Single-Player vs. Multiplayer AI: Why Most Companies Are Getting AI Wrong | AI For The C-Suite EP 74

1
00:00:01,420 --> 00:00:03,312
Right, and we are recording. Perfect.
2
00:00:03,312 --> 00:00:07,315
I'm Chad Harvey,
3
00:00:07,315 --> 00:00:09,156
and this is AI for the C-suite,
4
00:00:09,156 --> 00:00:13,610
the show for senior leaders who know AI
matters and need to figure out what
5
00:00:13,610 --> 00:00:17,363
to do about it. Each episode, we dig into
what it all actually means
6
00:00:17,363 --> 00:00:20,606
for mid-market organizations, how it's
changing decisions,
7
00:00:20,606 --> 00:00:23,949
strategy, leadership, and the nature of
work.
8
00:00:23,949 --> 00:00:25,250
Let's get into it.
9
00:00:26,656 --> 00:00:31,391
My guest today has spent his whole career
on the unglamorous half of tech,
10
00:00:31,391 --> 00:00:33,433
the part that nobody puts in the keynote.
11
00:00:33,433 --> 00:00:38,237
Justin Watts started out in large scale
project management at IBM,
12
00:00:38,237 --> 00:00:40,460
then spent nearly eight years at Metalab,
13
00:00:40,460 --> 00:00:42,942
the product studio that helped design
Slack.
14
00:00:42,942 --> 00:00:46,886
When there, he shipped work for companies
like Uber and Amazon
15
00:00:46,966 --> 00:00:50,890
and eventually ran ops. These days he's
the co-founder of Switchboard.
16
00:00:51,202 --> 00:00:55,643
Where he and his team build the internal
software and AI automation that growing
17
00:00:55,643 --> 00:00:57,804
companies just can't buy off the shelf.
18
00:00:57,804 --> 00:01:00,404
And here's why he's here today.
19
00:01:00,404 --> 00:01:03,265
Everybody's chasing shiny AI tools,
20
00:01:03,265 --> 00:01:06,866
yet Justin spends his days on the thing
underneath all of it:
21
00:01:06,866 --> 00:01:10,647
your data, your processes, the plumbing.
22
00:01:10,647 --> 00:01:12,748
He's got a great way of putting it.
23
00:01:12,748 --> 00:01:17,289
Most companies are trying to run
multiplayer AI on a single-player setup,
24
00:01:17,289 --> 00:01:18,940
and we're gonna dig into that first.
25
00:01:18,940 --> 00:01:20,650
So as we get into it,
26
00:01:20,866 --> 00:01:23,284
Justin, welcome to AI for the C suite.
27
00:01:24,246 --> 00:01:25,938
What a tee up. Thanks, Chad. Good to be
here.
28
00:01:25,938 --> 00:01:26,938
~
29
00:01:26,586 --> 00:01:31,978
Absolutely. So I deliberately put that
line in your introduction here because
30
00:01:31,978 --> 00:01:34,730
you and I had some conversation before the
episode today.
31
00:01:34,730 --> 00:01:36,706
And I love the way you phrased it.
32
00:01:36,706 --> 00:01:43,013
You phrased it. You said that single
player versus multiplayer AI is what's
going
33
00:01:43,013 --> 00:01:47,845
on out there right now. And I want you to
walk us through what you mean when
34
00:01:47,845 --> 00:01:50,256
you say single player versus multiplayer
AI.
35
00:01:50,256 --> 00:01:51,737
Walk us through that distinction.
36
00:01:51,737 --> 00:01:53,838
What's it mean for organization and
37
00:01:54,210 --> 00:01:58,268
Why is it the thing that is quietly
deciding whether AI initiatives
38
00:01:58,268 --> 00:01:59,314
are actually working or not?
39
00:02:00,792 --> 00:02:03,883
Yeah, think a lot of teams view AI still
as what it launched as,
40
00:02:03,883 --> 00:02:07,755
which is a chat box sitting there for one
person to plug whatever they want into
41
00:02:05,549 --> 00:02:06,549
Mm-hmm.
42
00:02:07,755 --> 00:02:09,406
and use whatever they get back from it.
43
00:02:09,406 --> 00:02:11,617
And so that is wonderful and all,
44
00:02:11,617 --> 00:02:13,127
but that's not how businesses work.
45
00:02:13,127 --> 00:02:19,930
And so the the thing I'm seeing execs and
leaders grasp is if I can do this,
46
00:02:19,930 --> 00:02:23,472
how does my team do this? And a chat box
is not the answer to that.
47
00:02:23,472 --> 00:02:28,614
If to overly simplify, if a business is an
assembly line of process and you know,
48
00:02:28,614 --> 00:02:29,614
steps
49
00:02:29,512 --> 00:02:31,644
One, four, and nine are done by one
person.
50
00:02:31,644 --> 00:02:33,886
Steps two, six, and twelve are done by
someone else.
51
00:02:33,886 --> 00:02:38,109
And then there's a step in between where
you gotta send an email to chase
something,
52
00:02:38,109 --> 00:02:41,832
and then some other step you're plugging
something into a spreadsheet somewhere.
53
00:02:41,832 --> 00:02:45,755
As you can probably imagine, AI is not
built to magically know which step
54
00:02:45,755 --> 00:02:48,138
is active right now, who needs to do what
and when.
55
00:02:48,138 --> 00:02:51,861
And so businesses who are saying we we
want AI,
56
00:02:51,861 --> 00:02:56,715
~ and viewing it as this magical silver
bullet through the lens of a chat box,
57
00:02:56,715 --> 00:02:58,144
and that's all they're used to are
58
00:02:58,968 --> 00:03:02,280
quickly becoming disenchanted with how AI
works in their business because they're
59
00:03:02,280 --> 00:03:04,321
not looking at it in that multiplayer way.
60
00:03:04,321 --> 00:03:05,942
Just like video game equivalent of,
61
00:03:05,942 --> 00:03:10,825
you know, you're playing Tomb Raider or or
Mario.
62
00:03:10,825 --> 00:03:13,847
It's you playing Mario and you're going
through each level by yourself
63
00:03:13,887 --> 00:03:19,330
is very different than Fortnite ~ or
Roblox as some folks may play themselves
64
00:03:19,330 --> 00:03:22,082
or their kids play. Like it's them with a
hundred other people doing something
65
00:03:22,082 --> 00:03:24,043
together. And that's how business
operates.
66
00:03:24,043 --> 00:03:26,074
And AI needs to be thought of that way as
in
67
00:03:26,424 --> 00:03:28,126
How do we have steps one, three,
68
00:03:28,126 --> 00:03:29,948
and six done by one person and steps two,
69
00:03:29,948 --> 00:03:33,453
five, and twelve by someone else ~
process-wise,
70
00:03:33,453 --> 00:03:35,555
have their AI kind of working together
that way?
71
00:03:35,555 --> 00:03:37,788
So that's a bit of a ramble of a tee up to
it.
72
00:03:37,788 --> 00:03:40,321
We can go down any of those rabbit holes I
probably just opened up,
73
00:03:40,321 --> 00:03:42,454
but that's how I I frame it to folks to
think about.
74
00:03:43,222 --> 00:03:44,522
No, I think that's a good framing.
75
00:03:44,522 --> 00:03:48,063
And it makes me wonder whether that makes
this podcast the equivalent
76
00:03:48,063 --> 00:03:50,124
of people that watch other people playing
video games.
77
00:03:50,124 --> 00:03:54,065
~ I don't know. So all right, be that as
yeah,
78
00:03:52,193 --> 00:03:55,808
That's all market to itself. Twitch is
Twitch is a multi billion dollar company.
79
00:03:54,065 --> 00:03:57,385
I know. It certainly is.
80
00:03:57,385 --> 00:04:01,406
~ all right. So w l let's let's go in
here.
81
00:04:01,406 --> 00:04:08,328
~ how does that land ~ when you have an
executive ~ who thinks that they're
82
00:04:08,328 --> 00:04:12,019
already using AI, but it's really single
player mode like you've described.
83
00:04:12,019 --> 00:04:13,019
Is that ~
84
00:04:13,451 --> 00:04:16,106
A difficult conversation? Is that an aha
moment?
85
00:04:16,106 --> 00:04:17,504
Is it a variety of reactions?
86
00:04:18,422 --> 00:04:21,363
It's usually a difficult conversation that
leads to an aha moment.
87
00:04:21,363 --> 00:04:25,994
As in, you know, a use case for a lot of
people is whatever their business might
be,
88
00:04:25,994 --> 00:04:29,405
let's say professional services of some
shape or form,
89
00:04:29,405 --> 00:04:31,226
they have someone who starts out as a
lead,
90
00:04:31,226 --> 00:04:34,787
that lead becomes a client or a customer
who's getting a service or a product.
91
00:04:34,787 --> 00:04:36,887
That service or product is fulfilled,
92
00:04:36,887 --> 00:04:38,868
and then there's a financial transaction
for that.
93
00:04:38,868 --> 00:04:42,619
All of those different steps, the the hard
part of the conversation usually
94
00:04:42,619 --> 00:04:44,360
is what does it look like to do those?
95
00:04:44,360 --> 00:04:45,400
What does your data look like?
96
00:04:45,838 --> 00:04:49,160
Do you have a CRM that calls the customer
one thing and a financial system that's
97
00:04:49,160 --> 00:04:51,681
disconnected from the CRM calling the
customer something different?
98
00:04:51,681 --> 00:04:55,473
Do you have nine spreadsheets in between
for all the people fulfilling
99
00:04:55,473 --> 00:04:59,315
the service or the product to keep track
of whatever customer support needs versus
100
00:04:59,315 --> 00:05:01,847
operations versus sales, whatever that
might be?
101
00:05:01,847 --> 00:05:04,448
That's usually the hard part of the
conversation is well,
102
00:05:04,448 --> 00:05:08,040
how can you point AI at what's the status
of this customer when
103
00:05:08,040 --> 00:05:10,951
the customer had lives in six different
systems that aren't disconnected?
104
00:05:10,951 --> 00:05:12,752
They're called four different things,
105
00:05:12,752 --> 00:05:14,473
might be a legal name in one system,
106
00:05:14,473 --> 00:05:15,494
a shorthand in another.
107
00:05:15,938 --> 00:05:18,359
their public name in another, whatever it
looks like.
108
00:05:18,359 --> 00:05:22,431
The hard part is to deduce, is your data
ready for AI?
109
00:05:22,431 --> 00:05:25,502
Because AI will get very confused very
quickly if you say,
110
00:05:25,502 --> 00:05:28,263
what's the status of Chad Harvey Inc.
111
00:05:28,263 --> 00:05:31,945
but Chad is spelt six different ways in
six different places.
112
00:05:31,945 --> 00:05:35,406
So that's usually what leads to the aha
moment of,
113
00:05:35,406 --> 00:05:40,048
well, you can't just point AI at
spreadsheets that live on people's local
computers,
114
00:05:40,048 --> 00:05:43,209
process that lives in people's heads and
has never been documented,
115
00:05:43,209 --> 00:05:44,600
or five systems where
116
00:05:44,931 --> 00:05:48,044
Half the job people's job are just
repeating information between these
systems
117
00:05:48,044 --> 00:05:49,044
because they're not connected.
118
00:05:49,033 --> 00:05:52,477
So once you climb that mountain and have
that aha moment,
119
00:05:52,477 --> 00:05:54,339
it then becomes very easy to see.
120
00:05:54,339 --> 00:05:56,145
Okay, well, if we get our data in order,
121
00:05:56,145 --> 00:05:58,782
it actually becomes pretty quick to
install,
122
00:05:58,782 --> 00:06:00,243
so to speak, AI in our company,
123
00:06:00,243 --> 00:06:03,506
because you're not spending all your time
jerrying,
124
00:06:03,506 --> 00:06:08,851
jerry raking these workarounds and nuances
that you have to bake in to say,
125
00:06:05,656 --> 00:06:06,656
Mm-hmm.
126
00:06:08,851 --> 00:06:11,783
well, when we say this, we say we mean
look at that system.
127
00:06:11,783 --> 00:06:12,954
When we say this, we mean
128
00:06:13,698 --> 00:06:17,066
that kind of service. If everything is
consistent and connected data wise,
129
00:06:17,066 --> 00:06:19,772
AI starts to become a lot easier to adopt
and implement.
130
00:06:19,772 --> 00:06:23,762
And really what the goal is is to see an
actual ROI and benefit out of
131
00:06:24,992 --> 00:06:28,885
One ring to rule them all. Yes.
132
00:06:26,637 --> 00:06:27,637
Yes.
133
00:06:28,885 --> 00:06:34,450
This is fascinating because I'm I'm always
very intrigued when someone
134
00:06:34,570 --> 00:06:38,563
~ has a conversation with me and we touch
on all these different pieces because
135
00:06:38,563 --> 00:06:40,334
you're you're talking about business
processes,
136
00:06:40,334 --> 00:06:41,766
you're talking about data integrity,
137
00:06:41,766 --> 00:06:43,357
you're talking about multiple systems,
138
00:06:43,357 --> 00:06:46,499
then we're talking about AI. And and I'm
wondering,
139
00:06:46,499 --> 00:06:48,831
when you walk into an organization,
140
00:06:48,831 --> 00:06:54,135
~ do you ever find that it is just simply
more than you can handle?
141
00:06:54,135 --> 00:06:55,135
Because
142
00:06:54,808 --> 00:06:56,279
They don't have their processes mapped.
143
00:06:56,279 --> 00:06:59,813
They've got the multiple systems with the
different ~ kind of form fields that
144
00:06:59,813 --> 00:07:02,705
you're talking about where the same data
is called different things.
145
00:07:02,705 --> 00:07:08,581
How do you know when that possible
engagement that you've been thinking about
146
00:07:08,581 --> 00:07:11,384
is just ~ that rarest of of materials,
147
00:07:11,384 --> 00:07:14,847
it's unobtainium, right? How how do you
know that when you walk in there
148
00:07:13,187 --> 00:07:14,187
Yeah.
149
00:07:14,847 --> 00:07:15,858
and where do you draw the line?
150
00:07:17,058 --> 00:07:20,661
Yeah, the conversation ~ I often have with
execs is like,
151
00:07:20,661 --> 00:07:22,762
what is the surface area of where you want
AI?
152
00:07:22,762 --> 00:07:26,495
Is it is there a is there a problem child
in your business of a department
153
00:07:23,086 --> 00:07:24,086
Mm.
154
00:07:26,495 --> 00:07:29,197
or a team where it's like you know that if
you doubled in revenue,
155
00:07:29,197 --> 00:07:32,390
you'd have to double the headcount because
it's just such mayhem and
156
00:07:32,470 --> 00:07:34,321
not driven by any sort of formal process,
157
00:07:34,321 --> 00:07:37,756
technology, data, et cetera? Or are you
looking for a complete reinvention
158
00:07:37,756 --> 00:07:42,087
of how the business operates? Or is there
kind of a few bottleneck areas that
159
00:07:42,087 --> 00:07:43,087
you want to focus on?
160
00:07:43,417 --> 00:07:45,968
That's why I try to dissuade people from
the notion of like we want
161
00:07:45,968 --> 00:07:48,098
to reinvent the business as quickly as
possible.
162
00:07:48,098 --> 00:07:50,509
~ because very few are prepared to do
that.
163
00:07:50,509 --> 00:07:53,269
We are helping a couple teams do that
right now,
164
00:07:53,269 --> 00:07:55,440
but it is not a three-month, you know,
165
00:07:55,440 --> 00:08:00,851
quick win mini project. It is a year-long
effort ~ to turn process
166
00:08:00,851 --> 00:08:04,432
in people's heads and data scattered
across different places and tools that
don't
167
00:08:04,432 --> 00:08:08,733
talk to each other into a cohesive system
with a cohesive process throughout that
168
00:08:08,733 --> 00:08:10,994
requires effort. ~ and so
169
00:08:11,298 --> 00:08:14,020
For that reason, to answer your question
more plainly,
170
00:08:14,020 --> 00:08:16,362
it really depends on that surface area.
171
00:08:16,362 --> 00:08:18,114
What do you want to change in your
business?
172
00:08:18,114 --> 00:08:20,126
If it's everything, then we can do that,
173
00:08:20,126 --> 00:08:23,268
but are you ready for it? ~ and inversely,
174
00:08:21,485 --> 00:08:22,485
Right.
175
00:08:23,268 --> 00:08:25,460
usually I'm not trying to, you know,
176
00:08:25,460 --> 00:08:27,852
pitch some grand transformation project.
177
00:08:27,852 --> 00:08:31,235
People want to see wins in months and
quarters,
178
00:08:31,235 --> 00:08:34,558
not years and years. So that's how I
approach it.
179
00:08:35,278 --> 00:08:37,409
That and and what you just described,
180
00:08:37,409 --> 00:08:39,399
~ I cannot remember the name of the model,
181
00:08:39,399 --> 00:08:43,470
but there's a a pyramid model with three
levels where it talks about transformation
182
00:08:43,470 --> 00:08:45,381
at the top where almost nobody is yet,
183
00:08:45,381 --> 00:08:50,092
and then we've got cost savings and ~
dramatic increases in efficiency
184
00:08:50,092 --> 00:08:52,002
at kind of those bottom two pyramid
levels.
185
00:08:52,002 --> 00:08:58,301
And what I wonder is as we begin to to
move into
186
00:08:58,301 --> 00:09:00,415
~ a more sophisticated approach to AI,
187
00:09:00,415 --> 00:09:02,485
which you are clearly taking with your
clients,
188
00:09:02,485 --> 00:09:04,946
do you feel that many of them
189
00:09:05,346 --> 00:09:09,148
will stop at the efficiency and the cost
savings that they
190
00:09:09,148 --> 00:09:11,149
are achieving through these
implementations?
191
00:09:11,149 --> 00:09:15,771
Or do you think that some will eventually
become ready for that true transformation
192
00:09:15,771 --> 00:09:19,612
that you talked about there? Because I I
see the potential with a lot
193
00:09:19,612 --> 00:09:22,594
of what's ~ being the the train tracks,
194
00:09:22,594 --> 00:09:26,365
if you will, that are being laid down
right now for transformative,
195
00:09:26,365 --> 00:09:29,247
massive transformative efforts for
organizations,
196
00:09:29,247 --> 00:09:31,968
but I'm not sure that the appetite is
actually there.
197
00:09:33,357 --> 00:09:36,339
Yeah, I think we're we're seeing two
things play out right now.
198
00:09:36,339 --> 00:09:39,800
Or I'm in my experience with all the teams
that we get the the privilege
199
00:09:37,912 --> 00:09:38,912
Yeah.
200
00:09:39,800 --> 00:09:42,812
of getting to see how they're approaching
it and what we get to do together.
201
00:09:41,752 --> 00:09:42,752
Yeah.
202
00:09:42,812 --> 00:09:44,453
I'm seeing two things play out.
203
00:09:44,453 --> 00:09:48,285
One is the technology is just inherently
getting better and these things
204
00:09:48,285 --> 00:09:49,816
are getting easier and easier to do.
205
00:09:49,816 --> 00:09:52,257
~ as in the transformation itself,
206
00:09:52,257 --> 00:09:55,819
you don't have to put as much guardrails
and harnessing around
207
00:09:55,819 --> 00:09:58,280
the models because the models themselves
are getting so capable.
208
00:09:58,280 --> 00:10:01,012
So I think just the natural technological
progression,
209
00:09:58,703 --> 00:09:59,703
Mm.
210
00:10:01,012 --> 00:10:02,012
which is why
211
00:10:02,166 --> 00:10:04,168
If you think back to the digital
transformation era,
212
00:10:04,168 --> 00:10:08,992
it was a huge lift both financially and
time-wise to go from email on
213
00:10:08,992 --> 00:10:11,804
a server in your closet to in the cloud.
214
00:10:11,804 --> 00:10:15,668
And now anyone can spin up a Microsoft 365
or Google Workspace account
215
00:10:15,668 --> 00:10:18,620
in two minutes. So the technology part is
getting better.
216
00:10:18,620 --> 00:10:23,965
But the thing that has stood out to me ~
for teams wanting just
217
00:10:23,965 --> 00:10:26,947
the short-term ROI benefits versus those
who are rethinking
218
00:10:26,947 --> 00:10:29,970
and transforming their entire business is
historically.
219
00:10:30,304 --> 00:10:34,966
Every era of digital transformation has
been if if you get on board,
220
00:10:34,966 --> 00:10:36,707
you get some benefits, but if you don't,
221
00:10:36,707 --> 00:10:41,309
it's just a margin hit. So it's kind of
not not opting into getting your file
server
222
00:10:41,309 --> 00:10:44,540
out of the closet into the cloud was a tax
on your business.
223
00:10:43,502 --> 00:10:45,044
Yeah. Yeah.
224
00:10:44,540 --> 00:10:47,001
It was, you know, I'm maybe making five,
225
00:10:47,001 --> 00:10:51,073
10, 15% less margin, but my competitors
that didn't do that either
226
00:10:51,073 --> 00:10:55,075
are also fine with, you know, spreadsheets
that are I have to VPN into a server
227
00:10:55,075 --> 00:10:57,346
in a closet. So it used to be a tax.
228
00:10:58,072 --> 00:10:59,333
But now it's become a countdown.
229
00:10:59,333 --> 00:11:00,823
If you're not willing to transform,
230
00:11:00,823 --> 00:11:05,735
we're seeing teams start from day one
competing with businesses that have been
231
00:11:05,735 --> 00:11:08,797
around for 20 years and they are within a
year approaching
232
00:11:08,797 --> 00:11:12,749
the same capabilities with 25 people as a
business of 250 people.
233
00:11:12,749 --> 00:11:16,990
So to avoid the transformation era of
digital transformation,
234
00:11:16,990 --> 00:11:18,391
cloud transformation, et cetera,
235
00:11:18,391 --> 00:11:21,362
that might have been a tax, but now it's
becoming a countdown where
236
00:11:21,362 --> 00:11:24,074
if you don't participate, then you're
gonna get run out of business.
237
00:11:24,074 --> 00:11:25,074
And that's not
238
00:11:25,016 --> 00:11:28,297
trying to be fear-mongering or the snake
oil of AI will solve everything.
239
00:11:28,297 --> 00:11:31,958
It won't, but it presents an opportunity
to completely change
240
00:11:31,958 --> 00:11:33,929
how a business operates and becomes more
efficient.
241
00:11:33,929 --> 00:11:37,840
So I I think right now it's totally fine
if teams are just looking
242
00:11:37,840 --> 00:11:39,390
for that first rung or two on the ladder,
243
00:11:39,390 --> 00:11:41,851
as you mentioned, but that rung three,
244
00:11:41,851 --> 00:11:44,612
if that's true transformation of the
business process and model
245
00:11:44,612 --> 00:11:49,624
and and operational infrastructure in
three or four years,
246
00:11:49,624 --> 00:11:50,894
people are going to have no choice.
247
00:11:52,995 --> 00:11:58,078
I appreciate how you couched your remarks
~ against the idea of fear mongering
248
00:11:58,199 --> 00:12:01,671
and I I absolutely love the idea of a tax
versus a countdown.
249
00:12:01,671 --> 00:12:06,484
I think that's a very ~ stark visual that
should drive it home for a lot of people.
250
00:12:06,484 --> 00:12:12,508
And and I agree that we are in a era here
where you need
251
00:12:12,508 --> 00:12:17,341
to take the longer term and by longer term
it could be even five or ten years,
252
00:12:17,341 --> 00:12:20,724
~ as opposed to fifteen or twenty
implications of this very seriously.
253
00:12:20,724 --> 00:12:21,724
So
254
00:12:21,587 --> 00:12:24,453
I I really like that. I haven't heard
anybody phrase it quite that way in terms
255
00:12:24,453 --> 00:12:28,963
of a tax versus a countdown. So kudos to
you for that little visual there.
256
00:12:28,963 --> 00:12:32,130
Yeah.
257
00:12:29,912 --> 00:12:32,154
The if I can add one more, the way I've
I've thought
258
00:12:32,154 --> 00:12:34,045
about as well, more visually or
anecdotally,
259
00:12:34,045 --> 00:12:36,107
if it drives it home for people more,
260
00:12:36,107 --> 00:12:38,789
is anytime you see a montage in a movie,
261
00:12:38,789 --> 00:12:41,712
if you pick picture Rocky punching a
punching bag,
262
00:12:41,712 --> 00:12:43,433
running up the stairs in his tracksuit,
263
00:12:43,433 --> 00:12:46,195
like you could pay to have that done for
you previously.
264
00:12:46,195 --> 00:12:48,277
You could pay a consultant to come in and
say,
265
00:12:48,277 --> 00:12:50,809
Hey, we want to move our email to the
cloud,
266
00:12:50,809 --> 00:12:52,571
whatever it might be. This era,
267
00:12:52,571 --> 00:12:53,962
you can't pay for the montage.
268
00:12:53,962 --> 00:12:56,354
You have to put in the work to transform
your business because it's
269
00:12:54,253 --> 00:12:55,253
Mm-hmm.
270
00:12:56,354 --> 00:12:59,096
not just a switch that flips and suddenly
271
00:12:59,320 --> 00:13:01,852
things are transformed, you actually have
to reinvent the business kind
272
00:13:01,852 --> 00:13:05,695
of top down and bottom up. It's a it's a
grassroots level because obviously there's
273
00:13:05,695 --> 00:13:08,077
many folks scared for the impacts to their
job.
274
00:13:08,077 --> 00:13:11,020
But so far we've not seen jobs being
eliminated.
275
00:13:11,020 --> 00:13:13,922
It's that people are creating new jobs
within business just because of
276
00:13:13,922 --> 00:13:15,163
the time that they're freed up from,
277
00:13:15,163 --> 00:13:18,215
you know, people didn't they everyone grew
up with the dream of astronaut,
278
00:13:18,215 --> 00:13:20,848
doctor, lawyer. No one grew up with the
dream of I want to move sales around
279
00:13:20,848 --> 00:13:23,780
spreadsheets and reply to Roy in
accounting for the ninth time to chase
280
00:13:23,780 --> 00:13:27,563
an invoice status. So the the montage you
used to be able to pay for,
281
00:13:27,563 --> 00:13:28,744
but you can't really pay for it anymore.
282
00:13:30,031 --> 00:13:33,946
Yeah, fair enough. And my my dream was
kind of just the first part.
283
00:13:33,946 --> 00:13:36,419
It was the astronaut part, but I'm also
still salty.
284
00:13:36,419 --> 00:13:38,691
I've said this before, I I never got my
flying car.
285
00:13:38,691 --> 00:13:41,615
I was promised a flying car and jet packs
when I was a kid.
286
00:13:40,428 --> 00:13:41,725
Where? Yeah.
287
00:13:41,615 --> 00:13:42,615
Where are they?
288
00:13:42,571 --> 00:13:44,661
Yeah. Jetsons, come on, you lied to us.
289
00:13:45,167 --> 00:13:49,789
Exactly. All right. Let me circle back
real quick to where we started because
290
00:13:49,789 --> 00:13:54,161
I I don't want to ~ completely drive by
cause we got into some great stuff,
291
00:13:54,161 --> 00:13:57,207
but I don't want to completely drive by
your single player versus multiplayer.
292
00:13:57,207 --> 00:13:58,819
I had one or two other questions there.
293
00:13:58,819 --> 00:14:04,714
What point do you think a company crosses
into multiplayer
294
00:14:04,714 --> 00:14:07,496
~ levels and is that a danger zone?
295
00:14:09,662 --> 00:14:12,903
It is a danger zone if you treat it as the
Wild West.
296
00:14:12,903 --> 00:14:17,284
~ so a lot of times, like if you picture a
sales team,
297
00:14:13,198 --> 00:14:14,198
Mm.
298
00:14:17,284 --> 00:14:21,726
there's ~ you know, let's call it a a
10-person sales team in a business.
299
00:14:21,726 --> 00:14:24,856
There might be a playbook of generally
this is,
300
00:14:24,856 --> 00:14:27,297
you know, the point at which we do
discovery and this is how we
301
00:14:27,297 --> 00:14:29,818
do discovery and this is how we build a
proposal and how we close
302
00:14:29,818 --> 00:14:34,489
the business and kick it off. But very
rarely will all 10 of those people have
303
00:14:34,489 --> 00:14:36,630
the exact same process for how they get
there.
304
00:14:37,186 --> 00:14:39,677
And so that's where it starts to turn into
the Wild West,
305
00:14:39,677 --> 00:14:41,087
where I've seen some teams just say,
306
00:14:41,087 --> 00:14:45,438
Hey, here's the equivalent of a
16-year-old with a a learner's permit.
307
00:14:45,438 --> 00:14:47,349
You just give them the keys to the car and
say good luck.
308
00:14:47,349 --> 00:14:50,800
Instead, what is successful is we're gonna
take you and show you
309
00:14:50,800 --> 00:14:53,691
how to parallel park and shoulder check
and user blinkers and whatnot,
310
00:14:53,691 --> 00:14:56,852
and giving those guardrails to a team
where a you're not letting it
311
00:14:56,852 --> 00:15:00,883
be the Wild West, as in there's now six AI
agents all trying to do
312
00:15:00,883 --> 00:15:03,244
the same thing differently, then you're
back to square one of what are
313
00:15:03,244 --> 00:15:04,544
we doing here if we're still
314
00:15:04,780 --> 00:15:07,131
you know, all in silos, that single player
experience.
315
00:15:07,131 --> 00:15:09,872
and B, we're all working towards the same
thing,
316
00:15:09,872 --> 00:15:12,173
which is to take the work off of our
plates,
317
00:15:12,173 --> 00:15:13,974
not just speed up little parts.
318
00:15:13,974 --> 00:15:16,235
It's, you know, that entire proposal
process.
319
00:15:16,235 --> 00:15:19,466
If you've got a good transcript from a
call and you've got the context
320
00:15:19,466 --> 00:15:20,737
of what a good proposal looks like,
321
00:15:20,737 --> 00:15:22,338
the proposal is written for you.
322
00:15:22,338 --> 00:15:25,359
And yes, you as a human still need to be
in the loop to review it and whatnot.
323
00:15:25,359 --> 00:15:28,781
But letting it be the Wild West is where I
see teams get frustrated
324
00:15:28,781 --> 00:15:31,742
at their progress because they never
really gave themselves guardrails to say,
325
00:15:32,108 --> 00:15:33,442
Yes, there's ten people on this team,
326
00:15:33,442 --> 00:15:36,311
but it doesn't mean we should have 10
people building their own AI system.
327
00:15:36,311 --> 00:15:39,902
It should be that single system for a
multiplayer environment.
328
00:15:40,726 --> 00:15:44,548
Mm-hmm. I think that's ~ that's a really
fair point.
329
00:15:44,548 --> 00:15:46,709
And what when you say Wild West,
330
00:15:46,709 --> 00:15:49,641
I I think something that often comes to my
mind too,
331
00:15:49,641 --> 00:15:51,222
and I'm I'm segueing here a little bit,
332
00:15:51,222 --> 00:15:55,354
but one of the things that I often think
of is what you touched on ~ earlier
333
00:15:55,354 --> 00:15:58,085
in our conversation, companies where
everything's a spreadsheet,
334
00:15:58,085 --> 00:16:01,737
right? So we've got that kind of is the
wild west over here.
335
00:16:01,737 --> 00:16:05,340
And then on the other side, we've got
overstructured shops that are living
336
00:16:05,340 --> 00:16:07,140
on these big enterprise platforms.
337
00:16:07,140 --> 00:16:09,772
And I I'm sure that you've seen both.
338
00:16:10,270 --> 00:16:14,222
And I'm sure that there are benefits and
drawbacks to both.
339
00:16:14,222 --> 00:16:17,795
So maybe we can spend a little time
talking about that as well,
340
00:16:17,795 --> 00:16:23,358
because this idea of ~ mid-market being
more nimble and being able to move faster,
341
00:16:23,358 --> 00:16:25,569
I think is legitimate in many ways.
342
00:16:25,569 --> 00:16:29,342
However, there's a lot more wild westing
that's going on with those types
343
00:16:29,342 --> 00:16:31,983
of organizations. So with all that in
mind,
344
00:16:31,983 --> 00:16:34,705
~ I don't know. I there was a question in
there somewhere.
345
00:16:34,705 --> 00:16:35,705
Your thoughts.
346
00:16:35,776 --> 00:16:36,776
Yeah.
347
00:16:38,008 --> 00:16:43,181
I think of like most people don't realize
Excel came out the same year that
348
00:16:43,181 --> 00:16:44,332
the movie Back to the Future did.
349
00:16:44,332 --> 00:16:46,234
Like that's how old that that product is.
350
00:16:44,865 --> 00:16:50,034
my god. ~
351
00:16:46,234 --> 00:16:51,217
I know. makes me feel old. But but my
point being of like,
352
00:16:51,217 --> 00:16:54,219
yes, Excel has stood the test of time for
good reason because it is
353
00:16:54,219 --> 00:16:56,281
so flexible and it's easy to understand.
354
00:16:56,281 --> 00:17:00,394
The drawback is most businesses run on it
without a realization of
355
00:17:00,394 --> 00:17:02,205
how much they're holding themselves back
by,
356
00:17:02,205 --> 00:17:03,346
you know, ~
357
00:17:03,426 --> 00:17:07,007
This this notion of our operating system
is an Excel model or
358
00:17:07,007 --> 00:17:09,909
an Excel spreadsheet that gets passed
around and synced between 10 people,
359
00:17:09,909 --> 00:17:12,569
but they're still spending most of their
day just filling out fields.
360
00:17:12,569 --> 00:17:15,011
What if you could have all of that filled
out for you?
361
00:17:15,011 --> 00:17:17,492
And what if the process was invisible for
getting there?
362
00:17:17,492 --> 00:17:20,992
The the custom software side is that's
what's fascinating to watch.
363
00:17:20,992 --> 00:17:22,983
Is the mental math used to be,
364
00:17:22,983 --> 00:17:26,055
well, if I want to build something that
replaces Excel,
365
00:17:26,055 --> 00:17:28,627
it's either s some sort of SAS off the
shelf,
366
00:17:28,627 --> 00:17:32,128
which most people are increasingly re
367
00:17:32,249 --> 00:17:35,842
Held from because the SaaS model is meant
to A,
368
00:17:35,842 --> 00:17:37,633
B, C-based pricing and gets expensive,
369
00:17:37,633 --> 00:17:40,324
but more so it's never built for you.
370
00:17:40,324 --> 00:17:42,897
It's built for lowest common denominator
of how many businesses can we
371
00:17:42,897 --> 00:17:45,229
get to sign up. But on the flip side,
372
00:17:45,229 --> 00:17:47,391
building custom software for yourself used
to be,
373
00:17:47,391 --> 00:17:50,153
you know, call it 18 months, five
developers,
374
00:17:50,153 --> 00:17:51,714
and three quarters of a million dollars.
375
00:17:51,714 --> 00:17:54,576
The mental math for that is very quickly
changing,
376
00:17:54,576 --> 00:17:56,658
where, you know, it's easy to have a
conversation about,
377
00:17:56,658 --> 00:17:59,750
well, you can replace this with a whole
call it operating system
378
00:17:59,750 --> 00:18:01,922
for your company that does a lot of your
process for you.
379
00:18:02,316 --> 00:18:05,849
Has actual features that that help you do
your work instead of updating
380
00:18:05,849 --> 00:18:07,811
19 formulas a week to make it work again.
381
00:18:07,811 --> 00:18:12,094
and that mental math is becoming easy
because what used to be 18 months
382
00:18:12,094 --> 00:18:15,397
and five developers is four months and two
developers.
383
00:18:15,397 --> 00:18:18,640
So it's I encourage everyone,
384
00:18:15,617 --> 00:18:16,617
Mm-hmm.
385
00:18:18,640 --> 00:18:23,924
maybe as a something more actionable than
pontificating ~ to folks,
386
00:18:23,924 --> 00:18:27,598
is take an Excel spreadsheet that your
business runs on,
387
00:18:27,598 --> 00:18:30,110
throw it into Claud or Chat GPT and say,
388
00:18:30,464 --> 00:18:33,076
Is this a candidate for a product that
would help us internally?
389
00:18:33,076 --> 00:18:35,438
If so, what would that look like?
390
00:18:33,549 --> 00:18:34,549
Mm-hmm.
391
00:18:35,438 --> 00:18:37,189
What features would it have? How would we
use it?
392
00:18:37,189 --> 00:18:39,220
Like spending that 20 minutes doing that,
393
00:18:39,220 --> 00:18:43,013
I've encouraged many folks to do that as a
means to like their early days,
394
00:18:43,013 --> 00:18:44,884
they're not ready to work with us for an
example.
395
00:18:44,884 --> 00:18:46,065
And I say, just try this then.
396
00:18:46,065 --> 00:18:50,028
And they come back and go, we didn't
realize like we are running this business
397
00:18:50,028 --> 00:18:52,800
so inefficiently because it's four
spreadsheets duct taped together,
398
00:18:52,800 --> 00:18:55,141
and we could have one system that does a
lot of this for us.
399
00:18:55,141 --> 00:18:58,594
So as an actionable thing, just try that
with Claud or Chat GPT one day.
400
00:18:59,169 --> 00:19:01,290
Yeah. I love it. That's a great
suggestion.
401
00:19:01,290 --> 00:19:05,621
So let's talk about the other side ~ of
the the equation,
402
00:19:05,621 --> 00:19:07,372
the the big enterprise platforms,
403
00:19:07,372 --> 00:19:10,092
right? You you kind of touched on that a
little bit and talked about SaaS.
404
00:19:10,092 --> 00:19:13,514
When you've got an overstructured shop
that's using one of these big
405
00:19:13,574 --> 00:19:16,775
~ ERP systems or one of these big SaaS
platform systems,
406
00:19:16,775 --> 00:19:22,657
is that ~ difficult in a different way for
you to work with the clients?
407
00:19:22,657 --> 00:19:23,657
What do you see there?
408
00:19:24,750 --> 00:19:29,172
~ it's it's better in that the data is
usually more structured and you
409
00:19:29,172 --> 00:19:32,994
can you can get to AI faster because
you're not pointing it at a bunch
410
00:19:31,938 --> 00:19:32,938
Mm.
411
00:19:32,994 --> 00:19:35,415
of people's C drives that you're hoping
they updated something that
412
00:19:35,415 --> 00:19:39,117
day on their computer. So in that way it's
better on the data side,
413
00:19:36,341 --> 00:19:37,341
Ha.
414
00:19:39,117 --> 00:19:41,599
where the drawback happens is the process.
415
00:19:41,599 --> 00:19:44,920
So the flexibility of something like Excel
allows you to do kind
416
00:19:42,188 --> 00:19:43,188
Mm-hmm.
417
00:19:44,920 --> 00:19:46,531
of whatever you want with it. On the flip
side,
418
00:19:46,531 --> 00:19:48,102
these giant organizations
419
00:19:48,781 --> 00:19:51,192
They need to make sure that their software
works for everyone.
420
00:19:51,192 --> 00:19:53,212
So if you say, Can we have this feature?
421
00:19:53,212 --> 00:19:56,033
Can we change how that works? They're
gonna usually the answer is yeah,
422
00:19:56,033 --> 00:19:57,733
we'll put on the roadmap. That's a good
idea.
423
00:19:57,733 --> 00:19:59,174
And you never hear from them again.
424
00:19:59,174 --> 00:20:03,465
So the rigidity of those platforms is
inherent in the business model
425
00:19:59,459 --> 00:20:00,459
Mm-hmm.
426
00:20:03,465 --> 00:20:05,755
of their trying to appeal to as many folks
as possible.
427
00:20:05,755 --> 00:20:09,486
The drawback for each business that uses
it is if it doesn't work for you,
428
00:20:09,486 --> 00:20:10,877
there's not much more you can do with it.
429
00:20:10,877 --> 00:20:12,977
And so that's where people are again,
430
00:20:12,977 --> 00:20:15,678
kind of that custom stuff. If you think
about the middle ground of
431
00:20:16,320 --> 00:20:21,915
it built for you off the shelf perfectly
or an Excel model that you've jerry
rigged.
432
00:20:21,915 --> 00:20:24,737
~ that if that custom software is
somewhere in the middle,
433
00:20:22,389 --> 00:20:23,389
Yeah.
434
00:20:24,737 --> 00:20:27,289
then that tends to also be the answer to
that.
435
00:20:27,289 --> 00:20:30,422
Cause again, it's becoming clear napkin
math to say,
436
00:20:30,422 --> 00:20:34,084
well, we're gonna pay Salesforce 60 grand
a year,
437
00:20:34,084 --> 00:20:38,178
or we can build something custom once for
a hundred grand and have
438
00:20:38,178 --> 00:20:41,150
no more ongoing costs other than a small
amount of hosting,
439
00:20:41,150 --> 00:20:45,194
which is usually a hell of a lot less than
than ~ seat-based costs.
440
00:20:45,194 --> 00:20:46,194
So
441
00:20:46,068 --> 00:20:47,549
It depends on what you need as a business.
442
00:20:47,549 --> 00:20:51,681
The the one other thing I would add is a
lot of SaaS companies
443
00:20:51,681 --> 00:20:54,883
are having that reckoning of do we let
others use our data?
444
00:20:54,883 --> 00:21:00,037
As in, you know, everything in a business
doesn't run out of one piece of software.
445
00:21:00,037 --> 00:21:01,357
It usually runs out of five, ten,
446
00:21:01,357 --> 00:21:05,140
fifteen pieces of software. ~ and for
every one piece of software that
447
00:21:05,140 --> 00:21:08,052
is incredibly flexible and lets you plug
that piece of software into the rest
448
00:21:08,052 --> 00:21:09,472
of the ones that your business uses,
449
00:21:09,472 --> 00:21:11,033
there's always five that are saying,
450
00:21:11,033 --> 00:21:12,033
no, that's
451
00:21:12,216 --> 00:21:14,417
we're gonna charge an arm and a leg for
API access,
452
00:21:14,417 --> 00:21:17,239
or we're just not gonna let you because
their goal is to keep you
453
00:21:17,239 --> 00:21:18,809
in their ecosystem and keep giving them
money.
454
00:21:18,809 --> 00:21:22,531
But that I think ultimately is going to
work against them because people need that
455
00:21:22,531 --> 00:21:26,383
flexibility. There there isn't one magical
piece of software that every business
456
00:21:26,383 --> 00:21:27,674
can run on. So
457
00:21:30,253 --> 00:21:33,876
So you've raised a point there that I I've
seen in a variety
458
00:21:33,876 --> 00:21:36,908
of different ways because you you talked
about kind of the multiple systems
459
00:21:36,908 --> 00:21:39,240
and you you talked about a number of
different things there.
460
00:21:39,240 --> 00:21:43,744
And what I often find when I go into
medium and and larger organizations
461
00:21:43,784 --> 00:21:46,466
is they love to say that we have a source
of truth,
462
00:21:46,466 --> 00:21:51,670
right? ~ and what I find personally
typically is that maybe
463
00:21:51,931 --> 00:21:55,453
one department or one functionality has a
source of truth,
464
00:21:55,453 --> 00:21:57,315
but there's usually a dozen source of
truths.
465
00:21:57,315 --> 00:21:59,807
~ what do you find when you walk in?
466
00:22:01,068 --> 00:22:05,161
Yeah, it goes back to that earlier ~ point
that we were chatting about,
467
00:22:05,161 --> 00:22:09,285
you know, the the professional service
business that has a lead that turns into
468
00:22:09,285 --> 00:22:11,767
a customer that gets a something fulfilled
for them.
469
00:22:11,767 --> 00:22:14,603
There's usually a bunch of spreadsheets,
470
00:22:14,603 --> 00:22:15,900
a bunch of tools where everyone says,
471
00:22:15,900 --> 00:22:17,441
I know exactly who that customer is.
472
00:22:17,441 --> 00:22:21,905
But if you were to zoom out, view it as
plane altitudes at the ground foot level,
473
00:22:21,905 --> 00:22:23,456
yes, everyone has their own source of
truth.
474
00:22:23,456 --> 00:22:25,478
But if you were at 30,000 foot view as a
plane,
475
00:22:25,478 --> 00:22:29,071
you would see nine different neighborhoods
with vastly different architecture
476
00:22:29,071 --> 00:22:30,071
of houses and
477
00:22:30,400 --> 00:22:31,950
No connecting roads to each other.
478
00:22:31,950 --> 00:22:34,261
I'm really into analogies today,
479
00:22:32,236 --> 00:22:33,236
Okay.
480
00:22:34,261 --> 00:22:35,261
sorry.
481
00:22:34,891 --> 00:22:40,835
I love it. We're gonna have to retain an
illustrator to to insert some cartoons
into
482
00:22:37,492 --> 00:22:43,934
But that's kind of sold.
483
00:22:40,835 --> 00:22:43,541
this episode's videos. So yeah.
484
00:22:43,934 --> 00:22:46,785
but that's kind of the the way that most
businesses operate.
485
00:22:46,785 --> 00:22:51,136
And so the the thing that execs have to
reckon with is what actually
486
00:22:51,136 --> 00:22:53,357
is our source of truth and who do we give
the keys to?
487
00:22:53,357 --> 00:22:57,278
As in, is does this become a ~ engineering
or IT function?
488
00:22:57,278 --> 00:22:58,278
Does this become
489
00:22:58,560 --> 00:23:01,492
one department to rule them all to use
your your analogy from earlier.
490
00:23:01,492 --> 00:23:03,153
Like it's a hard decision to make,
491
00:23:03,153 --> 00:23:06,845
but in the absence of that, you go back to
Wild West of well,
492
00:23:06,845 --> 00:23:09,287
there is no decision made, therefore
everyone does it their own way
493
00:23:09,287 --> 00:23:10,668
and it's a hodgepodge.
494
00:23:12,829 --> 00:23:15,531
So let's talk about the Wild West a little
bit more.
495
00:23:15,531 --> 00:23:19,014
~ and let's change the names to protect
the guilty.
496
00:23:19,014 --> 00:23:23,168
~ if you reflect back on companies that
you've been working with over
497
00:23:23,168 --> 00:23:24,879
the last couple of years in this space,
498
00:23:24,879 --> 00:23:29,132
think about a company where you walked in
and the data was a genuine mess.
499
00:23:29,132 --> 00:23:32,965
It was absolutely, you know, ~ it was just
it was horrible,
500
00:23:32,965 --> 00:23:35,047
as bad as it could get, because I'm sure
you've been there.
501
00:23:35,047 --> 00:23:36,929
What did you find? ~
502
00:23:37,409 --> 00:23:41,804
What did you do? What actually changed for
the company ~ after you were done with
503
00:23:41,804 --> 00:23:43,676
them? And again, we don't need names.
504
00:23:43,676 --> 00:23:45,137
We'll we'll protect the guilty here.
505
00:23:45,137 --> 00:23:48,331
~ but yeah, kinda kinda channel that
experience for us.
506
00:23:45,387 --> 00:23:46,387
Yeah.
507
00:23:49,187 --> 00:23:51,359
Yeah, we one that as you're asking that,
508
00:23:51,359 --> 00:23:56,543
one that came to mind right away was we
had a a team where they provide services
509
00:23:56,543 --> 00:23:59,855
for someone and they they grew very
organically.
510
00:23:59,855 --> 00:24:02,407
And so as is the case with most
businesses,
511
00:24:02,407 --> 00:24:05,690
you you keep an eye as a leader on the
things that are breaking and
512
00:24:05,690 --> 00:24:07,434
the things that seem like they're running
along smoothly.
513
00:24:07,434 --> 00:24:09,303
You're just like, Great, I won't pay
attention to it.
514
00:24:09,303 --> 00:24:12,375
The thing they didn't pay attention to is
because they grew organically in
515
00:24:12,375 --> 00:24:18,250
a in a great way overall, was they hired
all these account managers and they had
516
00:24:18,782 --> 00:24:21,723
rate card as a business, as in whatever
service is being provided,
517
00:24:21,723 --> 00:24:23,203
we're gonna quote them our rate.
518
00:24:23,203 --> 00:24:28,305
The thing that they never stopped to take
a look at was what are our rates?
519
00:24:28,305 --> 00:24:30,846
Are they consistent? And when we came in,
520
00:24:30,846 --> 00:24:32,866
they said we want to automate our quoting,
521
00:24:32,866 --> 00:24:35,627
like our kind of cash to quote cycle.
522
00:24:35,627 --> 00:24:37,948
We want to automate the quoting and
proposal creation because
523
00:24:37,948 --> 00:24:40,729
our account managers spend endless hours a
week kind of creating all
524
00:24:40,729 --> 00:24:42,609
of these for every every request that
comes in.
525
00:24:42,609 --> 00:24:44,940
So we said great, let's take a look at the
rate cards.
526
00:24:45,976 --> 00:24:47,887
Their rate cards were spread across,
527
00:24:47,887 --> 00:24:50,830
I think it was nine spreadsheets and were
the equivalent when we brought
528
00:24:50,830 --> 00:24:54,332
it all into one was 14,000 rows in one
spreadsheet.
529
00:24:54,332 --> 00:24:56,274
And this is like a mid-market business.
530
00:24:56,274 --> 00:25:01,337
This isn't, you know, a a manufacturing
business with 91 types of sockets
531
00:25:01,337 --> 00:25:05,540
and 51 types of screws. Like they had
about 10 services,
532
00:25:05,540 --> 00:25:08,813
but somehow their rate card was the
equivalent of 14,000 rows.
533
00:25:08,813 --> 00:25:11,405
And that's because the account management
team had no guardrails
534
00:25:11,425 --> 00:25:13,146
and they grew organically. So they just
said,
535
00:25:13,612 --> 00:25:16,363
I'll copy so-and-so spreadsheet and make
it my own for my customers.
536
00:25:16,363 --> 00:25:18,093
And then they copied and et cetera,
537
00:25:18,093 --> 00:25:20,044
et cetera. So, as you can probably
imagine,
538
00:25:20,044 --> 00:25:22,895
if a customer comes in and picks one of 10
services and says,
539
00:25:22,895 --> 00:25:25,506
I want to automate the quoting and
proposal part,
540
00:25:25,506 --> 00:25:29,607
well, which rate do we pick when there's
only 10 services but 14,000 rows of rates?
541
00:25:29,607 --> 00:25:35,809
And so that's just one example of that ~
absence of guardrails.
542
00:25:35,809 --> 00:25:38,860
If you view it as as bowling, there's
there was no bumpers on the bowling,
543
00:25:36,215 --> 00:25:37,215
Yeah.
544
00:25:38,860 --> 00:25:40,290
and every time it could go into the
gutter.
545
00:25:40,706 --> 00:25:42,529
How can you have automation AI say,
546
00:25:42,529 --> 00:25:46,036
well, this is what we're going to charge
them when there's 900 variations
547
00:25:46,036 --> 00:25:50,323
of one service kind of thing? So that's
just one extreme example of
548
00:25:50,324 --> 00:25:53,920
10 services does not equate to 14,000 rows
for rate sheet.
549
00:25:54,638 --> 00:26:00,555
Indeed. What was the timeline on when they
started to see I I hate
550
00:26:00,555 --> 00:26:03,828
the re the phrase ROI, but I I don't know
what else to use.
551
00:26:03,828 --> 00:26:07,763
So how long did it take them to see a
return on that investment and that work?
552
00:26:08,886 --> 00:26:12,087
Well, we spent about six months together
building all of that because we
553
00:26:12,087 --> 00:26:16,208
had to to your earlier point of the the
rungs of the transformation pyramid,
554
00:26:16,208 --> 00:26:21,560
so to speak, they were kind of forced into
this transformation because
555
00:26:21,560 --> 00:26:24,231
the way that they they did it couldn't
sustain itself.
556
00:26:24,231 --> 00:26:28,302
Like there was there was no process and
there was a 14,000 row rate card,
557
00:26:28,302 --> 00:26:31,563
et cetera. And so for them it was about
six months of both
558
00:26:31,563 --> 00:26:34,824
the business process mapping, and by that
I mean getting the right people
559
00:26:34,824 --> 00:26:35,824
in the room to say,
560
00:26:36,642 --> 00:26:38,784
From step one to step 10, how does it
work?
561
00:26:38,784 --> 00:26:40,785
What are all the edge cases? Where does it
go wrong?
562
00:26:40,785 --> 00:26:43,978
What do we need to factor in? And half of
our job is just to get people
563
00:26:43,978 --> 00:26:45,549
in the room together to say, how does this
work?
564
00:26:45,549 --> 00:26:48,922
Because and why I say that's half the job
is usually getting them into
565
00:26:48,922 --> 00:26:50,743
the room and getting them to talk about
their own process.
566
00:26:50,743 --> 00:26:51,828
They realize for themselves, like,
567
00:26:51,828 --> 00:26:55,977
how did we let this happen? Why does a 40
step process need to be
568
00:26:55,977 --> 00:26:58,730
40 steps when logically, now that we're
playing it out loud,
569
00:26:58,730 --> 00:27:00,691
it actually could be 20. And of those 20,
570
00:27:00,691 --> 00:27:01,782
10 of them could be automated.
571
00:27:02,284 --> 00:27:05,196
So that ROI to answer your question was
kind of that six months of
572
00:27:05,196 --> 00:27:07,608
us doing the mapping, getting all the data
figured out.
573
00:27:07,608 --> 00:27:11,582
Because the rate card example was one of
four or five examples where how
574
00:27:11,582 --> 00:27:15,164
did it get this way? But that allowed us
to build a consistent system
575
00:27:15,605 --> 00:27:19,428
and layer in automation and AI into it so
that at that six,
576
00:27:19,428 --> 00:27:21,650
seven month mark, we could start to launch
it to their team.
577
00:27:21,650 --> 00:27:23,711
And ROI came within, you know,
578
00:27:23,711 --> 00:27:26,594
a few quarters because they just save so
much time.
579
00:27:26,594 --> 00:27:31,928
The way that I look at ROI of a lot of AI
is I've I've yet to see any revenue.
580
00:27:32,131 --> 00:27:35,654
generating AI be successful, as I'm sure
some people listening have tried
581
00:27:35,654 --> 00:27:37,936
the AI SDR and the AI salesperson,
582
00:27:37,936 --> 00:27:41,538
like they're all my opinion, largely
terrible.
583
00:27:41,538 --> 00:27:46,643
but the operating side of the business is
where you get the the benefit.
584
00:27:46,643 --> 00:27:48,604
And so a lot of times the ROI,
585
00:27:48,604 --> 00:27:51,697
the way that I think about it with execs
and they come to to learn is a good
586
00:27:51,697 --> 00:27:52,697
way to think about it is like,
587
00:27:52,567 --> 00:27:55,430
how much could we grow and not have to
raise headcount?
588
00:27:55,430 --> 00:27:56,971
As in if our margin
589
00:27:57,493 --> 00:27:59,695
at ten million is, you know, ten percent.
590
00:27:59,695 --> 00:28:02,117
At twenty million, it could actually be
twenty five percent,
591
00:28:02,117 --> 00:28:06,100
not just twenty percent. ~ or whatever the
met the math is for their business.
592
00:28:06,100 --> 00:28:10,824
But the time that you get back tends to
equate to those people have more time
593
00:28:10,824 --> 00:28:13,116
for more important things. And if we
doubled revenue,
594
00:28:13,116 --> 00:28:15,318
doesn't mean that we have to double head
count accordingly.
595
00:28:18,157 --> 00:28:21,298
I think one of the themes that emerges in
a lot of the conversations I've
596
00:28:21,298 --> 00:28:25,530
had with folks in spaces ~ like yours or
adjacent to yours that
597
00:28:25,530 --> 00:28:32,023
are working with clients around AI is that
AI no longer allows
598
00:28:32,023 --> 00:28:35,985
you to hide the fact that your digital
change management initiative failed.
599
00:28:35,985 --> 00:28:39,896
It doesn't allow you to hide the fact that
your processes aren't documented
600
00:28:39,896 --> 00:28:43,968
and that they're a mess. It doesn't allow
you to hide because your data
601
00:28:43,988 --> 00:28:45,909
is in fifty thousand different places.
602
00:28:45,909 --> 00:28:46,909
And
603
00:28:46,507 --> 00:28:51,247
That that's really what came through for
me just now ~ in your anecdote again,
604
00:28:51,247 --> 00:28:57,129
is that we can't apply this technology
unless those other things
605
00:28:57,129 --> 00:28:58,129
get addressed first.
606
00:29:00,352 --> 00:29:02,373
Yeah, and that's that's the painful part.
607
00:29:02,373 --> 00:29:07,026
I think that's why when people say AI is
gonna take everyone's jobs
608
00:29:07,026 --> 00:29:09,297
and it's gonna change every company over
the next eighteen months,
609
00:29:09,297 --> 00:29:13,579
I roll my eyes at this point because human
nature and the history
610
00:29:13,579 --> 00:29:17,532
of technological evolution ~ indicates
that is never going to be the case.
611
00:29:17,532 --> 00:29:22,194
Yes, the technology is is progressing
extremely quickly and I still have
612
00:29:22,194 --> 00:29:24,516
jaw drop moments, you know, every month at
this point.
613
00:29:24,516 --> 00:29:26,697
It used to used to feel like every year
and then every quarter,
614
00:29:26,207 --> 00:29:28,053
Yeah. Yeah.
615
00:29:26,697 --> 00:29:28,478
and now it's every month. ~
616
00:29:28,866 --> 00:29:31,727
But that doesn't change the fact that you
could have all the advancement
617
00:29:31,727 --> 00:29:34,808
in the world tomorrow, but it takes a long
time for humans to
618
00:29:34,808 --> 00:29:37,038
be comfortable with change, to be willing
to change.
619
00:29:37,038 --> 00:29:38,839
And that is where most of our time is
spent.
620
00:29:38,839 --> 00:29:44,370
Like the I think McKinsey has the stat of
like something like 65 or
621
00:29:44,370 --> 00:29:46,731
70% of digital transformation projects
fail.
622
00:29:46,731 --> 00:29:50,602
What they fail to include in that is about
90% of that is in
623
00:29:50,602 --> 00:29:53,953
the change management stage. It you can
build the coolest thing in the world,
624
00:29:53,953 --> 00:29:55,243
but if people don't understand it,
625
00:29:55,243 --> 00:29:57,804
both the benefit and the why of it and the
how and the what.
626
00:29:58,361 --> 00:30:01,973
They're not going adopt it. So as as I
think the AI,
627
00:30:01,973 --> 00:30:06,235
the tech bros have come to call it of the
technological overhang,
628
00:30:06,235 --> 00:30:09,797
as in the overhang of like the technology
has progressed a lot more than people
629
00:30:09,797 --> 00:30:12,708
realize. But the ability for humans,
630
00:30:12,708 --> 00:30:15,479
which companies are just made up of humans
working together,
631
00:30:15,479 --> 00:30:18,511
the ability of companies and therefore the
people inside of them to change
632
00:30:18,511 --> 00:30:20,352
is is going to be a lot slower than people
think.
633
00:30:22,089 --> 00:30:26,710
I I've spoken with a number of change
management experts on ~ this show before
634
00:30:26,710 --> 00:30:30,781
and ~ worked with a number of others on ~
overlapping engagements
635
00:30:30,781 --> 00:30:36,033
and in with clients. And the the question
I keep coming back to,
636
00:30:36,033 --> 00:30:38,464
which you you just kind of teed up very
nicely,
637
00:30:38,464 --> 00:30:41,774
and I I don't expect an answer for this
because it's a little bit more esoteric,
638
00:30:41,774 --> 00:30:47,766
but I wonder whether the companies that
are born first and are AI first
639
00:30:47,766 --> 00:30:49,477
and are new to the market now,
640
00:30:49,477 --> 00:30:50,997
you touched on some of those earlier.
641
00:30:51,445 --> 00:30:55,227
I wonder whether they are going to outrun
all of these other organizations because
642
00:30:55,387 --> 00:30:58,808
you hit the nail on the head when you
talked about people being the problem here
643
00:30:58,808 --> 00:31:01,419
and the change management ~ component
being a problem.
644
00:31:01,419 --> 00:31:06,222
And I wonder can existing organizations
change enough and change fast enough
645
00:31:06,222 --> 00:31:09,323
to outrun the competitors that are already
spinning up right now?
646
00:31:11,000 --> 00:31:13,871
Yeah, I think it goes back to that's why I
call it a countdown.
647
00:31:13,871 --> 00:31:18,693
Like it's it's unavoidable because we're
already seeing it with even in our space,
648
00:31:18,693 --> 00:31:21,654
like we're ~ not to talk about our our
shop.
649
00:31:21,654 --> 00:31:23,035
I'm not here to pitch what we do,
650
00:31:23,035 --> 00:31:26,503
but it it still blows my mind when we
share our prices and timelines
651
00:31:26,503 --> 00:31:28,977
and there are people like, there's no way
you can do it at a good quality
652
00:31:29,198 --> 00:31:30,314
for that price and that timeline.
653
00:31:30,314 --> 00:31:34,080
I'm like, watch. and it's because Yeah,
654
00:31:32,558 --> 00:31:34,823
Hold my beer. Yeah.
655
00:31:34,080 --> 00:31:36,901
yeah. And it's because we're we're not,
656
00:31:36,901 --> 00:31:39,842
you know, carrying the baggage of a
Deloitte McKinsey Accenture.
657
00:31:40,887 --> 00:31:44,150
you know, half the half of people's week
is spent in politics and a meeting about
658
00:31:44,150 --> 00:31:45,871
a meeting about an email from six weeks
ago.
659
00:31:45,871 --> 00:31:50,054
We are building everything as kind of AI
first and keeping humans in the loop.
660
00:31:50,054 --> 00:31:55,218
But by nature of that, we're out competing
those businesses and winning work because
661
00:31:55,218 --> 00:31:57,950
of that. And so I use that as a very
personal anecdote,
662
00:31:57,950 --> 00:32:00,052
but I think we're seeing it in many
spaces.
663
00:32:00,052 --> 00:32:04,115
I think tutoring, if people look at kind
of the tutoring and education space,
664
00:32:04,115 --> 00:32:07,728
there's a lot of of startups coming out of
like Y Combinator that
665
00:32:07,728 --> 00:32:09,970
are surpassing the revenue at
666
00:32:10,062 --> 00:32:13,413
25, 30 people of businesses that have been
around that they're competing with who've
667
00:32:13,413 --> 00:32:17,304
been around for 20 years. And so it goes
back to that countdown of it's
668
00:32:17,304 --> 00:32:19,154
not to fearmonger, it's more of yes,
669
00:32:19,154 --> 00:32:24,046
you can avoid making changes and
transforming and and preparing and diving
into this
670
00:32:24,046 --> 00:32:25,906
era. And right now that's okay.
671
00:32:25,906 --> 00:32:26,907
But a couple of years from now,
672
00:32:26,907 --> 00:32:30,527
when more and more to your point companies
start and are born out of this era,
673
00:32:30,527 --> 00:32:33,668
they inherently have the advantage of they
don't have the baggage of,
674
00:32:33,668 --> 00:32:37,179
you know, so and so who's been here for
twenty years and they've opted
675
00:32:37,179 --> 00:32:39,630
to run their department off a spreadsheet
that only they know in the
676
00:32:39,832 --> 00:32:42,945
Ten other people on their team have no
idea how it works is very different than
677
00:32:43,325 --> 00:32:47,529
someone who said, I get multiplayer AI and
skills and harness this with AI,
678
00:32:47,529 --> 00:32:49,651
and we built our company around that from
day one.
679
00:32:49,651 --> 00:32:51,572
They just inherently cannot compete you.
680
00:32:51,572 --> 00:32:53,122
Yeah.
681
00:32:52,023 --> 00:32:53,274
Yeah, agreed.
682
00:32:54,455 --> 00:32:56,917
Since we're touching on the issue of
change here,
683
00:32:56,917 --> 00:33:00,600
~ I hear the phrase bandied about a lot,
684
00:33:00,600 --> 00:33:06,135
AI ready. And I'm sure you've walked into
potential client meetings before
685
00:33:06,135 --> 00:33:11,069
and heard that phrase. What is the biggest
misconception that execs have about being
686
00:33:11,130 --> 00:33:14,973
AI ready and and where do these smart
leaders get this wrong?
687
00:33:17,752 --> 00:33:20,643
Think the thing I've had to reckon with
and and get over my own
688
00:33:20,643 --> 00:33:23,494
~ resentments or chips on my shoulder,
689
00:33:23,494 --> 00:33:29,895
if I'm being honest, is the I had the
benefit of working only with tech startups
690
00:33:29,895 --> 00:33:33,266
and and call it Silicon Valley companies
for nearly a decade.
691
00:33:33,266 --> 00:33:39,128
And when you say API and you say agile and
you say like all of these terms,
692
00:33:39,128 --> 00:33:44,909
that's just common language. I was it was
a rude awakening working with everyone
693
00:33:45,089 --> 00:33:46,930
now, working with every company.
694
00:33:47,785 --> 00:33:51,177
or every company we work with is not in
Silicon Valley and doesn't come from tech.
695
00:33:51,177 --> 00:33:53,278
And so there was a bit of a chip on my
shoulder of like,
696
00:33:53,278 --> 00:33:56,370
do you not get this? Like, do you not get
why you should have data
697
00:33:56,370 --> 00:33:57,780
and tools connected and whatnot?
698
00:33:57,780 --> 00:34:01,172
and so that is kind of like the change
part,
699
00:34:01,172 --> 00:34:04,444
the AI ready part is I really encourage
leaders.
700
00:34:04,444 --> 00:34:06,215
My my resentment is gone. I get it.
701
00:34:06,215 --> 00:34:07,565
Everyone is running their own business.
702
00:34:07,565 --> 00:34:11,527
You know, we've got founders who who know
these niche businesses better than anyone.
703
00:34:11,527 --> 00:34:14,918
And someone from Silicon Valley thinks
they can come in and take over.
704
00:34:14,918 --> 00:34:16,590
It's like, no, you don't get the nuance
and
705
00:34:16,992 --> 00:34:19,783
and requirements or compliance needs or
whatever it might be.
706
00:34:19,783 --> 00:34:23,495
So there's many folks who they have every
right to to be successful
707
00:34:23,495 --> 00:34:27,166
in their business. What shocks me is their
level of technical understanding
708
00:34:27,166 --> 00:34:31,208
is so low that the change and AI
readiness,
709
00:34:28,397 --> 00:34:29,397
Mm.
710
00:34:31,208 --> 00:34:32,608
they don't even know what that looks like.
711
00:34:32,608 --> 00:34:36,320
So all that is to say, you know,
712
00:34:36,320 --> 00:34:42,282
if if the peak of technical performance in
your company is someone that
713
00:34:42,282 --> 00:34:45,934
can build a pivot table, then you probably
need some help on getting AI ready.
714
00:34:46,550 --> 00:34:51,091
~ and if the peak in your your
organization is we have a CRM
715
00:34:51,091 --> 00:34:53,312
and that's everything else is run out of
spreadsheets,
716
00:34:53,312 --> 00:34:57,803
you probably need to be AI ready
culturally more than you do technically.
717
00:34:57,803 --> 00:35:00,214
Because the technical part is now pretty
easy.
718
00:35:00,214 --> 00:35:03,285
We we're building with tools that have
been around for five,
719
00:35:03,285 --> 00:35:06,396
ten, fifteen, twenty years. It's the
people side of it.
720
00:35:06,396 --> 00:35:09,607
So ~ th this is a bit of a I'll get off my
soapbox,
721
00:35:09,607 --> 00:35:12,778
but but the AI readiness
722
00:35:10,315 --> 00:35:13,184
No no, keep going. Let's put another
soapbox on top of it.
723
00:35:14,488 --> 00:35:16,559
The AI readiness part, if anything,
724
00:35:16,559 --> 00:35:19,171
that I encourage leaders to do is start
playing around with this stuff.
725
00:35:19,171 --> 00:35:23,264
Start understanding it for yourselves and
not just taking the the LinkedIn,
726
00:35:23,264 --> 00:35:27,407
you know, slop that is everything changed
today and AI is here.
727
00:35:27,407 --> 00:35:30,569
There is no AI silver bullet. There is no
one AI tool.
728
00:35:30,569 --> 00:35:33,431
There is no one AI model that will solve
all of this for you.
729
00:35:33,431 --> 00:35:37,184
Being AI ready is more about a cultural
readiness for change of how
730
00:35:37,184 --> 00:35:41,177
you do business and how you operate than
it is a technical adoption.
731
00:35:41,177 --> 00:35:43,118
The technical adoption is part of it.
732
00:35:43,512 --> 00:35:45,754
But it's always gonna fail if you're not
ready culturally.
733
00:35:45,754 --> 00:35:47,102
And that starts at the top.
734
00:35:48,770 --> 00:35:52,162
And I I love those two succinct litmus
tests that you threw out there
735
00:35:52,162 --> 00:35:53,653
in terms of how did you phrase it?
736
00:35:53,653 --> 00:35:56,763
If the peak of your technical prowess in
your company is somebody who
737
00:35:56,763 --> 00:35:58,775
can build a pivot table, that's a problem,
738
00:35:58,775 --> 00:36:01,937
right? So not AI ready. ~
739
00:36:00,170 --> 00:36:02,091
We ask folks. We've
740
00:36:02,091 --> 00:36:05,364
learned to ask folks when we're getting
into a company that says we don't have
741
00:36:05,364 --> 00:36:09,096
a strategy and we're hiring you to help us
figure out what is our even our strategy.
742
00:36:09,096 --> 00:36:11,458
We send out a feeler of, you know,
743
00:36:11,458 --> 00:36:13,519
who are the who are the most technically
proficient.
744
00:36:13,519 --> 00:36:15,721
And it's it's always wonderful when
they're like,
745
00:36:15,721 --> 00:36:18,153
we have an engineering team and chat with
our VP of engineering and
746
00:36:18,153 --> 00:36:20,924
we hop right into a great conversation
with them or a CTO type.
747
00:36:20,924 --> 00:36:24,287
But most mid-market businesses don't have
that kind of team.
748
00:36:24,287 --> 00:36:25,998
And so it's not a knock on them.
749
00:36:25,998 --> 00:36:27,759
It's not none of this is meant as a slam.
750
00:36:27,759 --> 00:36:28,759
It's more so
751
00:36:29,248 --> 00:36:32,329
We we ask to get introduced to those folks
so we understand where they're
752
00:36:32,329 --> 00:36:34,980
at and and we quickly learn, you know,
753
00:36:34,980 --> 00:36:36,900
yeah, you should chat with with so and so.
754
00:36:36,900 --> 00:36:40,151
They they understand technology and we
meet with them and we say,
755
00:36:40,151 --> 00:36:41,732
We hear we hear you're the pro,
756
00:36:41,732 --> 00:36:42,742
we'd love to see what you've built.
757
00:36:42,742 --> 00:36:44,283
And they pull up an Excel sheet and
they're like,
758
00:36:44,283 --> 00:36:46,603
I can use macros in Excel. And we're like,
759
00:36:46,603 --> 00:36:48,364
That's great. That's not a bad thing.
760
00:36:48,364 --> 00:36:53,075
But if that's your kind of pinnacle of
technical ability in the organization,
761
00:36:53,075 --> 00:36:55,746
you are you're operating in a different
era.
762
00:36:56,349 --> 00:36:59,510
Yeah. Well said. Well said.
763
00:36:59,510 --> 00:37:02,051
You touched on this before too.
764
00:37:02,051 --> 00:37:07,294
You you kind of talked about ~ some of the
pitfalls with ~ the classic strategy deck
765
00:37:07,294 --> 00:37:09,525
approach. I think you kind of channeled
the specter of some of
766
00:37:09,525 --> 00:37:13,956
the big consulting companies. ~ and the
fact that ~ instead of you know,
767
00:37:13,956 --> 00:37:15,417
a hundred and twenty-page report,
768
00:37:15,417 --> 00:37:17,718
you guys build something that people can
see.
769
00:37:17,718 --> 00:37:24,449
And I'm interested in your perspective on
why seeing it changes the conversation and
770
00:37:24,875 --> 00:37:29,749
What should execs insist on being shown
before they commit a real budget instead
771
00:37:29,749 --> 00:37:32,842
of a hundred twenty-page report or a
really fancy slide deck?
772
00:37:32,842 --> 00:37:35,674
So again, my two questions here are are
basically,
773
00:37:35,674 --> 00:37:37,776
why does seeing it change the convo?
774
00:37:37,776 --> 00:37:42,480
And what should an exec insist on seeing
before they ~ even authorize
775
00:37:42,480 --> 00:37:43,741
a single dollar of spend?
776
00:37:44,909 --> 00:37:49,123
Yeah, I I don't take for granted that I
was spoiled getting to work
777
00:37:49,123 --> 00:37:52,586
at a product design agency that got to
work with the coolest startups
778
00:37:52,586 --> 00:37:55,949
and and enterprise a hundred or ten even.
779
00:37:53,214 --> 00:37:54,214
huh.
780
00:37:55,949 --> 00:37:59,953
And it was always fascinating to watch
where you know we had
781
00:38:00,413 --> 00:38:04,327
we had product management, we had projects
being run really well,
782
00:38:04,327 --> 00:38:07,119
we had great UX research, we had great
engineers,
783
00:38:07,119 --> 00:38:10,202
but the thing where the execs
784
00:38:10,644 --> 00:38:12,265
Eyes lit up, the founder's eyes lit up,
785
00:38:12,265 --> 00:38:16,178
and everyone really crystallized around
what this thing is and how it
786
00:38:16,178 --> 00:38:18,960
is going to work and function was always
in the design phase.
787
00:38:18,960 --> 00:38:22,022
That was always the most fun phase because
there's a very
788
00:38:22,022 --> 00:38:24,944
big difference between when contracts
signed,
789
00:38:24,944 --> 00:38:27,345
spin up project management template,
790
00:38:27,345 --> 00:38:29,467
inform someone with a notification,
791
00:38:29,467 --> 00:38:31,278
do the checklist item for them,
792
00:38:31,278 --> 00:38:33,740
that kind of thing. There's a very big
difference between that
793
00:38:33,740 --> 00:38:37,682
and watching someone press a button and
seeing the template spin up
794
00:38:37,682 --> 00:38:39,073
and fill in automatically for you.
795
00:38:39,073 --> 00:38:40,073
And then a
796
00:38:40,366 --> 00:38:43,667
Teams channel spin up for you and ~ all
that kind of stuff that
797
00:38:42,677 --> 00:38:43,677
Mm.
798
00:38:43,667 --> 00:38:45,648
you know automation does on paper,
799
00:38:45,648 --> 00:38:49,100
you can explain it, but there's a very big
difference between this is what
800
00:38:49,100 --> 00:38:51,231
it would actually look like for you to be
using it tomorrow.
801
00:38:51,231 --> 00:38:53,732
And so that's why we take the prototype
first approach.
802
00:38:53,732 --> 00:38:58,514
~ and it again, we're we're spoiled in
this era being born out of the AI
803
00:38:58,514 --> 00:39:02,465
era of a prototype because we're not
building a complex backend and having
804
00:39:02,465 --> 00:39:05,697
to build all this infrastructure and worry
about security and whatnot,
805
00:39:05,697 --> 00:39:08,338
it's non it's non-functional and not using
real data.
806
00:39:08,758 --> 00:39:11,820
and and whatnot. So we can spin it up in a
week kind of thing.
807
00:39:11,820 --> 00:39:14,922
So a prototype really crystallizes the
idea.
808
00:39:14,922 --> 00:39:17,764
And so to your point, that's kind of the
what and why.
809
00:39:17,764 --> 00:39:22,878
to your question around like what execs
should ask for is a prototype.
810
00:39:22,878 --> 00:39:25,810
Like what will this actually look like if
we were to roll this out tomorrow
811
00:39:26,251 --> 00:39:28,162
in our organization as a system,
812
00:39:28,162 --> 00:39:29,543
whether that's something as small as,
813
00:39:29,543 --> 00:39:32,815
you know, call it a utility app where you
drop in something
814
00:39:32,815 --> 00:39:36,158
and it automates something for you all the
way to we want an OS
815
00:39:36,673 --> 00:39:38,654
orchestration layer system for our
business,
816
00:39:38,654 --> 00:39:41,046
asking for a prototype is I think where
people should start.
817
00:39:41,046 --> 00:39:46,320
And maybe I'm a bit biased, but having
been through an era where you you
818
00:39:46,320 --> 00:39:49,903
map out the workflow on paper and you're
in a Visio or Miro type app
819
00:39:49,903 --> 00:39:51,534
and kind of figuring out what it is,
820
00:39:51,534 --> 00:39:53,175
when you actually get to building it,
821
00:39:53,175 --> 00:39:56,257
there's always way more change than you
think there will be because it's
822
00:39:56,257 --> 00:39:58,879
not until people actually see how they're
going to use it that it clicks for them.
823
00:39:59,512 --> 00:40:04,984
Yeah. I am personally very guilty of using
jargon and driving by terms that
824
00:40:04,984 --> 00:40:06,525
I understand and my guests understand,
825
00:40:06,525 --> 00:40:08,846
but sometimes the ~ the listener does not.
826
00:40:08,846 --> 00:40:12,487
And one of the mu the things that I wanted
to ask you as a follow-up
827
00:40:12,687 --> 00:40:17,859
was ~ what's the smallest version of
something that a a resource constrained
company
828
00:40:17,859 --> 00:40:20,370
could try themselves? And you mentioned a
utility app.
829
00:40:20,370 --> 00:40:24,492
And I'm I'm thinking that that might be a
good way for us to kind of explain what
830
00:40:24,492 --> 00:40:26,743
is a utility app, number one, and then
number two.
831
00:40:27,105 --> 00:40:30,964
What's the smallest version of one of
those that's most cost effective that
you've
832
00:40:30,964 --> 00:40:33,782
seen that would demonstrate to a leader
the potential and
833
00:40:33,782 --> 00:40:35,775
the possibility that you're speaking to
right now?
834
00:40:36,918 --> 00:40:42,450
Yeah, if if you think if you think about
your business as Henry Ford
835
00:40:42,450 --> 00:40:48,322
era assembly line, ~ and there's call it
forty core steps
836
00:40:48,742 --> 00:40:52,443
or forty core process in your business
that's that kind of follow an assembly
line.
837
00:40:52,443 --> 00:40:55,994
Utility app is zooming in on three or four
of those steps
838
00:40:55,994 --> 00:40:57,565
and building something that would solve
for that.
839
00:40:57,565 --> 00:41:01,766
Mo again to your earlier point of why ~
why folks kind of
840
00:41:02,134 --> 00:41:05,025
Either are intimidated by or fail with
these kind of AI rollouts
841
00:41:05,025 --> 00:41:08,225
and deployments and transformations is
because they try to boil the ocean
842
00:41:08,225 --> 00:41:11,377
and they want to they want to solve for
all 40 steps of the assembly line.
843
00:41:11,377 --> 00:41:13,957
And so we refer to it as a utility app,
844
00:41:13,957 --> 00:41:16,528
as in there's a utility provided to an
area of the business,
845
00:41:16,528 --> 00:41:19,219
not orchestration of the entire business.
846
00:41:19,219 --> 00:41:24,900
And so a a good example is a lot of
businesses will get a request for a
proposal.
847
00:41:24,900 --> 00:41:30,362
And so they'll get an email saying we're
looking for XYZ on ABC timeline.
848
00:41:31,184 --> 00:41:36,049
whatever the inputs might be, how do you
build a utility that intakes that looks
849
00:41:36,049 --> 00:41:38,291
at your call it parameters internally,
850
00:41:38,291 --> 00:41:40,673
your matrix of services, cost,
851
00:41:40,673 --> 00:41:42,164
resourcing, whatever it might be,
852
00:41:42,164 --> 00:41:45,387
and automatically intake that and generate
a first pass of the proposal
853
00:41:45,387 --> 00:41:47,960
and quote for them. That is a good example
of a utility app.
854
00:41:47,960 --> 00:41:50,252
It's not solving for the entire sales
cycle.
855
00:41:50,252 --> 00:41:54,236
It's solving for the utility of that
proposal part that happens somewhere in,
856
00:41:54,236 --> 00:41:56,608
you know, those three or four steps in
that assembly line.
857
00:41:57,740 --> 00:42:02,412
That's a perfect example. And I think it's
a it's a good springboard into
858
00:42:02,592 --> 00:42:04,573
~ the decision making process.
859
00:42:04,573 --> 00:42:06,631
So in your your hypothetical there,
860
00:42:06,631 --> 00:42:08,854
you you you talked about a 40 step
process,
861
00:42:08,854 --> 00:42:13,276
right? Assuming that the data in an
organization is in good enough shape,
862
00:42:13,276 --> 00:42:15,327
right, with those 40 steps. Let's just
assume that.
863
00:42:15,327 --> 00:42:17,097
I know it's a big assumption, but let's
assume that.
864
00:42:17,097 --> 00:42:19,538
~ assuming the data is in good enough
shape,
865
00:42:19,538 --> 00:42:25,681
how do you decide where AI actually
belongs in a workflow like that versus
where
866
00:42:25,681 --> 00:42:26,681
it's just a gimmick?
867
00:42:26,920 --> 00:42:27,920
~ or hype.
868
00:42:29,036 --> 00:42:33,769
Yeah, the the thing I encourage folks to
think about first and foremost,
869
00:42:33,769 --> 00:42:37,991
if it's call it the Eisenhower matrix of
like cost or effort versus value,
870
00:42:37,991 --> 00:42:42,733
~ is what is the most repetitive and what
involves the most people having to do it.
871
00:42:42,733 --> 00:42:45,735
So we have folks sometimes will come to us
and they say,
872
00:42:45,735 --> 00:42:50,568
~ our CEO wants an AI automation that
summarizes his email and calendar
873
00:42:50,568 --> 00:42:51,638
for the week every Monday morning.
874
00:42:51,638 --> 00:42:55,651
It's like, okay, here's a YouTube video
and try out Zapier.
875
00:42:55,651 --> 00:42:57,852
~ that's great, that's valuable for them,
876
00:42:56,301 --> 00:42:57,301
Yeah. Huh.
877
00:42:57,852 --> 00:42:58,852
but like
878
00:42:58,412 --> 00:43:02,985
The effort versus value is really not
there versus you've got 10-person sales
team
879
00:43:03,165 --> 00:43:06,748
each spending, call it 15 hours a week
generating quotes and proposals for
people.
880
00:43:06,748 --> 00:43:11,591
That on that effort versus value matrix is
up and to the right probably pretty high.
881
00:43:11,591 --> 00:43:16,084
~ because especially when salespeople are
revenue driving and their their time
882
00:43:16,084 --> 00:43:18,926
is better spent closing business and
forming better relationships
883
00:43:18,926 --> 00:43:22,468
and following up with people. So that's
how I encourage folks to look
884
00:43:22,468 --> 00:43:24,159
at it is what's the effort versus value?
885
00:43:24,159 --> 00:43:27,962
If it's something that you do once a
quarter and one person does it for half a
day.
886
00:43:28,438 --> 00:43:29,438
if it's some sort of, you know,
887
00:43:29,390 --> 00:43:30,983
call it quarterly reconciliation,
888
00:43:30,983 --> 00:43:34,940
that's probably not worth it versus
something that five people are doing every
889
00:43:34,940 --> 00:43:37,595
day for five hours a day or five hours a
week even.
890
00:43:37,595 --> 00:43:40,972
That adds up pretty quickly. So it's kind
of that effort versus value calculation.
891
00:43:41,868 --> 00:43:44,720
I I love that we just channeled the
Eisenhower matrix in this interview.
892
00:43:44,720 --> 00:43:51,396
This is great. so where where do you see
people forcing AI into steps where
893
00:43:45,537 --> 00:43:46,537
Yeah.
894
00:43:51,396 --> 00:43:54,478
it doesn't belong? So if we're gonna step
back and we're gonna think about
895
00:43:54,478 --> 00:43:57,511
the Eisenhower matrix and think about your
client engagements,
896
00:43:57,511 --> 00:44:01,474
I'm sure you've seen this. Where are
people forcing it where it just should not
be?
897
00:44:02,636 --> 00:44:06,077
Yeah, the the complexity of like human
decision making is where I see a
898
00:44:06,077 --> 00:44:09,419
lot of folks trying to do it. So a an
analogy that I've come to use with folks
when
899
00:44:09,419 --> 00:44:11,770
they're overthinking it or trying to
overengineer something that should
900
00:44:11,770 --> 00:44:17,382
not be is NASA does not hire one rocket
scientist to build a rocket.
901
00:44:17,382 --> 00:44:19,973
They hire rocket scientists who builds a
team around them and
902
00:44:19,973 --> 00:44:23,064
one person is responsible for the metal on
the outside.
903
00:44:23,064 --> 00:44:25,975
One person's responsible for the fuel mix
in the engines.
904
00:44:25,975 --> 00:44:29,437
One like I could go on, you get the the
point of like there's not
905
00:44:29,437 --> 00:44:32,038
one person that does it. And so if you're
trying to
906
00:44:32,526 --> 00:44:37,138
Turn a process that involves ten people
who have to make very nuanced decisions
907
00:44:37,138 --> 00:44:40,619
every day or every week into one single AI
solution.
908
00:44:40,619 --> 00:44:43,530
That's probably a bad place to start and
you're probably gonna be disappointed.
909
00:44:43,530 --> 00:44:47,552
But if you're looking at it as this rocket
scientist as a team of 10 people,
910
00:44:47,552 --> 00:44:51,594
well, two out of those ten people probably
have something that's pretty rote,
911
00:44:51,594 --> 00:44:56,306
pretty repetitive. You could explain in
two minutes on paper what needs to be
done.
912
00:44:56,306 --> 00:45:00,297
And yes, maybe it's done twenty times a
week at an hour apiece.
913
00:45:00,297 --> 00:45:01,798
And so that's the part where we
914
00:45:02,018 --> 00:45:04,446
We encourage folks to think about adding
AI to first.
915
00:45:04,446 --> 00:45:06,664
Don't solve for the rocket scientists from
day one,
916
00:45:06,664 --> 00:45:07,664
solve for the team around them.
917
00:45:09,015 --> 00:45:11,116
Does that also well let me back up.
918
00:45:11,116 --> 00:45:15,879
~ that that was that was a perfect ~
perfectly understandable answer.
919
00:45:15,879 --> 00:45:19,972
And I'm I'm wondering whether that answer
is also an extension
920
00:45:19,972 --> 00:45:25,134
of your internal rules for w how you
decide what stays human versus what
belongs
921
00:45:25,134 --> 00:45:26,936
with an AI or an automated process.
922
00:45:26,936 --> 00:45:27,936
Yes?
923
00:45:27,660 --> 00:45:28,660
Yeah,
924
00:45:27,960 --> 00:45:31,601
like human in the loop is probably the
term that my team is sick of hearing me
use.
925
00:45:31,601 --> 00:45:33,752
Like I use it in every every discussion,
926
00:45:31,831 --> 00:45:32,831
Yeah.
927
00:45:33,752 --> 00:45:37,663
it feels like. Cause we don't we're we're
not here to to blow smoke
928
00:45:34,199 --> 00:45:35,199
Yeah.
929
00:45:37,663 --> 00:45:40,594
or sell people snake oil ~ only to be
disappointed by it.
930
00:45:40,594 --> 00:45:42,614
Like we want to build stuff we're proud of
and that actually works
931
00:45:42,614 --> 00:45:43,734
and moves the needle for people.
932
00:45:43,734 --> 00:45:45,125
So that's what I'm always saying.
933
00:45:45,125 --> 00:45:46,926
My team of like, okay, that's great.
934
00:45:46,926 --> 00:45:48,746
Where where's where can this go wrong?
935
00:45:48,746 --> 00:45:51,417
Where is the edge case? Where is the human
in the loop needed?
936
00:45:51,417 --> 00:45:54,258
And so that becomes I find usually a
937
00:45:54,319 --> 00:45:59,245
pretty easy thing to do if a biz if a team
sits down with a a person,
938
00:45:59,245 --> 00:46:01,177
a department, or everyone and says,
939
00:46:01,177 --> 00:46:04,411
let's map our process out, you can usually
spot pretty quickly
940
00:46:04,411 --> 00:46:08,306
the bottlenecks in human decision making
that's what's needed,
941
00:46:08,306 --> 00:46:10,038
what's not in that decision making.
942
00:46:10,038 --> 00:46:13,790
With that in mind, and we you you you
touched on this subject a little
943
00:46:13,790 --> 00:46:18,553
bit earlier in our conversation when you
talked about ~ not seeing layoffs,
944
00:46:18,553 --> 00:46:22,485
right? So there's a lot of fear mongering
out there that AI is gonna take jobs.
945
00:46:22,485 --> 00:46:26,558
~ I'm I'm gonna put a TBD on on that
entire conversation.
946
00:46:26,558 --> 00:46:30,710
~ but for right now, we are not seeing
that for the most part.
947
00:46:30,710 --> 00:46:35,193
What I am interested in ~ hearing from you
with your clients,
948
00:46:35,193 --> 00:46:37,364
when work does get automated.
949
00:46:38,238 --> 00:46:43,306
What actually is happening to the people
and to their jobs in your experience?
950
00:46:43,905 --> 00:46:49,226
Yeah, a lot of folks are getting
repurposed to if if most businesses have
some
951
00:46:49,226 --> 00:46:54,858
external ~ stakeholder to their business
that makes them successful,
952
00:46:54,858 --> 00:46:56,799
freeing up time to spend more time with
them.
953
00:46:56,799 --> 00:46:58,489
So whether you're selling a product,
954
00:46:58,489 --> 00:47:01,860
doing a service, you have users,
955
00:47:01,860 --> 00:47:03,661
if you're you're a technical app,
956
00:47:03,661 --> 00:47:08,212
that kind of thing, all of the time freed
up that I see is usually spent
957
00:47:08,212 --> 00:47:09,652
on investing in those relationships.
958
00:47:09,652 --> 00:47:12,263
So how do we retain people?
959
00:47:12,639 --> 00:47:14,150
As customers or clients for longer?
960
00:47:14,150 --> 00:47:16,121
How do we close business faster?
961
00:47:16,121 --> 00:47:17,742
How do we support them better?
962
00:47:17,742 --> 00:47:20,033
All of the repurposing of roles,
963
00:47:20,033 --> 00:47:24,766
because there has been some cases where
you most businesses have
964
00:47:24,766 --> 00:47:25,847
a never-ending amount of work.
965
00:47:25,847 --> 00:47:28,558
So their time is freed up and there's
other things now added to their plate
966
00:47:28,558 --> 00:47:32,731
to take its place. But for those cases
where they're in some sort of role that
967
00:47:32,731 --> 00:47:34,392
is connected to an external stakeholder,
968
00:47:34,392 --> 00:47:37,353
that time is repurposed and their
responsibility becomes well,
969
00:47:37,353 --> 00:47:40,255
your your new measurement is your
retention of your.
970
00:47:40,577 --> 00:47:42,670
your customers should be higher because
you now have more time
971
00:47:42,670 --> 00:47:45,023
to spend hearing their needs, solving
their challenges,
972
00:47:45,023 --> 00:47:49,213
whatever it might be. that's usually where
I see a lot of the repurposing from is,
973
00:47:49,213 --> 00:47:51,404
you know, what are the external
stakeholders to our business and how
974
00:47:51,404 --> 00:47:52,404
can we spend more time with them?
975
00:47:54,166 --> 00:47:57,768
I appreciate your frontline perspective on
that because it's impossible
976
00:47:57,768 --> 00:47:59,909
to ~ does anybody open a newspaper
anymore?
977
00:47:59,909 --> 00:48:03,571
I it's impossible to look at a a digital
article or an any headline about
978
00:48:03,571 --> 00:48:06,983
AI without every second or third one
screaming about job layoffs and,
979
00:48:06,983 --> 00:48:08,214
you know, this, that, the other thing.
980
00:48:08,214 --> 00:48:12,106
And I think it's important to get ~ real
information.
981
00:48:12,106 --> 00:48:15,308
Yeah, we've seen the studies that that
debunk a lot of these headlines,
982
00:48:15,308 --> 00:48:18,770
but I think it's important to get that
firsthand experience from from folks that
983
00:48:18,770 --> 00:48:22,262
are out there with clients and sitting and
f working with folks and
984
00:48:22,848 --> 00:48:26,173
What was behind that question in my mind
is when we had talked before,
985
00:48:26,173 --> 00:48:30,318
I think you had said to me that fewer than
one in twenty of your conversations with
986
00:48:30,318 --> 00:48:35,135
clients involved the idea of cutting
headcount and the other nineteen
987
00:48:35,135 --> 00:48:37,128
are are more in line with what you're
talking about here.
988
00:48:37,793 --> 00:48:40,545
Yeah, like if you sit down with the
average exec team,
989
00:48:40,545 --> 00:48:44,157
their five year roadmap is always full.
990
00:48:44,157 --> 00:48:46,358
Like there's never and it always gets
pushed out,
991
00:48:44,894 --> 00:48:47,037
Right. Yeah.
992
00:48:46,358 --> 00:48:49,990
and their OKRs are never achieved and
their metrics are never exactly where they
993
00:48:49,990 --> 00:48:55,368
want them to be. Like there is we live in
a capitalistic society with consumption
994
00:48:55,368 --> 00:48:59,425
out the wazoo. Like it there is never
going to be a shortage of of things that
995
00:48:59,425 --> 00:49:01,857
we make up to do for ourselves and each
other.
996
00:49:01,857 --> 00:49:05,319
And so I that sounds maybe too esoteric,
997
00:49:05,319 --> 00:49:06,319
but
998
00:49:05,965 --> 00:49:08,477
Point being of like every time we've seen
these technical changes,
999
00:49:08,477 --> 00:49:10,548
you know, new jobs have come with it.
1000
00:49:10,548 --> 00:49:13,760
If you even if you look at, I think when
you ask the average,
1001
00:49:13,760 --> 00:49:16,402
you know, elementary school student in the
70s,
1002
00:49:16,402 --> 00:49:18,604
80s, 90s, they said, I want to be a
teacher,
1003
00:49:18,604 --> 00:49:20,630
a doctor, a lawyer. You ask them today,
1004
00:49:20,630 --> 00:49:22,517
I want to be a YouTube content creator.
1005
00:49:22,517 --> 00:49:23,828
I want to like, I wanna be like Mr.
1006
00:49:23,828 --> 00:49:26,690
Beast. Like these jobs didn't exist 30
years ago.
1007
00:49:26,690 --> 00:49:30,232
What the hell was YouTube? But now these
are the number one most desired jobs
1008
00:49:30,232 --> 00:49:32,724
of of the child who's has the world open
to them.
1009
00:49:32,724 --> 00:49:33,975
~ so it's
1010
00:49:34,359 --> 00:49:37,805
That to me is just a very small anecdote
that is indicative of how we
1011
00:49:37,845 --> 00:49:40,760
act as a culture of every change has
brought new jobs and new ways
1012
00:49:40,760 --> 00:49:43,154
of working and living and that creates new
jobs,
1013
00:49:43,154 --> 00:49:44,154
et cetera.
1014
00:49:44,896 --> 00:49:46,787
I think that's a fantastic perspective.
1015
00:49:46,787 --> 00:49:50,498
you said something there that spurred a
thought with me too.
1016
00:49:50,498 --> 00:49:52,679
~ and I can't remember the exact verbiage
you used,
1017
00:49:52,679 --> 00:49:55,420
but it was basically we're very inventive
creatures,
1018
00:49:55,420 --> 00:49:57,131
right? We're gonna continue to invent
stuff.
1019
00:49:57,131 --> 00:50:00,502
And one of the things that I see an awful
lot is that people love
1020
00:50:00,502 --> 00:50:04,493
to invent complexity. ~ they're really bad
at simplifying things.
1021
00:50:04,493 --> 00:50:09,035
And I'm I'm wondering whether you're
seeing any ramifications from that with
clients
1022
00:50:09,175 --> 00:50:13,377
that have maybe engaged in some AI pilots
or some different work with their folks
1023
00:50:13,377 --> 00:50:14,377
because
1024
00:50:14,239 --> 00:50:18,053
What I have seen is a voluminous amount of
new information,
1025
00:50:18,053 --> 00:50:21,446
right? And I am also seeing in certain
instances a lot of complexity
1026
00:50:21,446 --> 00:50:25,130
~ that doesn't need to be there as a
result of people having access
1027
00:50:25,130 --> 00:50:27,882
to tools and kind of extending their reach
beyond where they should be.
1028
00:50:27,882 --> 00:50:30,345
Does that resonate with you? You see
anything like that?
1029
00:50:31,713 --> 00:50:34,304
I th I think there's two sides to that
coin that I often see.
1030
00:50:34,304 --> 00:50:37,154
One is going back to the cultural change
that leaders need to kind
1031
00:50:37,154 --> 00:50:39,535
of get their their teams pointed at,
1032
00:50:38,113 --> 00:50:39,113
Mm.
1033
00:50:39,535 --> 00:50:43,336
which is ~ a lot of times complexity is
born out of well,
1034
00:50:43,336 --> 00:50:44,336
we've always done it that way.
1035
00:50:44,277 --> 00:50:47,538
That is one of my both favorite and least
favorite phrases to hear when
1036
00:50:47,538 --> 00:50:50,228
we get into a new, new working
relationship with the team,
1037
00:50:50,228 --> 00:50:53,709
where we start to question their process
and we come we try to come from
1038
00:50:53,709 --> 00:50:56,790
it as tactfully as possible. When it's
very obvious that it doesn't make sense,
1039
00:50:56,790 --> 00:50:59,371
the answer usually is well, we've always
done it that way.
1040
00:51:00,065 --> 00:51:02,586
And it's like, well, so you've never
questioned your complexity?
1041
00:51:02,586 --> 00:51:03,836
Like, does it need to be complex?
1042
00:51:03,836 --> 00:51:05,616
What if it was to be simpler? What would
that look like?
1043
00:51:05,616 --> 00:51:07,357
I don't know. I never really thought about
that,
1044
00:51:07,357 --> 00:51:10,788
is usually the answer. So just giving
people permission to say,
1045
00:51:08,811 --> 00:51:09,811
Huh huh.
1046
00:51:10,788 --> 00:51:12,338
like, does something need to be this
complex,
1047
00:51:12,338 --> 00:51:13,699
or have you always just done it this way?
1048
00:51:13,699 --> 00:51:14,999
And usually the answer to that is,
1049
00:51:14,999 --> 00:51:16,629
well, yeah, we have always done it that
way.
1050
00:51:16,629 --> 00:51:17,720
We've never really rethought it.
1051
00:51:17,720 --> 00:51:19,820
On the flip side of the spectrum,
1052
00:51:19,820 --> 00:51:26,082
the the interesting thing to observe is
giving teams just a small sliver
1053
00:51:26,082 --> 00:51:29,523
of AI functionality and wins in their
company turns.
1054
00:51:30,436 --> 00:51:33,698
Kind of like the matrix when at the end he
can see the code and it's like,
1055
00:51:33,698 --> 00:51:35,188
I can see the matrix. I'm not just in it,
1056
00:51:35,188 --> 00:51:37,430
I can see it now. When you give them a
small slice,
1057
00:51:37,430 --> 00:51:38,970
they start to see opportunity everywhere.
1058
00:51:38,970 --> 00:51:41,992
And they go, Well, before when we were
gonna go after,
1059
00:51:41,992 --> 00:51:45,083
I'm I'm overusing sales as a motion or an
example,
1060
00:51:45,083 --> 00:51:48,005
but if we're a sales team and we want to
go after a certain segment,
1061
00:51:48,005 --> 00:51:51,276
we didn't have time to do the right
research to have the right conversations
with
1062
00:51:51,276 --> 00:51:53,027
the right people. But now we can.
1063
00:51:53,027 --> 00:51:57,209
Like we that was never it never felt open
to us because our our 40 hours a week,
1064
00:51:57,209 --> 00:52:00,131
50 hours a week was always filled to the
max with what we already had to do.
1065
00:52:00,503 --> 00:52:03,686
But now that we can actually have someone
aka I,
1066
00:52:03,686 --> 00:52:05,267
I I don't mean to humanize it,
1067
00:52:05,267 --> 00:52:08,841
but now that we can have a thing do a lot
of that work for us,
1068
00:52:05,473 --> 00:52:06,473
Mm.
1069
00:52:08,841 --> 00:52:11,243
this feels like a whole world that is open
to us.
1070
00:52:11,243 --> 00:52:13,355
And now we've got a whole new job to do
because we're getting
1071
00:52:13,355 --> 00:52:17,579
so much better information that we can
have better conversations with quote
unquote
1072
00:52:17,579 --> 00:52:20,452
better people or more suited people for
our business,
1073
00:52:20,452 --> 00:52:21,463
whatever that might be.
1074
00:52:22,689 --> 00:52:26,830
That's that's good. I I appreciate the
thoughtfulness behind that answer.
1075
00:52:26,830 --> 00:52:30,270
All right. Time always runs quickly.
1076
00:52:30,270 --> 00:52:32,852
~ we've got ~ limited time left here.
1077
00:52:32,852 --> 00:52:38,053
So let's go back to where we began for
maybe our our final conversation thread
1078
00:52:38,053 --> 00:52:42,175
or two here, which is data ~ and my data
house being a mess.
1079
00:52:42,175 --> 00:52:45,896
So let's put ourselves in the shoes of one
of our listeners here.
1080
00:52:45,896 --> 00:52:47,876
I'm a mid market leader, I'm listening to
this,
1081
00:52:47,876 --> 00:52:50,677
I'm thinking, honestly, I know my data
house is a mess.
1082
00:52:51,103 --> 00:52:54,000
What's the first thing that they need to
be doing when they walk in
1083
00:52:54,000 --> 00:52:55,037
the office the next day?
1084
00:52:56,961 --> 00:53:01,224
They should think about what are some of
the common pieces of data that span their
1085
00:53:01,224 --> 00:53:03,917
business. So again, going back to,
1086
00:53:03,917 --> 00:53:07,269
you know, ~ revenue, operating margin,
1087
00:53:07,269 --> 00:53:12,414
profitability, ~ even something as simple
as our customer name,
1088
00:53:12,414 --> 00:53:17,688
going to going to, you know, let's say you
have in a world in which you have Nike
1089
00:53:17,688 --> 00:53:23,042
as a client. In a world ~ where actually
Nike is too simple of a name.
1090
00:53:18,611 --> 00:53:21,253
In a world yes. Sorry. ~
1091
00:53:23,042 --> 00:53:26,055
Most businesses don't work with with
companies with that popular or simple
1092
00:53:26,055 --> 00:53:27,055
of a name.
1093
00:53:26,872 --> 00:53:32,216
But you work with ABC Corp. And so if you
go to sales and you go
1094
00:53:32,216 --> 00:53:36,199
to operations and you go to finance and
you go to customer support,
1095
00:53:36,199 --> 00:53:39,241
like go to all the bit the departments in
your business and say,
1096
00:53:39,241 --> 00:53:41,773
what is your source of truth for our
client list?
1097
00:53:41,773 --> 00:53:45,405
And what is the name of ABC Corp in your
your system of record?
1098
00:53:45,405 --> 00:53:48,568
If you don't have the same answer across
the board,
1099
00:53:48,568 --> 00:53:51,049
then you up you have a data problem.
1100
00:53:51,049 --> 00:53:53,151
And that problem could be a simple solve,
1101
00:53:53,151 --> 00:53:54,382
it could be a complex solve.
1102
00:53:54,904 --> 00:53:57,366
But if there was something that I would
encourage folks to do tomorrow
1103
00:53:57,366 --> 00:53:59,727
as leaders is go to your departments and
say,
1104
00:53:59,727 --> 00:54:02,899
what is your source of truth for whatever
the data point is and what does
1105
00:54:02,899 --> 00:54:05,170
it say for give them an example.
1106
00:54:05,170 --> 00:54:09,193
And usually they will find it's almost
always a different source of truth.
1107
00:54:09,193 --> 00:54:12,705
And more often than not, it's a different
name for that data point.
1108
00:54:14,143 --> 00:54:15,596
I love the simplicity of that.
1109
00:54:15,596 --> 00:54:20,055
Is there also a test that's that simple
that somebody could use
1110
00:54:20,055 --> 00:54:21,749
to identify their weakest process?
1111
00:54:24,065 --> 00:54:27,517
That one's a bit harder. I I I won't say
it's super simple,
1112
00:54:25,502 --> 00:54:26,502
Ha ha ha.
1113
00:54:27,517 --> 00:54:31,359
but again, going to department heads and
just saying,
1114
00:54:31,359 --> 00:54:35,781
could you list out all the processes that
your team have in the next 24 hours?
1115
00:54:35,781 --> 00:54:37,702
And if they can't, then that probably
means they haven't.
1116
00:54:37,702 --> 00:54:40,333
Like a a good not a good team,
1117
00:54:40,333 --> 00:54:43,365
because it's it's I don't want to knock
people who haven't defined their process
1118
00:54:43,365 --> 00:54:46,066
and mapped it out. That's not a lot of
people's jobs to do all day.
1119
00:54:46,066 --> 00:54:47,777
But as a leader of a department,
1120
00:54:48,203 --> 00:54:50,364
If you have no idea how the department
functions beyond,
1121
00:54:50,364 --> 00:54:52,665
well, we've always done it that way and we
have a we have a tribal
1122
00:54:52,686 --> 00:54:56,528
or institutional knowledge with each other
and it's not written down anywhere,
1123
00:54:56,528 --> 00:54:59,339
then how can you apply AI to it?
1124
00:54:59,339 --> 00:55:03,081
Like what are you going to give to AI to
explain how it should do
1125
00:55:03,081 --> 00:55:05,553
the process that you're asking it to do
that a human did before?
1126
00:55:05,553 --> 00:55:09,075
And so a a quick, you know, litmus test is
could you provide
1127
00:55:09,075 --> 00:55:14,298
me written documentation or a workflow map
or any sort of documentation
1128
00:55:14,338 --> 00:55:16,199
of how the process in your department
runs?
1129
00:55:17,003 --> 00:55:18,783
within twenty four hours. And if they say
no,
1130
00:55:18,783 --> 00:55:21,977
I'd have to create it, then that's step
one.
1131
00:55:21,977 --> 00:55:24,549
You should probably create it because when
we come in,
1132
00:55:24,549 --> 00:55:29,733
that's what we're gonna do. So any company
that they work with externally
1133
00:55:25,609 --> 00:55:26,609
Right.
1134
00:55:29,733 --> 00:55:32,815
as a consultant is going to want to do
that because we all know working with
1135
00:55:32,815 --> 00:55:35,747
AI every day, if you can't just say we're
not there yet at least,
1136
00:55:35,747 --> 00:55:38,179
where read those people's minds and figure
out how it works.
1137
00:55:38,124 --> 00:55:39,124
Yeah
1138
00:55:38,494 --> 00:55:40,655
yeah, not not quite yet. Not quite yet.
1139
00:55:40,655 --> 00:55:46,036
All right. I I'm always hesitant to ask
future casting questions,
1140
00:55:46,036 --> 00:55:49,778
~ but I I've got kind of a a two-parter
here for you as we wrap up.
1141
00:55:49,778 --> 00:55:55,569
What is something that you've been most
surprised by ~ in the last three
1142
00:55:55,569 --> 00:55:59,131
to six months with AI? Has anything come
across your desk that
1143
00:55:59,131 --> 00:56:00,391
you didn't anticipate?
1144
00:56:00,391 --> 00:56:03,012
As much as I expected the models to get
better,
1145
00:56:03,012 --> 00:56:05,734
I did not expect them to get better this
fast.
1146
00:56:05,734 --> 00:56:10,807
Like it is it it always is painful to hear
when someone's like,
1147
00:56:10,807 --> 00:56:13,108
AI is crap. And then you ask them,
1148
00:56:13,108 --> 00:56:17,501
How do you use it? Well, I signed up for
Chat GPT a year ago and it it felt like
1149
00:56:17,501 --> 00:56:18,872
a glorified Google. I'm like, yeah,
1150
00:56:18,872 --> 00:56:22,604
that's fair a year ago. I don't think
you've realized what has changed in last
year.
1151
00:56:22,604 --> 00:56:27,326
~ so the the surprising thing is just how
quickly they have advanced.
1152
00:56:27,326 --> 00:56:28,326
And by that I mean
1153
00:56:28,881 --> 00:56:33,663
its ability fable from anthropic is a good
recent example of you used
1154
00:56:33,663 --> 00:56:35,654
to just give it something as a request,
1155
00:56:35,654 --> 00:56:37,044
whatever that prompt might have been,
1156
00:56:37,044 --> 00:56:38,771
as simple or complex as it might have
been.
1157
00:56:38,771 --> 00:56:40,836
And it usually just did the thing you
asked.
1158
00:56:40,836 --> 00:56:43,097
And either it did it right and you were
happy with it,
1159
00:56:43,097 --> 00:56:45,728
or more often than not, it didn't do it
right because it didn't understand
1160
00:56:45,728 --> 00:56:47,659
the nuance. It didn't think what's the
edge cases,
1161
00:56:47,659 --> 00:56:48,929
what's the things that could go wrong?
1162
00:56:48,929 --> 00:56:50,200
What are the things I don't know?
1163
00:56:50,200 --> 00:56:52,471
And now when you ask Fable that kind
stuff,
1164
00:56:52,471 --> 00:56:57,843
it is wild to see where you you can open
up the thinking reasoning that it's doing
1165
00:56:57,843 --> 00:56:58,843
and you can see
1166
00:56:58,643 --> 00:57:00,995
Well, but if I do this, then that would be
impacted.
1167
00:57:00,995 --> 00:57:03,818
But if I don't do this, then they they
wouldn't know that it wasn't done.
1168
00:57:03,818 --> 00:57:06,789
Like its ability to think outside the box,
1169
00:57:06,789 --> 00:57:09,913
so to speak, is impressive.
1170
00:57:09,913 --> 00:57:14,297
If I were I'm trying to think of like just
one word that isn't too hypey,
1171
00:57:14,297 --> 00:57:16,069
but like it it's very impressive.
1172
00:57:17,359 --> 00:57:20,801
And I think that's a good spin into the
last question,
1173
00:57:20,801 --> 00:57:22,532
which is again that future casting.
1174
00:57:22,532 --> 00:57:25,312
It is very difficult for us as humans,
1175
00:57:25,312 --> 00:57:29,506
I think, to understand the implications of
an exponentially increasing technology.
1176
00:57:29,506 --> 00:57:31,076
And we haven't plateaued yet, right?
1177
00:57:31,076 --> 00:57:34,098
That's kind of what you're talking about
with your recent comments here
1178
00:57:34,098 --> 00:57:35,339
and the improvements in the models.
1179
00:57:35,339 --> 00:57:39,391
What do you expect we're gonna see ~
either in terms of models
1180
00:57:39,391 --> 00:57:41,022
or functionalities or breakthroughs?
1181
00:57:41,022 --> 00:57:44,154
What are you thinking we may see in the
next six to eight months?
1182
00:57:44,154 --> 00:57:46,465
We're recording this midway through 2026.
1183
00:57:46,903 --> 00:57:50,201
~ so we got half a year left. ~ what do
you what do you think?
1184
00:57:51,902 --> 00:57:55,234
I th I think the models themselves are
going to continue to get better
1185
00:57:55,234 --> 00:57:56,825
in the same ways they've gotten better,
1186
00:57:56,825 --> 00:57:58,396
as in they are better at reasoning,
1187
00:57:58,396 --> 00:58:03,560
they are better at understanding what you
really mean if you give
1188
00:58:03,560 --> 00:58:04,880
it something vague, that kind of thing.
1189
00:58:04,880 --> 00:58:09,423
What I think we'll see the most
advancement is the infrastructure around
the models.
1190
00:58:09,423 --> 00:58:13,916
And by that I mean the memory and the
context and the ability to connect them
1191
00:58:09,981 --> 00:58:10,981
Mm.
1192
00:58:13,916 --> 00:58:15,648
to multiple things at once. Right now,
1193
00:58:15,648 --> 00:58:19,170
you know, most of the companies are
competing for lock in,
1194
00:58:19,170 --> 00:58:20,731
as in we want you to use only our
1195
00:58:21,355 --> 00:58:23,877
And only the way that we we set it up for
you.
1196
00:58:23,877 --> 00:58:27,830
~ Because of the open source movement
around AI models and whatnot,
1197
00:58:27,830 --> 00:58:31,642
and just the kind of like in the cloud era
when AWS became a big thing,
1198
00:58:31,642 --> 00:58:35,204
all AWS is really infrastructure around a
bunch of data centers
1199
00:58:35,305 --> 00:58:39,517
and networking devices. They just built
really good infrastructure that propelled
1200
00:58:39,517 --> 00:58:43,330
them quickly. I think we're gonna see the
same thing over the next little while with
1201
00:58:43,330 --> 00:58:48,393
AI ~ and what what benefit that provides
to people who are listening beyond just
1202
00:58:48,393 --> 00:58:50,455
the like nerdy technical cool stuff
happening.
1203
00:58:51,075 --> 00:58:54,338
Is you don't have to re-explain the same
thing every time you start
1204
00:58:54,338 --> 00:58:59,522
a new chat session. You don't have to
spend a ton of time connecting different
1205
00:58:59,522 --> 00:59:01,984
sources and having it understand the
nuance of,
1206
00:59:01,984 --> 00:59:05,947
well, I've got revenue data over here and
operating cost data over here
1207
00:59:05,947 --> 00:59:07,929
and and employee data over here.
1208
00:59:07,929 --> 00:59:09,810
And I have to tell it how to think about
all that.
1209
00:59:09,810 --> 00:59:12,624
Like the models are getting so good and
the infrastructure around it
1210
00:59:12,624 --> 00:59:15,435
is getting so good that it understands,
1211
00:59:15,435 --> 00:59:19,788
well, this business revenue model plus
their operating model plus their staff
costs
1212
00:59:19,788 --> 00:59:20,788
means.
1213
00:59:20,741 --> 00:59:25,181
XYZ for next year's budget. And that
infrastructure is what makes this more
useful.
1214
00:59:25,181 --> 00:59:28,237
And so I think the big shifts that we're
seeing already right
1215
00:59:28,237 --> 00:59:30,692
now is the infrastructure is becoming a
lot more mature really quickly,
1216
00:59:30,692 --> 00:59:31,775
which is what really matters.
1217
00:59:32,369 --> 00:59:35,820
Yeah. I think that's as excellent of a
future cast as I've heard.
1218
00:59:35,820 --> 00:59:40,202
So ~ Justin, I have ~ tremendously enjoyed
our time together today.
1219
00:59:40,202 --> 00:59:42,573
This was a a fantastic conversation.
1220
00:59:42,573 --> 00:59:46,205
If folks that didn't have the privilege of
having a microphone in front
1221
00:59:46,205 --> 00:59:49,376
of them ~ are listening in and they want
to get in touch with
1222
00:59:49,376 --> 00:59:52,227
you for a conversation, how would they go
about doing that?
1223
00:59:53,299 --> 00:59:57,785
Yeah, or feel free to check out what we do
at with switchboard dot com.
1224
00:59:57,785 --> 01:00:00,178
~ and if you search me up on LinkedIn,
1225
01:00:00,178 --> 01:00:04,373
Justin Watt, W A T T, and I think it's
just LinkedIn slash Watt Justin,
1226
01:00:04,373 --> 01:00:07,417
~ name backwards. If anyone had any
specific questions or wanted to
1227
01:00:07,417 --> 01:00:10,652
go further down any of these rabbit holes
that we've we've started today,
1228
01:00:10,652 --> 01:00:11,652
then feel free to reach out.
1229
01:00:12,699 --> 01:00:15,034
And they are many and varied. you could
trip,
1230
01:00:15,034 --> 01:00:18,432
~ break an ankle or sink your entire leg
down those rabbit holes.
1231
01:00:18,432 --> 01:00:21,539
So Justin, again, it's been a pleasure.
1232
01:00:18,510 --> 01:00:19,510
Mm-hmm.
1233
01:00:21,539 --> 01:00:23,163
Thank you very much for your time today.
1234
01:00:23,667 --> 01:00:24,702
Goes both ways. Thanks for having me,
1235
01:00:24,702 --> 01:00:25,702
Chad.
1236
01:00:25,967 --> 01:00:30,140
And thank you to our listeners for once
again tuning in to AI for the C Suite.
1237
01:00:30,140 --> 01:00:34,002
If you liked what you heard, subscribe
wherever you get your pod on,
1238
01:00:34,002 --> 01:00:37,645
follow us on LinkedIn and check out
AIForthece Suite dot com.
1239
01:00:37,645 --> 01:00:40,667
And until next time, keep your algorithms
running,
1240
01:00:40,667 --> 01:00:43,789
your leadership evolving, and your AI in
check.