
if you're dabbling in places that formerly,
20:18
Play from here
Copy citation
you know, just rank and file people
20:20
Play from here
Copy citation
would never touch these data structures.
20:22
Play from here
Copy citation
And now we're like, hey, cool, go and reinvent your job.
20:26
Play from here
Copy citation
What could possibly go wrong, right?
20:28
Play from here
Copy citation
I mean, in your business, it could be millions of dollars.
20:29
Play from here
Copy citation
So how do you think about controls
20:33
Play from here
Copy citation
as far as access and project design and et cetera,
20:37
Play from here
Copy citation
in that respect?
20:38
Play from here
Copy citation
Yeah, good question.
20:39
Play from here
Copy citation
This came up actually over the last few weeks.
20:42
Play from here
Copy citation
So we've got a really smart guy on our Amazon team
20:45
Play from here
Copy citation
that runs the advertising and he uses some different tools.
20:49
Play from here
Copy citation
And some of these tools actually have their own MCP.
20:51
Play from here
Copy citation
So he's like, yeah, I threw this thing in this MCP
20:55
Play from here
Copy citation
and it told me this and here's what I got.
20:57
Play from here
Copy citation
And it's like in his Asana report, you know?
20:59
Play from here
Copy citation
And I'm like, well, did you verify it?
21:02
Play from here
Copy citation
How do you know it's right?
21:02
Play from here
Copy citation
It's like the simplest question, right?
21:04
Play from here
Copy citation
Like, how do you know what's right?
21:05
Play from here
Copy citation
So from the leadership point of view,
21:07
Play from here
Copy citation
and if you're trying to build this team,
21:08
Play from here
Copy citation
I think you gotta like ask that question.
21:10
Play from here
Copy citation
I mean, it sounds like so stupidly simple,
21:11
Play from here
Copy citation
but it's like, do you know what's right?
21:13
Play from here
Copy citation
And then you get that blank stare back, right?
21:16
Play from here
Copy citation
And it's not because the guy is a bad guy.
21:17
Play from here
Copy citation
It's not because he's trying to be,
21:19
Play from here
Copy citation
but he was excited about the thing
21:20
Play from here
Copy citation
and wanted to show me what he found.
21:21
Play from here
Copy citation
And it's like, well, that might be right,
21:23
Play from here
Copy citation
but it might not be right.
21:24
Play from here
Copy citation
And which part is not right?
21:25
Play from here
Copy citation
And so when you get into these things of like,
21:27
Play from here
Copy citation
I'm just gonna dump this like 50,000 row Excel file
21:31
Play from here
Copy citation
into like somebody else's MCP server
21:33
Play from here
Copy citation
that's gonna hopefully like give me some gold back.
21:36
Play from here
Copy citation
I mean, it's just like, okay.
21:38
Play from here
Copy citation
All right, let's check ourselves here for a second.
21:39
Play from here
Copy citation
So, but I don't wanna discourage the guy.
21:41
Play from here
Copy citation
I just like, okay, how do we find ways to validate it?
21:44
Play from here
Copy citation
And I think that's really, you know,
21:46
Play from here
Copy citation
kind of back to your earlier like three points of value.
21:50
Play from here
Copy citation
And I think one of the ways it's going
21:54
Play from here
Copy citation
and one of the ways I'm sure you see it going too
21:57
Play from here
Copy citation
is like finding the parts that you can use AI
22:00
Play from here
Copy citation
to make it deterministic faster.
22:02
Play from here
Copy citation
So it's like this whole big process, right?
22:04
Play from here
Copy citation
Which parts can we make deterministic?
22:06
Play from here
Copy citation
Okay, maybe it's only this 30, 40%.
22:09
Play from here
Copy citation
Can we use AI to make this 30, 40% deterministic
22:12
Play from here
Copy citation
every time right away?
22:13
Play from here
Copy citation
So that part doesn't change.
22:14
Play from here
Copy citation
We save our brain power for the judgment part that we need.
22:17
Play from here
Copy citation
And I think that's, so that's kind of how I think about it
22:20
Play from here
Copy citation
and what we've run into.
22:21
Play from here
Copy citation
But yeah, like I told the guys, like you got my credit card.
22:23
Play from here
Copy citation
So let's make sure that we do it right.
22:26
Play from here
Copy citation
So I wanna roll back to that point for just a second.
22:28
Play from here
Copy citation
We talked about deterministic versus probabilistic earlier.
22:31
Play from here
Copy citation
And that is really, really important
22:33
Play from here
Copy citation
when you're dealing with a really big process
22:36
Play from here
Copy citation
is how can you break it down into parts
22:38
Play from here
Copy citation
that are basically digitally mechanical as it were.
22:42
Play from here
Copy citation
And it's sort of the way that coding
22:44
Play from here
Copy citation
has worked up until recently
22:46
Play from here
Copy citation
and reserve the variability of the AI
22:50
Play from here
Copy citation
for the part that you need to reserve for the AI.
22:52
Play from here
Copy citation
So I like to call it like the spaghetti bowl approach
22:56
Play from here
Copy citation
to data.
22:57
Play from here
Copy citation
If you've got this giant bowl of like advertising data
23:00
Play from here
Copy citation
and yeah, you can have a large language model
23:03
Play from here
Copy citation
go look at that data and parse the data
23:06
Play from here
Copy citation
and make conclusions on it.
23:08
Play from here
Copy citation
It will probably do a pretty good job of that.
23:11
Play from here
Copy citation
And that's fine for developing some quick insights.
23:14
Play from here
Copy citation
But if you're going to do like a regimented
23:18
Play from here
Copy citation
data based approach to marketing analytics
23:23
Play from here
Copy citation
or something like that,
23:24
Play from here
Copy citation
you want to be doing more traditional approaches
23:28
Play from here
Copy citation
to data parsing to cut that data down
23:31
Play from here
Copy citation
such instead of being a spaghetti bowl,
23:33
Play from here
Copy citation
it's like one little meatball.
23:34
Play from here
Copy citation
And you're gonna focus on that one little meatball
23:37
Play from here
Copy citation
for the LLM level.
23:38
Play from here
Copy citation
And then you need to decide how risky is it to my business
23:42
Play from here
Copy citation
if this is way off.
23:44
Play from here
Copy citation
Marketing exercise,
23:46
Play from here
Copy citation
probably not the biggest deal in the world.
23:49
Play from here
Copy citation
FDA compliance, really big deal.
23:51
Play from here
Copy citation
And it has to be right every single time.
23:54
Play from here
Copy citation
And teaching people how to make that judgment
23:57
Play from here
Copy citation
when media in the world is telling them that,
24:00
Play from here
Copy citation
hey, it's just this magic machine
24:02
Play from here
Copy citation
that you can get any information from
24:04
Play from here
Copy citation
is it's a learning process for sure.
24:08
Play from here
Copy citation
Yeah, I was talking to an old employee
24:11
Play from here
Copy citation
about this the other day,
24:12
Play from here
Copy citation
sent an email to him and was like,
24:14
Play from here
Copy citation
some of the ways you can think about it is like,
24:16
Play from here
Copy citation
am I okay with 85% right?
24:18
Play from here
Copy citation
If it's 85% right, am I okay with that?
24:20
Play from here
Copy citation
And if your job's on the line
24:22
Play from here
Copy citation
and you're like trying to report to somebody,
24:25
Play from here
Copy citation
if you're like a, whatever, VP of whatever
24:27
Play from here
Copy citation
and you're trying to report to this or that
24:29
Play from here
Copy citation
and like, is 85% good enough for your job?
24:31
Play from here
Copy citation
And maybe certain types of jobs,
24:34
Play from here
Copy citation
maybe it could be, right?
24:36
Play from here
Copy citation
Like, so where's the fail safe, right?
24:37
Play from here
Copy citation
Like customer service.
24:39
Play from here
Copy citation
Okay, on any given email is 85% good enough.
24:42
Play from here
Copy citation
It's like, maybe on any given email it is
24:44
Play from here
Copy citation
because if you go off the rails
24:46
Play from here
Copy citation
and somebody complains, you just refund their money
24:48
Play from here
Copy citation
and it's like, okay, we're all happy now anyway.
24:50
Play from here
Copy citation
No harm, no foul.
24:51
Play from here
Copy citation
So maybe 85% is okay.
24:52
Play from here
Copy citation
But to your point, like FDA compliance
24:55
Play from here
Copy citation
or something related to analytics
24:57
Play from here
Copy citation
that other people are dependent on
25:00
Play from here
Copy citation
and that 85% that other people downstream
25:03
Play from here
Copy citation
are depending, depending, depending, depending,
25:05
Play from here
Copy citation
maybe 85% is not good enough.
25:07
Play from here
Copy citation
And that's, we actually,
25:09
Play from here
Copy citation
so we're redoing our inventory system
25:11
Play from here
Copy citation
with this program we're in.
25:12
Play from here
Copy citation
That's an example of that, right?
25:14
Play from here
Copy citation
That's hundreds of thousands of dollars,
25:16
Play from here
Copy citation
millions of dollars across the year
25:18
Play from here
Copy citation
of decisions being made based on this thing.
25:20
Play from here
Copy citation
So we're not leaving that up to the LLM.
25:22
Play from here
Copy citation
We're trying to make sure that each part
25:24
Play from here
Copy citation
of the puzzle is deterministic in its right way.
25:27
Play from here
Copy citation
Okay, so you just sort of teed off
25:29
Play from here
Copy citation
what's a sort of pretty hot button topic,
25:32
Play from here
Copy citation
which is the idea of the SaaSpocalypse
25:34
Play from here
Copy citation
as people like yourself are looking at your processes.
25:38
Play from here
Copy citation
And it's not like you dislike some of the SaaS partners
25:42
Play from here
Copy citation
and vendors that you have in place.
25:44
Play from here
Copy citation
You just recognize that you can do the thing
25:46
Play from here
Copy citation
that you've always wanted to do now.
25:48
Play from here
Copy citation
And what do you think that that means as far as,
25:52
Play from here
Copy citation
if you were to prognosticate on the future
25:54
Play from here
Copy citation
of sort of mid markets like SaaS,
25:58
Play from here
Copy citation
are people like you going to nibble out
26:01
Play from here
Copy citation
the bottom of the market
26:03
Play from here
Copy citation
or will SaaS find a way to reinvent itself
26:08
Play from here
Copy citation
such that it's headless or whatever the approach is
26:12
Play from here
Copy citation
that allows it to persist and continue to add value?
26:14
Play from here
Copy citation
Yeah, I don't want to give a non-answer,
26:17
Play from here
Copy citation
but it is kind of all of it.
26:18
Play from here
Copy citation
So let me say why I think that.
26:20
Play from here
Copy citation
Number one, like creating software is really hard.
26:22
Play from here
Copy citation
It doesn't matter if Cloud Code is grinding five agents
26:25
Play from here
Copy citation
at a time to do it, it's really hard.
26:27
Play from here
Copy citation
Because, well, I heard this guy the other day say,
26:31
Play from here
Copy citation
maybe you know who it says,
26:32
Play from here
Copy citation
who's the famous person who said this,
26:33
Play from here
Copy citation
but it's like, you can outsource the thinking,
26:36
Play from here
Copy citation
but you cannot outsource the understanding.
26:38
Play from here
Copy citation
And so it's like, yeah, you can have the agents
26:41
Play from here
Copy citation
all think about this stuff,
26:42
Play from here
Copy citation
but if you don't really understand a truly deep level
26:45
Play from here
Copy citation
of how exactly this thing is supposed to work,
26:48
Play from here
Copy citation
you're going to struggle making software.
26:50
Play from here
Copy citation
So it's software of any kind of scale, right?
26:53
Play from here
Copy citation
So I don't necessarily think that, you know,
26:56
Play from here
Copy citation
SaaS apocalypse is going to happen.
26:57
Play from here
Copy citation
I do think there's going to be like a margin compression
26:59
Play from here
Copy citation
for sure, you know, a lot of this is going to come down
27:02
Play from here
Copy citation
and people are going to want to pay
27:03
Play from here
Copy citation
for the connector to the data,
27:04
Play from here
Copy citation
like you're saying the headless.
27:05
Play from here
Copy citation
It's like, yeah, I'll pay so-and-so
27:07
Play from here
Copy citation
for the connector to the data
27:08
Play from here
Copy citation
and I'll do what I want to do myself
27:10
Play from here
Copy citation
because getting that data sometimes
27:12
Play from here
Copy citation
and organizing that data in a good way,
27:14
Play from here
Copy citation
it can be very difficult.
27:15
Play from here
Copy citation
Like, I'm just saying,
27:16
Play from here
Copy citation
here's a very basic example, time zones, right?
27:19
Play from here
Copy citation
You got Shopify coming out in Pacific time,
27:21
Play from here
Copy citation
you got Amazon, UTC, Google a lot of times use UTC,
27:25
Play from here
Copy citation
we live in central time.
27:26
Play from here
Copy citation
So anybody who I want looking at a report,
27:28
Play from here
Copy citation
it better be in central time, right?
27:29
Play from here
Copy citation
So you got all these different systems,
27:30
Play from here
Copy citation
all these different time zones,
27:31
Play from here
Copy citation
you better make sure that every database query
27:33
Play from here
Copy citation
is organizing them the right way.
27:34
Play from here
Copy citation
You better make sure,
27:35
Play from here
Copy citation
and that's just a very basic thing, right?
27:37
Play from here
Copy citation
Just so my marketing people can look at
27:39
Play from here
Copy citation
and trust what they're looking at.
27:41
Play from here
Copy citation
That expands out.
27:43
Play from here
Copy citation
Yeah, I think, yes, I think there's compression in SaaS.
27:46
Play from here
Copy citation
Yes, I think some of the, like,
27:48
Play from here
Copy citation
maybe the Zapiers of the world a little bit,
27:50
Play from here
Copy citation
like get hurt in the beginning
27:51
Play from here
Copy citation
because we can kind of make some of those automations
27:53
Play from here
Copy citation
on our own.
27:55
Play from here
Copy citation
But I think the true understanding of the business,
27:58
Play from here
Copy citation
the true understanding of what it should be,
28:01
Play from here
Copy citation
that's still highly valuable.
28:02
Play from here
Copy citation
No organization can rip out all of their SaaS
28:05
Play from here
Copy citation
and know exactly how it should be for all these tools.
28:08
Play from here
Copy citation
So I've been peppering you with questions,
28:11
Play from here
Copy citation
but I think it could be helpful
28:12
Play from here
Copy citation
for the other business owners who listen to this show
28:15
Play from here
Copy citation
to sort of hear the questions that you have.
28:16
Play from here
Copy citation
Again, you're pretty far forward in the continuum
28:21
Play from here
Copy citation
of people who are leaning into this
28:23
Play from here
Copy citation
and quote, doing it right,
28:25
Play from here
Copy citation
yet it's still super jagged.
28:26
Play from here
Copy citation
So I see dozens and dozens of companies.
28:30
Play from here
Copy citation
What sort of questions do you have for me
28:33
Play from here
Copy citation
that might be helpful?
28:34
Play from here
Copy citation
Yeah, I think a lot of my questions come funny about AI
28:37
Play from here
Copy citation
as it comes back to the people, right?
28:38
Play from here
Copy citation
So it's like, how do we get the people,
28:40
Play from here
Copy citation
how do we get the people to have those magic moments?
28:42
Play from here
Copy citation
And what have you seen in other organizations
28:44
Play from here
Copy citation
like when somebody goes from,
28:47
Play from here
Copy citation
have you ever seen somebody go from like very anti-AI
28:49
Play from here
Copy citation
to now being like an AI zealot?
28:51
Play from here
Copy citation
Where it's like,
28:52
Play from here
Copy citation
because sometimes I think those are the people
28:53
Play from here
Copy citation
that you really want.
28:54
Play from here
Copy citation
It's like the AI crusades, right?
28:57
Play from here
Copy citation
It's like, if you can get the programmer
28:59
Play from here
Copy citation
who's been doing it since you were in diapers
29:00
Play from here
Copy citation
that knows everything about this stuff, right?
29:02
Play from here
Copy citation
If you can get that guy to really buy in on AI,
29:05
Play from here
Copy citation
I feel like that is like magic.
29:06
Play from here
Copy citation
But have you seen that before?
29:08
Play from here
Copy citation
How has that gone?
29:09
Play from here
Copy citation
Sarah, I have a keynote slide
29:10
Play from here
Copy citation
that I call the Pyramid of Suck.
29:12
Play from here
Copy citation
And it is like at the very top of the pyramid
29:15
Play from here
Copy citation
are our business changing ideas,
29:17
Play from here
Copy citation
like reinventing the entire way that you do.
29:19
Play from here
Copy citation
system and saving millions of dollars a year, okay?
29:22
Play from here
Copy citation
And then at the bottom of the pyramid are simply things that people don't like to do.
29:26
Play from here
Copy citation
And from a human deployment perspective, we tend to hire people based on something
29:32
Play from here
Copy citation
that we see that you could say is there like individual genius, be it in marketing
29:37
Play from here
Copy citation
or in copy or an image or data or whatever it is.
29:41
Play from here
Copy citation
And then we load them down from an administrative perspective with a bunch
29:47
Play from here
Copy citation
of craptastic work that they really hate.
29:49
Play from here
Copy citation
You don't have like a lot of, well, you actually, you do have some B2B type
29:53
Play from here
Copy citation
salespeople around, a good salesperson wants to be selling all the time.
29:57
Play from here
Copy citation
They don't want to be updating their CRM.
29:58
Play from here
Copy citation
So one way to build champions is to solve that person's pain point and to
30:04
Play from here
Copy citation
teach them how to solve more of them.
30:06
Play from here
Copy citation
And once they, they break that seal, particularly with the generation
30:11
Play from here
Copy citation
of, of the agentic harnesses right now.
30:15
Play from here
Copy citation
It's like a golden age at this second that I think is only going to get
30:18
Play from here
Copy citation
better that all of a sudden cross-platform communication is like,
30:24
Play from here
Copy citation
it's, it's pretty wide open.
30:26
Play from here
Copy citation
So somebody can be bridging between Asana and HubSpot and Clod and
30:33
Play from here
Copy citation
be solving problems all at once.
30:35
Play from here
Copy citation
And that's where I'm seeing like that, that flash of light of, wow, I never
30:39
Play from here
Copy citation
have to do this TPS report ever again.
30:41
Play from here
Copy citation
Like that is fantastic.
30:44
Play from here
Copy citation
Have you seen in a particular, I don't know, like business vertical or
30:48
Play from here
Copy citation
particular, like department of a business where people tend to
30:51
Play from here
Copy citation
have, have really liked that more?
30:53
Play from here
Copy citation
Or, and I guess my question is, have you seen it work better where you
30:57
Play from here
Copy citation
have some of the tech guys, the programmer guys doing that deterministic
31:00
Play from here
Copy citation
automation for them and doing that faster and doing more of them?
31:03
Play from here
Copy citation
Is that, is that the part that's making like the HR department happier?
31:06
Play from here
Copy citation
Or is it more like the HR department now can, you know, ask the LLM because
31:11
Play from here
Copy citation
it has all this data that they don't have to go bother so-and-so they
31:14
Play from here
Copy citation
could just ask the LLM, like, is that the magic moment or is it both?
31:17
Play from here
Copy citation
Or how do you see that?
31:17
Play from here
Copy citation
I think it's a really important question right now.
31:20
Play from here
Copy citation
We in our, in our training capacity, we'll bring people up to a level of
31:24
Play from here
Copy citation
capability that they can start solving a lot of their own problems, but it
31:28
Play from here
Copy citation
runs right into that wall that we were talking about earlier, as far as risk,
31:33
Play from here
Copy citation
data access, platform access, all of these things.
31:35
Play from here
Copy citation
So there has to be a pressure release valve that allows the person to solve
31:42
Play from here
Copy citation
their problem because otherwise you're just going to take the wind right out of
31:45
Play from here
Copy citation
their sails, like we sit down, we teach them how to think about this,
31:49
Play from here
Copy citation
how to develop new solutions.
31:51
Play from here
Copy citation
They've come up with some brilliant idea and you know,
31:55
Play from here
Copy citation
it's brilliant on a continuum.
31:56
Play from here
Copy citation
It's going to be, have a positive ROI.
31:59
Play from here
Copy citation
They can absolutely see that.
32:01
Play from here
Copy citation
But the tech team of the company, particularly in larger companies, is so
32:05
Play from here
Copy citation
focused on improving their product and all of their rocket science and all of
32:09
Play from here
Copy citation
their dollars are being focused on, on the AI application at that top of the
32:16
Play from here
Copy citation
pyramid, that there aren't a lot of resources farther down to make these
32:20
Play from here
Copy citation
things happen.
32:21
Play from here
Copy citation
And you see both OpenAI and Anthropic leaning into this with their forward
32:26
Play from here
Copy citation
deployed engineers right now.
32:28
Play from here
Copy citation
And this is essentially what the big consulting firms are doing for
32:31
Play from here
Copy citation
enterprise, but OpenAI and Anthropic are now investing literally billions of
32:35
Play from here
Copy citation
dollars in these forward deployed engineers.
32:38
Play from here
Copy citation
And the idea is you can parachute them in to say a marketing team and they'll be
32:43
Play from here
Copy citation
there on a full-time basis and they'll be listening to the needs of the team and
32:49
Play from here
Copy citation
they'll have certain rights and privileges that other people in the organization
32:53
Play from here
Copy citation
don't.
32:53
Play from here
Copy citation
Okay.
32:54
Play from here
Copy citation
That's great.
32:55
Play from here
Copy citation
But neither you nor I is going to be hiring OpenAI forward deployed engineers,
33:00
Play from here
Copy citation
much less like Accenture or KPMG or the rest of them.
33:04
Play from here
Copy citation
And finding the right skill sets that you can plug into that in kind of a normal,
33:10
Play from here
Copy citation
like smaller company that's out there is really, really challenging.
33:14
Play from here
Copy citation
People are either completely full of crap or they're just like off the rails, like
33:19
Play from here
Copy citation
while doing stuff and have, they basically have no explainability to what they're
33:23
Play from here
Copy citation
doing or you try and build somebody.
33:26
Play from here
Copy citation
So you take Sally in accounting, who's, who's showing that she has like good
33:30
Play from here
Copy citation
ideas, like, okay, cool.
33:32
Play from here
Copy citation
Now you're going to go take Kevin's courses or you're going to like, I'm going
33:35
Play from here
Copy citation
to, I'm going to give you the CloudMax plan and off you're going to go.
33:39
Play from here
Copy citation
That is not like a gentle learning curve.
33:42
Play from here
Copy citation
So you have this gap.
33:44
Play from here
Copy citation
And so you now have identified the problem.
33:46
Play from here
Copy citation
You've identified the ROI of the problem.
33:49
Play from here
Copy citation
And there's really no great way of solving it at the second, short of Aaron
33:54
Play from here
Copy citation
spending yet another weekend, like vibe coding your way through and admit it.
33:58
Play from here
Copy citation
You really like it because it's super fun.
34:00
Play from here
Copy citation
Right.
34:00
Play from here
Copy citation
But, but it's not scalable.
34:02
Play from here
Copy citation
It's not sustainable.
34:03
Play from here
Copy citation
And people like us are actually like totally burning a candle at both ends
34:07
Play from here
Copy citation
because we're trying to solve all of these problems because in most
34:11
Play from here
Copy citation
organizations, it is that head of the organization who has the flexibility to
34:16
Play from here
Copy citation
make the decisions and has access to all of the tools who can do it.
34:20
Play from here
Copy citation
So it is, it is really, really hard right now to break through that.
34:24
Play from here
Copy citation
And I think that people there's, there's a risk of disillusionment that cool.
34:28
Play from here
Copy citation
I can do custom GPTs.
34:31
Play from here
Copy citation
Yeah.
34:32
Play from here
Copy citation
Okay.
34:33
Play from here
Copy citation
There's a lot more than custom GPTs that you can do, but your organization
34:36
Play from here
Copy citation
has to decide that that's okay.
34:38
Play from here
Copy citation
And they have to figure out what the guardrails are there.
34:40
Play from here
Copy citation
Yeah.
34:41
Play from here
Copy citation
I think it's, I think it's really, I think it is a huge challenge.
34:44
Play from here
Copy citation
And I think right now the, if I had to like wave a magic wand to get this stuff
34:48
Play from here
Copy citation
done right now, I think it would be take somebody like yourself who really knows
34:51
Play from here
Copy citation
like what's currently capable, right.
34:53
Play from here
Copy citation
Take a really excited young developer and then put those two with Sally from
34:59
Play from here
Copy citation
accounting and for Sally knows what it should be, quote unquote, the young,
35:04
Play from here
Copy citation
excited programmer, like hopefully loves working with all these tools.
35:08
Play from here
Copy citation
And then you got kind of the person who knows the guardrails of like,
35:10
Play from here
Copy citation
here's what can actually work.
35:12
Play from here
Copy citation
And I think it kind of takes somebody who's got to play those roles right now.
35:15
Play from here
Copy citation
And I'm actually been surprised at how many programs I have on like, like
35:20
Play from here
Copy citation
freelance guys, like on Upwork and stuff like that still have not
35:24
Play from here
Copy citation
sunk their teeth into all of this.
35:25
Play from here
Copy citation
Like, there's a huge amount of the tech community that is just still like
35:29
Play from here
Copy citation
wants to know part of all this stuff, which is just mind blowing to me.
35:31
Play from here
Copy citation
My uncle has got, he works at a software company.
35:34
Play from here
Copy citation
He's a really high level guy there.
35:36
Play from here
Copy citation
And he has a bunch of programmers, like, I don't know how many, 50 programmers
35:39
Play from here
Copy citation
underneath him or something like that.
35:40
Play from here
Copy citation
And, and even he says that a lot of his people just are so resistant to it.
35:45
Play from here
Copy citation
I mean, some of them want to talk about like, you know, whatever the ecological
35:49
Play from here
Copy citation
issues and some of them want to talk about the moral issues and then some of
35:53
Play from here
Copy citation
them want to talk about the tech issues.
35:54
Play from here
Copy citation
And it's just like, I mean, we're at, he's like basically at a software company
35:57
Play from here
Copy citation
and like a bunch of the software engineers don't want to use this stuff.
35:59
Play from here
Copy citation
It's like, what is going on?
36:00
Play from here
Copy citation
So I have a sort of a dirty trick I play.
36:02
Play from here
Copy citation
I, when I come in, that's not the dirty trick, but when I'm doing an
36:06
Play from here
Copy citation
assessment of a larger company and they'll say that the organization is
36:10
Play from here
Copy citation
telling leadership, yes, we're using AI.
36:13
Play from here
Copy citation
And you know, you go to the, the controller's office and ask for the
36:17
Play from here
Copy citation
general ledger account for their expenses on like subscriptions and tokens.
36:21
Play from here
Copy citation
And you'll see like a 30 person dev team that maybe has, you know, 15 or
36:28
Play from here
Copy citation
20 Claude hundred dollar max plans.
36:30
Play from here
Copy citation
And it's like, yeah, there's no chance that in a, in a team that big.
36:36
Play from here
Copy citation
You're not using many times that amount in tokens and use and whatever.
36:42
Play from here
Copy citation
And I'm not really talking about token math maxing.
36:44
Play from here
Copy citation
That's like, that's kind of a different deal, but being able to identify like
36:48
Play from here
Copy citation
what people are doing and what people aren't doing.
36:50
Play from here
Copy citation
It's one of the quantifiable ways that, that, that we approach
36:54
Play from here
Copy citation
it and CFOs love it, of course.
36:56
Play from here
Copy citation
Yeah.
36:56
Play from here
Copy citation
It's, it's amazing.
36:57
Play from here
Copy citation
I mean, he's just such fertile ground out there right now.
37:00
Play from here
Copy citation
You know, if you're excited about tech, like what a time to be alive.
37:03
Play from here
Copy citation
My goodness.
37:04
Play from here
Copy citation
Absolutely.
37:05
Play from here
Copy citation
It's, I mean, it is so fun.
37:07
Play from here
Copy citation
Like it really is to have all of these capabilities and you are a dev.
37:12
Play from here
Copy citation
So you've been able to build things most of your career, but it's just faster now.
37:18
Play from here
Copy citation
So you can, you can try things and fail really quick and experiment in new
37:24
Play from here
Copy citation
directions and it's, you know, super fun.
37:27
Play from here
Copy citation
Well, I think the one last point I would leave is on that is like, there
37:30
Play from here
Copy citation
seems to be this bifurcation I see.
37:32
Play from here
Copy citation
I'm curious if you see this in your organization, it's like the, the young,
37:36
Play from here
Copy citation
like just straight out of college or straight out, you know, out of high
37:39
Play from here
Copy citation
school and doesn't want to go to college.
37:41
Play from here
Copy citation
Like personally excited about that.
37:42
Play from here
Copy citation
I think they're very powerful.
37:43
Play from here
Copy citation
And I think the, the person who's maybe from our generation, you know, who
37:47
Play from here
Copy citation
like knows how hard it is to like create something out of nothing with the tech.
37:52
Play from here
Copy citation
And like, just seize the power because like, I mean, I know how many hours
37:56
Play from here
Copy citation
and hours of grind code myself, like my goodness.
37:59
Play from here
Copy citation
Uh, and now seeing how that springboards is so fast.
38:03
Play from here
Copy citation
And I think there's some of the folks in the middle of it are just
38:05
Play from here
Copy citation
kind of like maybe a little lost.
38:07
Play from here
Copy citation
I don't know.
38:08
Play from here
Copy citation
So I would tend to go for maybe more of the little on the older side, forties,
38:11
Play from here
Copy citation
fifties programmers and like the younger side, like the scrappy young twenties.
38:16
Play from here
Copy citation
And I think there's something about the folks in the thirties right now that
38:18
Play from here
Copy citation
are just like, they seem to be the most resistant for some reason.
38:21
Play from here
Copy citation
I don't know if you see that or not.
38:23
Play from here
Copy citation
So we, we, we both have kids sort of comparable ages and
38:26
Play from here
Copy citation
they're not in college yet.
38:27
Play from here
Copy citation
Mine are starting to apply for college.
38:30
Play from here
Copy citation
What are you telling your kids?
38:31
Play from here
Copy citation
Yeah, that's a hard one.
38:32
Play from here
Copy citation
I would really like to find a good, I mean, the kids are funny, right?
38:36
Play from here
Copy citation
Cause like no one wants to listen to dad.
38:37
Play from here
Copy citation
So I need another dad to like step in and tell him like, you got to do this
38:41
Play from here
Copy citation
cause they're not going to listen to me.
38:42
Play from here
Copy citation
Right.
38:43
Play from here
Copy citation
So I need somebody else's dad to like come in.
38:45
Play from here
Copy citation
But, um, yeah, I think they're really, I think that that's another really, really
38:49
Play from here
Copy citation
good, if somebody wants like a side project that what a great side project
38:52
Play from here
Copy citation
on how to teach kids the latest and greatest on AI to make it fun for them.
38:56
Play from here
Copy citation
So there's a, there's, they're opening an alpha school in park city here, here
39:01
Play from here
Copy citation
in Utah, uh, which is, it's going to be interesting to see whether
39:04
Play from here
Copy citation
or not that plays out, you know, what alpha school is, right?
39:07
Play from here
Copy citation
Okay.
39:07
Play from here
Copy citation
Alpha school is an AI forward, very well-funded startup that basically
39:12
Play from here
Copy citation
does self-directed AI education.
39:15
Play from here
Copy citation
And then combines that with like social activities.
39:19
Play from here
Copy citation
So you're synthesizing the information in the morning, as far as the arithmetic
39:23
Play from here
Copy citation
and the literature and whatever.
39:25
Play from here
Copy citation
And then the rest of the day is around cooperative and collaborative activities
39:29
Play from here
Copy citation
that, that pull the human back in.
39:32
Play from here
Copy citation
And I think it's probably pretty rough around the edges right now, but it's,
39:35
Play from here
Copy citation
it's also probably going to be very valuable because those human skills, as
39:41
Play from here
Copy citation
far as communication and, and like group management and whatever, ironically apply
39:46
Play from here
Copy citation
more to technology right now than, than they ever have before.
39:50
Play from here
Copy citation
Yeah.
39:50
Play from here
Copy citation
I'll have to check that out.
39:51
Play from here
Copy citation
So let me ask you, I have a closing question of what's the tool that you
39:55
Play from here
Copy citation
couldn't live without or a couple of tools that you couldn't live without?
39:58
Play from here
Copy citation
I mean, I'm in cloud code CLI all day, every day.
40:01
Play from here
Copy citation
It's that for sure.
40:02
Play from here
Copy citation
So cloud code, okay.
40:03
Play from here
Copy citation
The one, I mean, I mean, I like codex too, by the way, I like, I like them both.
40:07
Play from here
Copy citation
And actually I think that there's a lot, you've probably already seen it and
40:10
Play from here
Copy citation
talked about it, but the power of battling them against each other.
40:13
Play from here
Copy citation
I think it was just phenomenal.
40:14
Play from here
Copy citation
I mean, codex has been really good for me lately.
40:16
Play from here
Copy citation
Finding, finding bugs that cloud did not find, but then cloud
40:20
Play from here
Copy citation
can like grind out the code.
40:21
Play from here
Copy citation
So that's, that's what I like.
40:22
Play from here
Copy citation
Those are the ones that I really, I think we just draw the line there.
40:25
Play from here
Copy citation
For me, it's funny.
40:26
Play from here
Copy citation
I mean, I can't live without sort of the, the basic integrated chat bot experience.
40:33
Play from here
Copy citation
Since I run my day through a chief of staff, we were talking offline about that,
40:38
Play from here
Copy citation
about being able to, to do it through a, like a super app, but I, for me, it's
40:44
Play from here
Copy citation
dictation, I just can't stop talking about dictation and how important that it's been.
40:50
Play from here
Copy citation
I use whisper flow, no affiliation, but I just looked, I had dictated 1.4
40:55
Play from here
Copy citation
million words in the last like eight months or something like that, which is
40:59
Play from here
Copy citation
tomes and tomes that are coming out of me.
41:02
Play from here
Copy citation
Like literally my carpal tunnel has gone away because I'm like now just dictating
41:07
Play from here
Copy citation
everything like I'm working with Jarvis.
41:10
Play from here
Copy citation
Well, okay.
41:10
Play from here
Copy citation
And if you had any closing pieces of advice for other business owners out there
41:14
Play from here
Copy citation
as somebody who is on this ragged journey, what would it be?
41:18
Play from here
Copy citation
Wow.
41:18
Play from here
Copy citation
Closing piece of advice.
41:19
Play from here
Copy citation
I would say, find somebody that you can talk to about this stuff, you know, a small
41:25
Play from here
Copy citation
group, one-on-one person, hire a really good programmer, do, do, do something that
41:30
Play from here
Copy citation
you can, continue the excitement around this, because I think that excitement
41:35
Play from here
Copy citation
brings the energy and that energy helps you push through the obstacles of, of
41:39
Play from here
Copy citation
getting things done and figuring these things out.
41:41
Play from here
Copy citation
And I think for me that that's just a really important piece.
41:44
Play from here
Copy citation
So be part of some AI community or have a little bit close to AI buddy.
41:48
Play from here
Copy citation
Awesome.
41:48
Play from here
Copy citation
Aaron, thank you so much for joining and anybody else listening, if you have
41:52
Play from here
Copy citation
specific questions, please send them in.
41:55
Play from here
Copy citation
I'm Kevin at ascendlabs.ai.
41:58
Play from here
Copy citation
We want to be as practical, practical as we can.
42:01
Play from here
Copy citation
This is called AI at work after all.
42:03
Play from here
Copy citation
So thanks again, and we will see you next week.
42:07
Play from here
Copy citation
Thanks Kevin.
42:12
Play from here
Copy citation
Thanks for listening.
42:13
Play from here
Copy citation
If you'd like to send us a question to answer on the show, email me at
42:17
Play from here
Copy citation
Elijah at spark6.com that is S P A R K the number six dot com or Kevin at
42:23
Play from here
Copy citation
Kevin at ascendlabs.ai to learn more about AI strategy and training for your
42:29
Play from here
Copy citation
business.
42:29
Play from here
Copy citation
Visit Kevin's company ascend AI labs at ascendlabs.ai and to learn more about
42:36
Play from here
Copy citation
design and development of custom AI solutions, visit my agency at spark6.com.
42:42
Play from here
Copy citation
You can also sign up for my free newsletter, AI framework Friday by
42:46
Play from here
Copy citation
visiting spark6.com forward slash newsletter.