Innovating Under the Government Umbrella with Adam Carpenter

WEBVTT

1
00:00:03.940 --> 00:00:04.770
Kalia Garrido: Okay.

2
00:00:09.640 --> 00:00:11.480
Kalia Garrido: And now we wait.

3
00:00:13.760 --> 00:00:15.520
Kalia Garrido: And there we go. Okay.

4
00:00:15.520 --> 00:00:16.050
Mike Lampa: And.

5
00:00:16.059 --> 00:00:18.449
Kalia Garrido: Hey, Adrian Dino!

6
00:00:19.270 --> 00:00:21.800
Kalia Garrido: Nice to see your names coming on in.

7
00:00:23.483 --> 00:00:32.600
Kalia Garrido: Okay, well, we'll give it just another minute. Really cruise on to the top of the hour before we get things officially kicked off. Let a few more folks

8
00:00:33.200 --> 00:00:38.009
Kalia Garrido: join us and Adam and Mike thank you for joining me today.

9
00:00:38.240 --> 00:00:38.890
Adam Carpenter: Absolutely.

10
00:00:39.610 --> 00:00:49.099
Kalia Garrido: For our riveting discussion about what vitamins I'm taking. Which right before this, it's not all data, you know, it's or maybe it is like, that's a data point, too.

11
00:00:49.710 --> 00:00:53.279
Mike Lampa: Share your magic vitamin cocktail with the audience.

12
00:00:53.280 --> 00:00:53.979
Adam Carpenter: That's it.

13
00:00:53.980 --> 00:00:54.310
Adam Carpenter: That's right.

14
00:00:55.390 --> 00:00:57.810
Kalia Garrido: Next time we'll start by talking about how much water we drink.

15
00:00:57.840 --> 00:00:59.590
Adam Carpenter: Hashtag, not medical advice.

16
00:00:59.590 --> 00:01:02.669
Kalia Garrido: Yeah, definitely, not definitely, not

17
00:01:04.298 --> 00:01:16.030
Kalia Garrido: great. Okay. Alright. I see everybody coming in. So this is great, and we will officially get ourselves kicked off. I would like to. Let's just make sure. Yep, we're recording

18
00:01:16.300 --> 00:01:26.239
Kalia Garrido: so welcome. Thank you. Everybody for joining us today. My name is Kalia Garrido, and I head of marketing and events here at great data minds which is now a hike to company.

19
00:01:26.280 --> 00:01:35.940
Kalia Garrido: So if you haven't met us yet, great data mines is a collective of passionate data activists. And we are on a mission to modernize the world of data. And we do this in 2 different ways.

20
00:01:35.970 --> 00:01:44.620
Kalia Garrido: The 1st is that we have our services arm. This is where we do strategic planning education and the deployment of critical data projects that happens@hiketo.com.

21
00:01:44.630 --> 00:01:56.099
Kalia Garrido: And in addition to our data services, Hi, 2 is a fully functional best in class innovation consultancy that specializes in digital transformation, strategy, design and implementations.

22
00:01:56.210 --> 00:02:17.920
Kalia Garrido: Now, when it comes to our data and analytics, community content and conversation, just like what we're up to today. You can find more about what we're doing at Hi 2.com forward slash events, and you can also find us on eventbrite where? You'll find all of the things that we have coming up and follow us on Youtube. If you'd like to see the recordings of events just like this.

23
00:02:18.080 --> 00:02:37.129
Kalia Garrido: So a little bit of housekeeping before we get going. This is a webinar. So of course, your cameras and microphones are off, but we would love you to interact with us. We do encourage engagement via the chat, the Q&A. And if you would like to wait, we'll reserve a little bit of time. At the end of the session for a more formal kind of question and answer.

24
00:02:37.380 --> 00:02:45.569
Kalia Garrido: And of course we are recording. And like, I said, it's going to be posted to the Youtube channel. We always get questions about that. So, yes, we are recording today's session.

25
00:02:45.870 --> 00:03:04.009
Kalia Garrido: So allow me to do some introductions for our special guest. Today we have Adam Carpenter with us. He is the chief data officer at the State of Montana, and he's currently working to modernize Montana's data infrastructure by breaking down data, silos and improving citizen data accessibility.

26
00:03:04.220 --> 00:03:16.859
Kalia Garrido: Adam has 10 years of experience in machine learning. He's been guiding Montana's data maturity journey, and he is dedicated to leadership, mentoring and sharing his love for data science. Adam, thank you so much for joining us today.

27
00:03:17.210 --> 00:03:18.840
Adam Carpenter: Thank you for having me. It's a pleasure.

28
00:03:18.980 --> 00:03:33.189
Kalia Garrido: Let's do this alright. My partner in crime per use. Mr. Mike Lampa is joining us. He is a seasoned data expert in digital transformation. He focuses on modernizing enterprise and data analytics programs.

29
00:03:33.280 --> 00:03:39.620
Kalia Garrido: He's got over 20 years of experience and he excels in implementing advanced technology tools and platforms.

30
00:03:39.790 --> 00:03:51.659
Kalia Garrido: His expertise and dedication have driven successful transformations and empowered organizations to thrive in the ever changing digital era. Mike, thank you for joining me, and please do take it away.

31
00:03:51.970 --> 00:03:55.409
Mike Lampa: Thank you. Kalia and Adam. How are you, my friend? And Sarah.

32
00:03:55.410 --> 00:04:03.050
Adam Carpenter: I'm I'm well, yeah, good to chat with you. It's might as well put a camera on for what we would be doing over a dinner or something, anyway, right.

33
00:04:03.050 --> 00:04:07.168
Mike Lampa: Yeah, yeah, yeah. I enjoy our little fireside oriented chats.

34
00:04:07.580 --> 00:04:08.250
Adam Carpenter: Yeah.

35
00:04:08.250 --> 00:04:21.839
Mike Lampa: And to that point, let's just get into it. One of the things I'm going to offer up to the audience is over. The course of the last, I'll say 2 and a half years

36
00:04:22.120 --> 00:04:27.409
Mike Lampa: from my perspective, the States made some really impressive strides around establishing

37
00:04:27.430 --> 00:04:36.440
Mike Lampa: a data and analytics mindset, you know, with, you know, around literacy and and awareness and all those kinds of things. And and it's it's

38
00:04:36.540 --> 00:04:39.849
Mike Lampa: permeating across multiple agencies.

39
00:04:40.050 --> 00:04:47.550
Mike Lampa: How did you get this started. How did this come about? Because it, you know from like I said from my perspective, pretty impressive traction.

40
00:04:47.690 --> 00:04:48.230
Mike Lampa: Yeah.

41
00:04:48.230 --> 00:04:59.748
Adam Carpenter: It's you know I'm I'm I 1st I want to say I'm really proud of everybody that we work with at the State. All all of our partners and agencies, and so on. That's a big part of why we've gotten it done so quickly is we have so many wonderful people.

42
00:05:00.250 --> 00:05:02.639
Adam Carpenter: But you know it. Really, it starts with

43
00:05:03.170 --> 00:05:31.079
Adam Carpenter: a governor that really cares about this process. Right? Really cares about data analytics. This is a governor who started his own technology company and then sold it to Oracle before he got into politics. And it's a governor who knows how to run a large organization. And so he demands, metrics. And you know, planning like Smart and Ogsm, which requires you actually measure what you're doing and report on it, which is suddenly driven his top level management to.

44
00:05:31.338 --> 00:05:45.839
Adam Carpenter: you know, push down the chain this idea, that look, we have to actually start counting the numbers and telling people what we're doing. And so that helps that culture permeate very quickly when it comes all the way from the top and and is part of. You know the way you now operate, and then

45
00:05:46.795 --> 00:05:47.730
Adam Carpenter: it's

46
00:05:48.030 --> 00:05:55.909
Adam Carpenter: you know you mix that in with, you know I I brought my experience. I I started in AI and machine learning, as as was mentioned before. But I

47
00:05:55.950 --> 00:05:59.829
Adam Carpenter: I went to consulting a few years before joining the State, and.

48
00:06:00.610 --> 00:06:05.349
Adam Carpenter: and, you know, went there to lead their machine learning and AI practice for the Western United States.

49
00:06:05.450 --> 00:06:19.330
Adam Carpenter: and we pretty quickly realized almost none of our clients were actually ready to go on that journey. They needed to go on their data maturity journey first.st And so I really became this sort of Cdo on loan, and that allowed me to go. Industry to industry.

50
00:06:19.370 --> 00:06:25.010
Adam Carpenter: whether it's dod aerospace manufacturing or medical, or Hr. Hospitality. You name it

51
00:06:25.090 --> 00:06:37.299
Adam Carpenter: and realize they all have. If you zoom out enough, at least the same set of problems, with the same set of not exactly the same set of solutions, but the same set of best practice approaches right.

52
00:06:37.320 --> 00:06:47.959
Adam Carpenter: And and so the solution may look different from a vendor perspective or from an order, or it depends on what state they're in. But at the end of the day. The the end state you want to get to is is

53
00:06:48.020 --> 00:06:56.609
Adam Carpenter: pretty consistent across all these organizations, and so that I think helped a lot, because, as you get to the State.

54
00:06:57.170 --> 00:06:57.900
Adam Carpenter: it's

55
00:06:59.650 --> 00:07:07.640
Adam Carpenter: The breadth of it is is enormous. Right? There are 20 agencies in the executive branch. A number of constitutional agencies that are enormous.

56
00:07:08.147 --> 00:07:27.110
Adam Carpenter: You know, health and human services is as large as a fortune. 500 company is the largest insurance provider in the State, and the largest epidemiological research lab in the State. And that's 2 of their 18 divisions right? And so the the size is just enormous, and

57
00:07:27.550 --> 00:07:32.230
Adam Carpenter: and that led us to setting ourselves up as sort of a consultancy

58
00:07:32.290 --> 00:07:39.180
Adam Carpenter: right where, where the Cdo office is. This internal best practice center of excellence shop, and

59
00:07:40.000 --> 00:08:04.430
Adam Carpenter: by starting that way we allowed ourselves to get the people who were the early movers, the people who were going to be our best allies, anyway, right to come to us and get a win, and then go tell that story right. And when your best champions are out there telling that story it gets the next one in the door and the next one in the door. And you know, 2, 3 years later you've got 34 projects running concurrently, and

60
00:08:04.560 --> 00:08:12.280
Adam Carpenter: and the State looks a lot different than it did right. And so that's that's sort of the formula we've used is is, you know.

61
00:08:12.670 --> 00:08:16.489
Adam Carpenter: have have the support from the top, obviously, but but also just

62
00:08:17.034 --> 00:08:21.170
Adam Carpenter: set up an approach that's flexible, and that starts with your champions.

63
00:08:21.170 --> 00:08:36.780
Mike Lampa: Yeah, yeah. And that's a great model. I mean, certainly we are always striving to see that happen in in the private sector. You know where you help the the business domains which I'll make the logical equivalent to your- your agencies. I mean.

64
00:08:36.780 --> 00:08:37.100
Adam Carpenter: Yeah.

65
00:08:37.100 --> 00:08:40.130
Mike Lampa: When I look at when I look at a state

66
00:08:40.150 --> 00:08:46.720
Mike Lampa: it's huge, right? I mean, it's just this monster enterprise right? Yep.

67
00:08:46.980 --> 00:08:56.760
Mike Lampa: shift gears a little bit. You know one of the things that Julie Burroughs and I have always preached, and and Kaylee helps us get this message out is

68
00:08:56.770 --> 00:09:18.749
Mike Lampa: constantly lean in towards the more modern technologies in your data and analytics space. And when I look at the enabling technology stack that you put together, it definitely quacks of being, you know, on the modern end of the continuum. How'd you go about settling in and vetting the technologies.

69
00:09:19.530 --> 00:09:25.149
Adam Carpenter: Yeah, no, it's a great question. I mean, I think 1st you you can start with some high level assumptions.

70
00:09:25.580 --> 00:09:32.110
Adam Carpenter: you know one, the public sector kind of has no business being in the cutting edge space, if you will

71
00:09:32.450 --> 00:09:52.989
Adam Carpenter: it, it's not. Taxpayers don't really want us to experiment with their money. If that makes sense right? You know what I mean. So so we're sort of, we're not in the early Movers camp, but we should be right behind it, right where once something's been proven great, if we can, if we can afford to do it, and it actually improves our lives saves us money, etc. Then we should do it.

72
00:09:53.258 --> 00:10:04.379
Adam Carpenter: So that's to start with right? And that helps you eliminate like, okay? Well, we don't need any of the newest AI. We don't need, you know, and I call it AI, just because it's popular right now. But anything right, you know.

73
00:10:04.590 --> 00:10:06.359
Adam Carpenter: We'll wait till it gets proven.

74
00:10:06.390 --> 00:10:09.350
Adam Carpenter: But we also shouldn't be sitting on

75
00:10:09.440 --> 00:10:14.649
Adam Carpenter: old legacy system. So so to answer the question, part of it is back to the governor. In fact.

76
00:10:15.172 --> 00:10:21.869
Adam Carpenter: When he came into office he put forward a policy that was cots 1st consumer off the show

77
00:10:22.510 --> 00:10:32.039
Adam Carpenter: and cloud first.st Right? In other words, don't build. Don't custom. Build it yourself. If there's something out there that can do it and don't build another data center, just use the cloud and.

78
00:10:32.760 --> 00:10:42.534
Adam Carpenter: There. There's a few reasons I actually want to articulate about why that's really really important. Not just that, he said. The policy, but why, the policy itself is important.

79
00:10:43.350 --> 00:10:44.810
Adam Carpenter: and and that is

80
00:10:46.170 --> 00:10:57.089
Adam Carpenter: I don't. Somewhere 2017, 2018, I think, is when it really switched over to cloud being a more cost, effective approach. And and the reason is simply economies of scale. If you think about

81
00:10:57.387 --> 00:11:14.919
Adam Carpenter: how you have to build a data center, right. You you always have to build every server in your data center sort of for your worst day of the year. Right? So your your accounting system can't crash at end of year close, when the accountants are all hammering away at the system, so that that server's got to be built

82
00:11:15.277 --> 00:11:23.289
Adam Carpenter: powerful enough for that, or you know you can virtualize, etc. But but even then your your virtual server has to be set up for your worst day, right.

83
00:11:24.190 --> 00:11:24.880
Adam Carpenter: And

84
00:11:25.120 --> 00:11:32.260
Adam Carpenter: it's also when you invested a data center, every single asset you purchase is a depreciating asset. Except for the real estate.

85
00:11:32.460 --> 00:11:32.870
Mike Lampa: Hmm.

86
00:11:32.870 --> 00:11:40.460
Adam Carpenter: So you're essentially just leasing something, anyway, right? Because in in 5, 10 years you're gonna pay somebody to recycle it for you.

87
00:11:40.580 --> 00:11:41.810
Mike Lampa: And so.

88
00:11:41.950 --> 00:11:48.569
Adam Carpenter: You end up in this, you know no win situation where with the cloud you get

89
00:11:48.760 --> 00:11:57.569
Adam Carpenter: fixes for both of those problems. Right? You no longer have to be in the software management game. You don't have to pay electric bills and all of those things. You pay some overhead for that.

90
00:11:57.580 --> 00:12:03.090
Adam Carpenter: So the cost per say processor clock is a little bit more expensive.

91
00:12:03.310 --> 00:12:09.410
Adam Carpenter: But then you think about the fact that even in the best case for an on-prem server.

92
00:12:09.780 --> 00:12:17.869
Adam Carpenter: the majority of the day your server is sitting idle or more than half idle, right? So, even in the case of an E-com retailer on Black Friday

93
00:12:18.000 --> 00:12:25.260
Adam Carpenter: right? They have to plan for whatever their peak is on Black Friday. So 2 Am. On Black Friday, their server spending a lot of time doing nothing.

94
00:12:26.170 --> 00:12:32.470
Adam Carpenter: Because they're only getting a trickle of what they're gonna get when they get to their peak. And so the cloud allows you to

95
00:12:32.630 --> 00:13:00.750
Adam Carpenter: get into a place where you have as flexible a resource usage as you need. And the reason I'm sort of down in these weeds a little bit is because it really really matters to getting things done that you don't have to deal with these big capital expenditure problems that you do with on Prem, if you're making a choice, hey? We want to modernize data. Okay? Well, upfront. I'm gonna need 2 million dollars to build out your server farm or your Vm, or whatever it is. Okay. Well, never mind, we don't. We don't have it

96
00:13:01.370 --> 00:13:01.990
Adam Carpenter: right.

97
00:13:01.990 --> 00:13:09.180
Adam Carpenter: But if you can come up with in in this world of pay as you go. Use only what you need when you need it.

98
00:13:09.581 --> 00:13:24.039
Adam Carpenter: You can start with 10 grand or 20 grand and get something done right. And and so the often. What we do is we preach this look, start, start with 50 grand use 50 to save yourself 70, and you funded your next project right

99
00:13:24.060 --> 00:13:33.609
Adam Carpenter: with a little bit of overhead. And so you just keep. Keep that process rolling and people start getting excited about that. And and to give an example. So we had

100
00:13:35.360 --> 00:13:43.189
Adam Carpenter: a project in child Safety, where you know, child, Safety had 5 district offices along with the central office in the State.

101
00:13:43.420 --> 00:13:55.879
Adam Carpenter: and files were held in physical form, and filing cabinets spread across these 5 different regional offices. Now, I don't know how many know this, but Montana is the 4th biggest state in the United States right after California.

102
00:13:55.890 --> 00:14:06.060
Adam Carpenter: It's absolutely enormous and empty. It's the second emptiest state in the United States. Right? We have on average less than 6 people per square mile. So.

103
00:14:07.390 --> 00:14:19.109
Adam Carpenter: It. It's a long way between regional offices, and if you need a file for a child that's in danger again, this is the most vulnerable of the most vulnerable

104
00:14:19.598 --> 00:14:29.849
Adam Carpenter: you're gonna have to call and find the right district office, and they're gonna have to go dig it out of a filing cabinet and make copies and Fedex it to you. And right you, you start to see the the problem. And

105
00:14:29.880 --> 00:14:38.200
Adam Carpenter: for I think the total project cost was 208,000 we digitize 9 million records.

106
00:14:39.079 --> 00:14:46.090
Adam Carpenter: And now there is a portal you can go to. There is a call center. You can call that has access to the central store. If you don't want to. Self, serve

107
00:14:46.732 --> 00:14:55.190
Adam Carpenter: and response. Time went from days or weeks to hours or days, right? And and for a cost that is

108
00:14:55.930 --> 00:15:03.859
Adam Carpenter: it's, you know, essentially, it saved 400,000 per year in labor and 30,000 per year in Fedex costs alone.

109
00:15:04.190 --> 00:15:06.950
Adam Carpenter: all for an outlay that was

110
00:15:07.040 --> 00:15:19.299
Adam Carpenter: less than just the annual operating expense that it saved right. And that's what I mean by that sort of operating expense, capital expense model. If we'd had to shell out 2 million dollars, we're not going to see that return.

111
00:15:19.450 --> 00:15:21.769
Adam Carpenter: for in this case 2 and a half years right?

112
00:15:21.770 --> 00:15:22.370
Mike Lampa: Yeah.

113
00:15:22.830 --> 00:15:25.980
Adam Carpenter: And not counting maintenance on the system, by the way.

114
00:15:26.835 --> 00:15:28.610
Adam Carpenter: And and that

115
00:15:28.860 --> 00:15:38.139
Adam Carpenter: is just crushing that that stops, especially in public, but in private as well. That stops a lot of projects before they start, because you have to take a big risk

116
00:15:38.330 --> 00:15:40.240
Adam Carpenter: on what is at the time

117
00:15:40.250 --> 00:15:54.719
Adam Carpenter: an unknown reward. It's a black box often in technology, right? So let's yeah, let's shell out 2 million and try to invent a solution to this thing. Good luck, right? As opposed to. Okay, let's spend 50 grand and see if this thing works, you know, and if it doesn't

118
00:15:54.760 --> 00:16:03.429
Adam Carpenter: oh, well, we're out 50 grand. It's not the. It's not the worst thing in the world, and and usually it, it will work right? So it's a pretty good odds game. So

119
00:16:03.797 --> 00:16:10.729
Adam Carpenter: that's a big part of it, right? Is moving to that with tools like Snowflake, which is what the State is centralized on for data.

120
00:16:10.950 --> 00:16:16.810
Adam Carpenter: An agency can get in for $6,000 and have enough capacity for the full year.

121
00:16:16.850 --> 00:16:30.590
Adam Carpenter: and then, if they vastly expand their footprint and need to spend more, they're very, very willing, because they're the ones using it right? It got more expensive because they got value out of it. And so they end up coming. We never go to them and go. Hey, you're

122
00:16:30.610 --> 00:16:33.739
Adam Carpenter: you're getting to your limit. They come to us and they're like, Hey, we're

123
00:16:33.810 --> 00:16:45.790
Adam Carpenter: we want to do more and we're worried. We're hitting our limit. Can we have some more credits? Right? And so it it becomes this. People are happy, you know the the pay as you go. Models, I think, scare people

124
00:16:46.182 --> 00:16:51.150
Adam Carpenter: when they're approaching them the 1st time, but it ends up saving you a ton of money.

125
00:16:51.250 --> 00:16:59.040
Adam Carpenter: because you almost always overestimate. If you, if you're buying in bulk ahead of time, you always, almost always overestimate your need

126
00:16:59.350 --> 00:17:07.769
Adam Carpenter: and then end up with this big lump sitting at the end that you didn't use right. That's waste. And so you know this. This is a better approach. I think.

127
00:17:08.390 --> 00:17:15.710
Mike Lampa: Yeah, I love the the whole, the incremental. Let's take a little bite and see if we can generate a return on that, because.

128
00:17:15.710 --> 00:17:16.030
Adam Carpenter: Right.

129
00:17:16.030 --> 00:17:18.301
Mike Lampa: A self feeding model.

130
00:17:19.750 --> 00:17:29.199
Mike Lampa: and it generates, you know, to your earlier point, that's what starts to get people fired up and start to evangelize across the agencies. Second, hey.

131
00:17:29.200 --> 00:17:34.140
Adam Carpenter: Well, and Harvard Business Review spoke about this right. I mean they had an article. I want to say it was 2,007

132
00:17:34.450 --> 00:17:41.790
Adam Carpenter: that I think the headline was 86% of large scale. It projects go over to overtime or over budget.

133
00:17:41.790 --> 00:17:42.350
Mike Lampa: Yeah.

134
00:17:42.490 --> 00:17:43.120
Adam Carpenter: Right.

135
00:17:43.270 --> 00:17:46.759
Adam Carpenter: And you know, you read the article, and it's that's essentially

136
00:17:46.840 --> 00:17:51.550
Adam Carpenter: that's not a misleading headline. I mean, that's essentially what the research says, and what they found is.

137
00:17:51.570 --> 00:18:16.100
Adam Carpenter: it's large. The key word is large, right? If you break that project down into pieces that a human can sort of keep between their ears and manage. Then a Pm. Can keep all the balls in the air. But you know a juggler can only handle so many objects at once, and once it gets to be too many things start dropping, and it's the same with the project. Once it gets too big, it is simply inevitable that things will fall. Everything that falls is waste, and

138
00:18:16.170 --> 00:18:18.759
Adam Carpenter: you spend money on later right? So.

139
00:18:19.770 --> 00:18:31.879
Mike Lampa: I'm going to connect a couple of dots. You mentioned that prior to coming to the State, you you did a stint as a consultant, and you got exposed to a lot of different organizations and whatnot.

140
00:18:32.330 --> 00:18:42.909
Mike Lampa: And if I connect that to the fact that you do have some impressive adoption rates and literacy leveling up rates happening within the State.

141
00:18:43.640 --> 00:18:51.630
Mike Lampa: What is the model? What? What's your operating model out of the chief Data Officers office.

142
00:18:52.630 --> 00:18:54.510
Mike Lampa: the office of chief data.

143
00:18:54.780 --> 00:19:00.114
Adam Carpenter: Yes, indeed. Yes, or, as my kids call me, the the chief data officer.

144
00:19:00.860 --> 00:19:04.010
Adam Carpenter: the I keep my dad jokes in my database for the rest.

145
00:19:06.650 --> 00:19:08.670
Adam Carpenter: That's that joke was told to me by my children.

146
00:19:10.490 --> 00:19:12.320
Adam Carpenter: It's it's the best joke I had. By the way.

147
00:19:12.320 --> 00:19:13.349
Mike Lampa: Yeah, it's good.

148
00:19:13.480 --> 00:19:15.459
Adam Carpenter: It is. It's phenomenal. So

149
00:19:16.281 --> 00:19:24.989
Adam Carpenter: you know, and and you you sort of nailed it right? We act as an internal consulting agency. So we're here. We set ourselves up as the center of excellence. We put out a data strategy.

150
00:19:25.210 --> 00:19:34.800
Adam Carpenter: and we have some indirect power in the sense that you know, when we put out the data strategy, you know, we were brought in by the State Coo, who.

151
00:19:34.820 --> 00:20:02.200
Adam Carpenter: under direct orders from the Governor. So the Governor backed our play right, the Governor said. Look, this is the strategy. Everybody needs to follow it. So while we don't have the power to force you into a system. In general, people are required to follow or to work with us if you will, right, at least to listen. And so the 1st thing we did was and this is the 1st thing I'd recommend any sea level do, because I think sea levels do a lot of damage when they come in with a 1. Size, fits all approach.

152
00:20:02.200 --> 00:20:02.680
Mike Lampa: Hmm.

153
00:20:03.001 --> 00:20:06.540
Adam Carpenter: Is. We listened. I we spent, you know, 3 months just

154
00:20:06.650 --> 00:20:10.129
Adam Carpenter: listening, just talking to people, but you know I found I remember

155
00:20:11.125 --> 00:20:21.109
Adam Carpenter: what in the in the nineties? Right? Clinton was asked how he'd how he'd gotten so many votes in the in the 96 election, and his answer, I think, was, it's the economy stupid.

156
00:20:21.310 --> 00:20:25.110
Adam Carpenter: And what he meant by that was was essentially

157
00:20:25.200 --> 00:20:36.729
Adam Carpenter: that people may aren't always experts in all these obscure issues, right, all these secondary political issues. But what every single person is an expert in is, how am I doing

158
00:20:36.950 --> 00:20:43.310
Adam Carpenter: right? And so when you go around an organization. You start talking to people about their experiences. They are

159
00:20:43.390 --> 00:21:03.609
Adam Carpenter: eager even to tell you about all of the headaches and annoyances they have in relation to what it is. You do right? And and you know it for one. It's some catharsis, but 2. They're they want help. Everybody wants help. Nobody goes to work, wanting to have a frustrating day with an annoying system or process. And so

160
00:21:04.128 --> 00:21:07.960
Adam Carpenter: you go out and you listen, and and you have people

161
00:21:08.000 --> 00:21:16.459
Adam Carpenter: left, right and center that are eager and willing to tell you about their headaches, their pet peeves, the biggest problems they have. And you just start peeling those off

162
00:21:17.070 --> 00:21:27.170
Adam Carpenter: and and prioritizing them, basically by saying, like, Okay, well, what's the what's the biggest bang for the smallest buck? Right? And we just organize the list that way, and we work our way through it.

163
00:21:27.627 --> 00:21:30.340
Mike Lampa: Unfortunately, with this many organizations, we can.

164
00:21:30.450 --> 00:21:37.610
Adam Carpenter: Parallel a lot of things right that you know. Each org can work on one project at a time kind of thing. So that helps

165
00:21:39.230 --> 00:21:46.390
Adam Carpenter: But it's is also it. A big part of it is the cultural

166
00:21:46.780 --> 00:21:58.450
Adam Carpenter: change model, right? And you and I have talked about this before this. You can't change a culture quickly, you can't force it. If I had come in to the State and spoken to agencies about

167
00:21:58.500 --> 00:22:03.160
Adam Carpenter: data, their data, their data systems the way I can. Now

168
00:22:03.650 --> 00:22:12.230
Adam Carpenter: they would have glazed over and started snoring right in my face. Right? But in 3 years, because we have.

169
00:22:12.340 --> 00:22:15.979
Adam Carpenter: and we've taken this approach of consistently trying to

170
00:22:16.850 --> 00:22:36.439
Adam Carpenter: slightly raise the level of data literacy in every interaction, in every conversation. Right? So every time we present to the legislature. We we make the presentation a little. We add a couple concepts that we didn't that we didn't last time. And we we sort of slowly, almost. You know, it's the frog in the in the boiling water analogy, right where

171
00:22:36.890 --> 00:22:50.500
Adam Carpenter: they almost don't notice that they're becoming data Literate, because we're we're feeding it to them so slowly. And and that's the point. It shouldn't be painful, right? It should be this thing you just can learn through osmosis, at least to the level they need to. And so.

172
00:22:50.640 --> 00:22:56.719
Adam Carpenter: by constantly having these conversations and reinforcing that data literacy, we

173
00:22:56.760 --> 00:23:07.259
Adam Carpenter: we raise the ability for people in the State to have these more complex conversations to engage in this. And then, you know, pretty. Now we're at the point where

174
00:23:07.570 --> 00:23:21.229
Adam Carpenter: people come to us with ideas instead of us coming to them with an idea for a solution to their problem. People come to us and like, Hey, we were thinking Snowflake would probably do this, if we if and handle this problem that we've got and we're like, great, yeah, we'll help you do it.

175
00:23:21.670 --> 00:23:24.259
Adam Carpenter: And so part of it was

176
00:23:24.696 --> 00:23:37.319
Adam Carpenter: was that. And then I think the last thing I'll mention is when we set ourselves up as this sort of consultancy, this internal consultancy, the one of the 1st things we did was make a sort of menu of services, if you will.

177
00:23:37.995 --> 00:23:55.089
Adam Carpenter: And the idea was to let people know that they can come to us at any stage of the process. You know it. Do you have a data project that is at the 11th hour, and you just wanna make sanity check it. Do you have something? Are are you starting

178
00:23:55.090 --> 00:24:16.660
Adam Carpenter: from absolute 0 and need to have a discovery conversation on. If there's anything you should even do about data and everything in between right? And so we sort of laid out this this graphic, this walking path, the roadmap where where agencies can sort of see themselves on one of those nodes and go, okay, okay, they. They have a they have something for us right?

179
00:24:16.760 --> 00:24:19.620
Adam Carpenter: And that helped break down the

180
00:24:20.360 --> 00:24:41.689
Adam Carpenter: you know the well. We didn't want to bother you because we're not as bad off as that agency, or you know, or we're in such bad shape. We didn't even think it. You'd be able to help. And and all of those those things that cause people to hesitate, because at the end of the day the 1st problem you have to solve is the human problem, the the buy in problem right? People want better, but they don't necessarily think you can deliver it.

181
00:24:41.710 --> 00:24:43.560
Adam Carpenter: and especially in public.

182
00:24:43.710 --> 00:24:54.859
Adam Carpenter: They've been sold magic bullet after magic bullet after magic bullet that just didn't work right. And so some of them are sitting in 20, some odd year old systems that

183
00:24:55.220 --> 00:25:01.830
Adam Carpenter: you know, 3 replacements for which have failed, and that's why they're still in it, or whatever you know. And they're 2 million dollars into

184
00:25:01.830 --> 00:25:03.699
Adam Carpenter: big one. Sorry, right? Right?

185
00:25:04.350 --> 00:25:12.480
Adam Carpenter: Right? And and and this, you know, this happens a lot, and it's it is. It can be very difficult to get something to change right. And so

186
00:25:13.268 --> 00:25:39.090
Adam Carpenter: getting buy in is is a number one. So making yourself available and as simple as this sounds easy to work with. I mean, that's been, I think that's the primary review that we get to to the sort of management level above us is is, oh, they're just so easy to work with. Perfect. That's exactly what we needed, right, is it? We shouldn't be a chore right? We should be a relief.

187
00:25:39.280 --> 00:25:39.710
Mike Lampa: Yeah.

188
00:25:39.710 --> 00:25:40.390
Adam Carpenter: And so that's.

189
00:25:40.882 --> 00:25:42.359
Mike Lampa: Repeat Customer, yeah.

190
00:25:42.360 --> 00:25:47.459
Adam Carpenter: Yeah, absolutely that. And that's the point is like, people make yourself somebody people want to work with.

191
00:25:47.710 --> 00:26:07.420
Adam Carpenter: And people will want to work with you as simple as that sounds it it really it. It has a profound effect on how quickly you can get things done because people look forward to the meeting. They don't. They don't do that. So you've you've we've all done this where you like. That email comes through, and you're like, I don't have the bandwidth to deal with that right now, right.

192
00:26:07.930 --> 00:26:11.860
Adam Carpenter: And you're less likely to do that with somebody that you know just isn't going to be a headache.

193
00:26:11.980 --> 00:26:26.519
Adam Carpenter: Right? You're much more likely to get an email from me and go, hey, Adam, you know, and and answer real quick. And so it's just all these little delays get cut, and that speeds everything up. You know what I mean. And so it sounds simple, but it's kind of amazing how

194
00:26:26.700 --> 00:26:30.669
Adam Carpenter: how much, just being easy to work with and and decent.

195
00:26:30.750 --> 00:26:32.180
Mike Lampa: Makes a difference. Yeah.

196
00:26:32.180 --> 00:26:37.120
Mike Lampa: Well, and and leading with an empathy, you know. Interview model.

197
00:26:37.750 --> 00:26:38.320
Mike Lampa: Yeah.

198
00:26:38.320 --> 00:26:40.649
Adam Carpenter: Right? Well, yeah, what's bothering you again? That's

199
00:26:40.680 --> 00:26:55.619
Adam Carpenter: that's also people should do that also. Not just because it's the right approach. But it's it's also the least painful approach we've all been in those meetings where you're you're trying. You're going through this painful, awkward silence and discovery, trying to dig out what it is

200
00:26:55.900 --> 00:26:59.610
Adam Carpenter: we're doing here, and they want or need, or whatever.

201
00:26:59.750 --> 00:27:07.189
Adam Carpenter: and trying to get them to start talking. And like, I said, if you can just start asking them about what it is that drives them absolutely nuts.

202
00:27:07.420 --> 00:27:08.599
Adam Carpenter: In the morning.

203
00:27:08.770 --> 00:27:12.950
Adam Carpenter: They'll look. They've got a 10 min ramp loaded up. They're happy to tell you, right?

204
00:27:13.620 --> 00:27:25.229
Adam Carpenter: that's a different conversation. If you're asking people about what drives them crazy, they're loaded up, if you ask them, what what system do you think we need to build? They're like, I I don't know. I don't even know what the range of answers is right.

205
00:27:25.270 --> 00:27:26.040
Mike Lampa: Right, right.

206
00:27:26.325 --> 00:27:32.610
Adam Carpenter: And so, yeah, that having that conversation from an angle from the this perspective of empathy, I think, is much more effective.

207
00:27:32.610 --> 00:27:37.093
Mike Lampa: Yeah, yeah, there's a reason. Lean leaned into it.

208
00:27:38.120 --> 00:27:44.040
Mike Lampa: So the chief data officer and your office

209
00:27:44.240 --> 00:28:00.889
Mike Lampa: has a very consultative approach, right and the agencies have an opportunity to essentially buy in to the enabling technologies, you know, for so it's kind of it's a federated delivery model. Each agency has got kind of their own delivery organization.

210
00:28:00.990 --> 00:28:18.520
Mike Lampa: Does your does the extent of your consultation lean into helping the different agencies start to buy into consistent practices proven approaches. So that over time, from agency to agency, we're seeing recognized design patterns and whatnot.

211
00:28:19.010 --> 00:28:27.739
Adam Carpenter: No, I mean, you've given away the whole game, Mike. I mean, that's that's really it. Right is, that's entirely it, in fact, is is trying to

212
00:28:28.090 --> 00:28:31.589
Adam Carpenter: move from the. You know.

213
00:28:31.730 --> 00:28:38.950
Adam Carpenter: we know 2 things right? We know centralizing is impossible. There's no way an organization is this big can run centrally, it has to be distributed.

214
00:28:39.320 --> 00:28:49.289
Adam Carpenter: And and we know that there's a huge inefficiency that comes with just a fractured technology stack. Right? Just.

215
00:28:49.650 --> 00:28:56.990
Adam Carpenter: you know, some fracturing is probably diversification is important, to be sure, but but this having 20 different people.

216
00:28:57.250 --> 00:29:06.469
Adam Carpenter: Us. Collecting citizen data in 20 different systems, is definitely worse than having 20 of them use the same system of customer data.

217
00:29:06.960 --> 00:29:11.120
Adam Carpenter: you know, feeding into it and and pulling out from it clearly.

218
00:29:11.260 --> 00:29:14.859
Adam Carpenter: yeah, and so, yeah, that's absolutely the game. And and

219
00:29:15.230 --> 00:29:22.699
Adam Carpenter: and although you you, I think you phrased it well, there, right? Because it isn't necessarily getting on the same system, it's getting on the same

220
00:29:23.300 --> 00:29:32.510
Adam Carpenter: type of system, oftentimes. Right? So so, for example, in the State, we have by statute, each agency

221
00:29:32.840 --> 00:29:36.659
Adam Carpenter: owns its data, and the director essentially is the data owner.

222
00:29:37.158 --> 00:29:46.649
Adam Carpenter: Legally speaking. Now, obviously, that's the citizen state. It's held in trust. But but the agency is responsible for its safekeeping, so that the buck stops somewhere, right?

223
00:29:46.810 --> 00:29:53.020
Adam Carpenter: And what that means is one. Agencies are obviously very, very protective of the data for which they hold

224
00:29:53.370 --> 00:29:55.069
Adam Carpenter: but 2.

225
00:29:55.593 --> 00:30:12.579
Adam Carpenter: It also means that centralizing in the in the old model of like, we're gonna buy a data lake and put data from each agency into this lake was an absolute nonstarter. We knew we still wouldn't be 5 agencies in if that's the way we started right.

226
00:30:13.220 --> 00:30:19.510
Adam Carpenter: And so what we had to do is find something that allowed us to centralize virtually right. Snowflake had this, this

227
00:30:19.560 --> 00:30:31.990
Adam Carpenter: organizations model, where essentially each agency gets its own instance of snowflake, they fully control the firewall rules the identity management rules, etc. There, there are base requirements we have

228
00:30:32.380 --> 00:30:44.389
Adam Carpenter: centrally dictated. Right? You know. You have to. You know. Take a lease permissions approach, use single sign on, and 2 factor, and white list to the state network. Things like that. But but for the most part

229
00:30:44.400 --> 00:30:52.479
Adam Carpenter: they control it right? And and while we require those things, we don't actually log in and do those things, my office has access to nobody's snowflake instance, right?

230
00:30:54.410 --> 00:30:57.720
Adam Carpenter: And but because it's Snowflake

231
00:30:58.370 --> 00:31:02.050
Adam Carpenter: sharing that data agency to agency is now only

232
00:31:02.130 --> 00:31:08.080
Adam Carpenter: one click of them, or, you know a few clicks of the mouse or a couple small lines of SQL code, right?

233
00:31:08.580 --> 00:31:14.359
Adam Carpenter: And, practically speaking, that removes a huge barrier

234
00:31:14.400 --> 00:31:22.569
Adam Carpenter: from on on data sharing between agencies, right? Because it comes down to 3 things. Can I do it practically like, what is the development project to do this data share?

235
00:31:22.870 --> 00:31:28.140
Adam Carpenter: Can I do it ethically? Right? Is it the right thing to do? And can I do it legally right?

236
00:31:28.750 --> 00:31:29.670
Adam Carpenter: And

237
00:31:30.060 --> 00:31:38.959
Adam Carpenter: if if you know that you have a huge burden on Number One, we have a huge development project. Then I'm not going to bother paying to consider the other 2, which aren't cheap.

238
00:31:39.140 --> 00:31:46.980
Adam Carpenter: right? Oh, well, look, I'm not. Gonna I'm not gonna do 3 months of legal research so that I can do 3 months of data development to hand you something

239
00:31:47.040 --> 00:31:54.690
Adam Carpenter: for your benefit. Sorry other agency. That's just not how this works. If you remove that 1st one. Well, now, all of a sudden.

240
00:31:54.820 --> 00:31:58.779
Adam Carpenter: it's just a question of can I legally and ethically, do this right.

241
00:31:58.860 --> 00:32:16.446
Adam Carpenter: And you know there are ways to speed that up to, you know, good governance data dictionary, right? The the ability for the attorney to not have to go dig through obscure SQL. Servers located in random places in the network that they have to dig up. But just go look in a dictionary and see a preview.

242
00:32:16.780 --> 00:32:21.210
Adam Carpenter: but it's you know it. It also just

243
00:32:21.690 --> 00:32:26.470
Adam Carpenter: essentially gets you to the point of centralizing without centralizing.

244
00:32:27.130 --> 00:32:34.269
Adam Carpenter: It removes a ton of the problems that come with not centralizing as well. Right the data on thumb drives floating around and

245
00:32:34.300 --> 00:32:41.450
Adam Carpenter: file shares, and all these other things that each one of which is just a small breach in your security footprint. Right?

246
00:32:41.906 --> 00:32:49.640
Adam Carpenter: Now, the data is shared all entirely within this encrypted space. And so you know. That's just an example of

247
00:32:50.030 --> 00:33:00.709
Adam Carpenter: it. It it they don't all need to be in the same snowflake instance. It matters hugely that they are simply all on Snowflake. Right? And so, yeah, that is

248
00:33:00.750 --> 00:33:04.790
Adam Carpenter: for for data dictionary and for

249
00:33:07.228 --> 00:33:14.790
Adam Carpenter: data platform snowflake. We have this pick one right where it doesn't make sense to centralize on multiple data platforms.

250
00:33:14.820 --> 00:33:39.159
Adam Carpenter: But for everything else we we do what's called a managed choice model, right where we know that 8 various agencies have different needs, right? Some agencies have a huge need for Gis. Some agencies have a huge need for one type of dashboarding versus another, etc. And so when it comes to dashboarding tools, we we pick, you know, a number of them. We have tableau, we have power bi, we have

251
00:33:39.840 --> 00:33:42.489
Adam Carpenter: thoughtspot. We have a number of them, and

252
00:33:42.820 --> 00:33:48.959
Adam Carpenter: agencies can pick whatever it is they want, and and that especially matters in Montana, we have.

253
00:33:49.671 --> 00:33:54.040
Adam Carpenter: just over a million people in the State. It it is not easy to find.

254
00:33:54.180 --> 00:34:02.210
Adam Carpenter: you know, 5 tableau developers for every single agency. That just isn't a thing that's gonna happen in in the tiny town of Helena. And so

255
00:34:02.420 --> 00:34:15.009
Adam Carpenter: we have to be able to diversify that skill set or let agencies pick the tool that matches the skill set. They've happened to hire right? You. You hire an analyst, and maybe that analyst happens to know.

256
00:34:15.219 --> 00:34:17.760
Adam Carpenter: you know, click or or

257
00:34:17.929 --> 00:34:22.890
Adam Carpenter: you know, whatever other looker, whatever other tool is out there that you might want to use. Okay, okay, well.

258
00:34:22.920 --> 00:34:52.200
Adam Carpenter: fine. But let's let's what we do is then, is essential, obviously is sort of vet the best of those, put them through security, review, etc, etc, so that if an agency wants to grab one of those arrows out of the quiver. It's ready to go. It's sharp, it's feathered, it's it's ready to shoot right, and that's the point is to give them the option, you know. But but to have it to to have all the pre vetting done, because state procurement can be such a slog

259
00:34:52.230 --> 00:34:55.020
Adam Carpenter: that oftentimes it's like, well.

260
00:34:55.230 --> 00:35:12.059
Adam Carpenter: okay, why wait 9 months? Never mind, it's not. There's no point I need these results in a month, you know. Okay, well, we've got. Don't, don't. We've got a tool ready. It's already been reviewed. We've already got a sort of starter contract. You'll just have to amend it real quick to your capacity and go, you know.

261
00:35:12.170 --> 00:35:14.620
Adam Carpenter: And so that's that's a big part of the approach, too.

262
00:35:14.860 --> 00:35:16.849
Mike Lampa: Yeah, I like that managed choice model.

263
00:35:16.990 --> 00:35:31.739
Mike Lampa: Yeah. So so with the the creative juices that are continuing to percolate and produce some goodness across the agencies in the States. Without asking you to let the cat out of the bag.

264
00:35:32.190 --> 00:35:35.550
Mike Lampa: Where where are you guys going next? What's gonna you know

265
00:35:36.180 --> 00:35:38.820
Mike Lampa: kind of things are we going to be doing in the future?

266
00:35:40.220 --> 00:35:41.015
Adam Carpenter: Yeah,

267
00:35:41.840 --> 00:36:00.139
Adam Carpenter: it's a good question. I mean, we are moving into this phase where, as you've helped us with already, we're we're going to be doing more and more of going from. You know the raw and and you know, slightly refined layers of our of our data footprint or data lake to

268
00:36:00.517 --> 00:36:06.739
Adam Carpenter: you know, refined reporting data where? Where applicable at least it's not always necessary, as as you know. But

269
00:36:06.780 --> 00:36:14.121
Adam Carpenter: but you know this, the Kimball model right? This model of of star schema of,

270
00:36:15.930 --> 00:36:18.829
Adam Carpenter: you know, essentially flattening out.

271
00:36:18.850 --> 00:36:24.948
Adam Carpenter: denormalizing your data and then putting it into a format that's built for reads instead of built for, writes

272
00:36:25.310 --> 00:36:27.229
Adam Carpenter: you know. OLAP. If you will,

273
00:36:29.020 --> 00:36:33.770
Adam Carpenter: is, is is definitely our next step. We have a ton of data.

274
00:36:34.606 --> 00:36:50.370
Adam Carpenter: That is still in its very complex form, or even more often a ton of data that is a combination of data from a legacy system and data from the current system, or even multiple legacy systems in the current system, which means

275
00:36:50.410 --> 00:37:09.450
Adam Carpenter: even even the data isn't even self consistent. Right? It's you can't. You can't even go in and do a query out of a consistent Oltp system in a lot of these cases. And so you've got A. We've got to do the work of getting those data sets conformed Union together, if you will.

276
00:37:09.450 --> 00:37:21.069
Adam Carpenter: and then denormalized and put into a state where a human can actually get to them. And you and I have seen the magic that this process of

277
00:37:21.140 --> 00:37:33.580
Adam Carpenter: of dimensional modeling can do where a transactional database of 500 tables is simplified for reporting user down to a star schema with 7 tables or 10 tables right? And

278
00:37:34.140 --> 00:37:45.630
Adam Carpenter: you know, it's mind blowing. People can't believe the sort of you show them the the Uml side by side, and one of them is completely illegible, and one of them, you, even a layman sort of looks at and goes, oh, okay, I kind of get what's there right?

279
00:37:46.480 --> 00:37:52.150
Adam Carpenter: You know, sales by date and product, and you know, and and so

280
00:37:52.410 --> 00:38:01.950
Adam Carpenter: getting there, we've got a long way to go on that front. It. Part of that solution is, is folks like yourself. You've helped us do this a couple of times already.

281
00:38:02.140 --> 00:38:11.320
Adam Carpenter: Part of that solution is also, I think we're we're sort of waiting. There are a few people who are experimenting with AI dimensional modeling.

282
00:38:11.530 --> 00:38:16.990
Adam Carpenter: And you know, I think it's gonna be a long time before that gets to a hundred percent. But

283
00:38:17.140 --> 00:38:20.830
Adam Carpenter: I'm happy with Pareto. If you can get me to 80% and then

284
00:38:21.156 --> 00:38:28.039
Adam Carpenter: we can continue to hire great great partners like Hi 2, and so on, to come in and and take us. That last 20

285
00:38:28.050 --> 00:38:32.059
Adam Carpenter: do that bit of work that really takes the human mind with experience to do.

286
00:38:32.200 --> 00:38:38.909
Adam Carpenter: I think that's huge. But I I really am looking forward to seeing the advance of AI in that space, because

287
00:38:38.990 --> 00:38:48.440
Adam Carpenter: I think it's an easier problem to solve than large language modeling. Frankly, just from my AI experience, I think that was a harder problem to solve, although

288
00:38:48.470 --> 00:38:55.660
Adam Carpenter: a more popular one, which is why it went first.st But you know everybody wants to have a conversation with a new being right?

289
00:38:56.570 --> 00:39:01.589
Adam Carpenter: but I I think I think that'll be a huge step forward for the field.

290
00:39:01.590 --> 00:39:02.649
Adam Carpenter: Yeah, yeah.

291
00:39:02.790 --> 00:39:11.504
Mike Lampa: It'll be a huge win for you as well. What a brilliant as always! It's a brilliant dialogue that I get to have with you, Adam.

292
00:39:12.220 --> 00:39:16.549
Adam Carpenter: It's fun. I'm gonna scoot just a little because this sun is really coming in on me.

293
00:39:16.874 --> 00:39:27.250
Mike Lampa: So, Kaylee. I don't know if if we got any questions coming in from the audience or not, we can give give folks a few minutes to see if they they wanna

294
00:39:27.300 --> 00:39:32.079
Mike Lampa: ask Adam any questions of interest in their mind.

295
00:39:32.980 --> 00:39:38.620
Kalia Garrido: Absolutely. We do have a question. But, like Mike said, if there's any other questions, please feel free to use the chat or the.

296
00:39:38.620 --> 00:39:39.620
Mike Lampa: Can't, can't hear you.

297
00:39:39.620 --> 00:39:40.310
Kalia Garrido: In the future.

298
00:39:41.290 --> 00:39:42.679
Kalia Garrido: Can you hear me? Okay.

299
00:39:42.900 --> 00:39:43.260
Mike Lampa: It's.

300
00:39:43.633 --> 00:39:46.030
Adam Carpenter: Your mics, yeah. Echoing or.

301
00:39:47.550 --> 00:39:48.279
Kalia Garrido: What's just do that.

302
00:39:48.280 --> 00:39:48.840
Adam Carpenter: Thanks, right.

303
00:39:48.840 --> 00:39:50.460
Kalia Garrido: How's that? Better?

304
00:39:50.460 --> 00:39:50.929
Adam Carpenter: That's better.

305
00:39:50.930 --> 00:39:52.234
Mike Lampa: It's getting better. Yep.

306
00:39:52.560 --> 00:40:19.350
Kalia Garrido: I drifted, I went too far away. This is what happens. Okay, so we do have a question coming in. And like Mike said, if anybody has any additional questions, please feel free to use the Q&A function, or the chat, or, you know, kind of raise your hand and let us know. But you you had. You said a few interesting, I guess maybe not one liners, but it felt kind of like one liners, Adam. One of the things that you mentioned was.

307
00:40:19.350 --> 00:40:43.079
Kalia Garrido: public sector has no business in the cutting edge space. And so that sort of does beg the question. You know, how important is it for government entities, states, counties, cities to collaborate with and or take a page from the private sector, on what works and what doesn't we know they're able to work at the speed of whatever they want. But it's different when you're in the government. So how do you balance that line?

308
00:40:43.750 --> 00:40:50.339
Adam Carpenter: I mean, I think it where applicable, which is a big asterisk here. That's the the approach that public should always take.

309
00:40:50.560 --> 00:40:56.381
Adam Carpenter: What is the fortune 500 doing about this right on average. Right? You know.

310
00:40:57.260 --> 00:41:04.370
Adam Carpenter: it doesn't always work. There are a bunch of things that public does that. No private organization does.

311
00:41:04.610 --> 00:41:10.389
Adam Carpenter: you know? No, no private organization, you know, collects

312
00:41:11.060 --> 00:41:17.960
Adam Carpenter: disease samples from hospitals and has to liaise with the Cdc. That's just a completely unique

313
00:41:18.120 --> 00:41:29.543
Adam Carpenter: sort of, and maybe that's a bad example. But there are. There are, you know, livestock brands on. I mean, there are just so many things that are that are just utterly unique to public

314
00:41:29.920 --> 00:41:35.429
Adam Carpenter: And and so even there, I you know, it's it's

315
00:41:35.520 --> 00:41:39.510
Adam Carpenter: still partially the right approach, right? And we still looked at like, well, what are

316
00:41:39.680 --> 00:41:58.729
Adam Carpenter: we have a problem with with brands and handwritten brands, and it's all done on paper, you know, by a guy in muddy boots with a with a clipboard standing somewhere. That is an hour from cell service or internet. So we're not. Gonna we're not just solving this with service now, or something like that.

317
00:41:59.620 --> 00:42:23.150
Adam Carpenter: you know. And then you okay, well, what have other organizations just done about image processing? Then? Right? You're looking like the New York Times, used Google to process their entire photo basement, right? Their their vault of 100 150 years of photos, or something like that. And you know. And so you start seeing like, people are using the Big 3 for a lot of this image processing stuff. Okay, well, maybe we can do that. I mean.

318
00:42:23.720 --> 00:42:38.590
Adam Carpenter: You know, a hand drawn brand isn't the hardest image I think they've ever had to solve. And so that's what we're working. It's 1 of the things we're working on now and and so that's a big part of it. And I do want to be careful with my previous statement. I mean, I'll say

319
00:42:38.690 --> 00:42:42.770
Adam Carpenter: there is. A public has a place in the cutting edge

320
00:42:42.850 --> 00:42:55.739
Adam Carpenter: when it comes to research. I mean, the government is actually fairly good at research, but that has to be the stated purpose of the program. Right? The I think I would have meant it to say, the Government has no place

321
00:42:56.140 --> 00:42:57.770
Adam Carpenter: being on the cutting edge

322
00:42:58.280 --> 00:43:04.980
Adam Carpenter: to solve average everyday problems. Right? I mean, unless the program's purpose is to push the cutting edge.

323
00:43:05.170 --> 00:43:06.319
Adam Carpenter: they shouldn't be.

324
00:43:06.380 --> 00:43:12.040
Adam Carpenter: you know. Go buy something that's tested and true. Don't don't experiment with my tax money, please.

325
00:43:12.160 --> 00:43:26.720
Mike Lampa: Yeah, right? I love that, and I will. I'll put a plug in for livestock management. I thought I had seen all the coolest business models. and I was stunned at the level of complexity that goes into track.

326
00:43:26.720 --> 00:43:27.340
Adam Carpenter: It's incredible.

327
00:43:27.340 --> 00:43:28.600
Mike Lampa: Your livestock. Yes.

328
00:43:28.600 --> 00:43:31.540
Adam Carpenter: It's incredible. Well, yeah, this is, this is a.

329
00:43:31.870 --> 00:43:36.190
Adam Carpenter: it's wild. Because this is, you know, a hundreds year old industry.

330
00:43:37.273 --> 00:43:44.690
Adam Carpenter: Governments have been having to track livestock and manage property rights around it. Who owns what

331
00:43:44.860 --> 00:43:56.270
Adam Carpenter: bit of livestock, etc. For centuries? Yeah. And this is a huge, there's a huge pile of case law on this in the Uk and Europe all over the world. And so

332
00:43:56.280 --> 00:44:00.429
Adam Carpenter: yeah, it's actually it's with you. Get into it. You start talking to him.

333
00:44:00.780 --> 00:44:05.540
Adam Carpenter: people of livestock. And you realize, like, Oh, this is like one of the world's oldest industries like this

334
00:44:05.970 --> 00:44:27.279
Adam Carpenter: like this is ancient. This has been going on a long time, and a lot of it is still done that way, because they had to find solutions for this stuff. And and while I mean, maybe we've digitized it, or whatever like, the fundamentally the same thing is like, we're still branding cattle, right, we're still and it and it matters like it.

335
00:44:27.280 --> 00:44:27.830
Mike Lampa: Yeah, it does.

336
00:44:27.830 --> 00:44:33.490
Adam Carpenter: You know, one of the one of the most interesting things that we went through with livestock was going, learning how.

337
00:44:34.214 --> 00:44:43.860
Adam Carpenter: You know. So a huge part of how we learned how to do contact tracing during Covid was from these livestock groups, because they do this constantly, constantly

338
00:44:43.900 --> 00:44:50.670
Adam Carpenter: every time there's a Brucellosis outbreak, mad cow, etc, etc. You've got to be able to track

339
00:44:50.740 --> 00:44:53.260
Adam Carpenter: that brand. What other

340
00:44:53.570 --> 00:45:00.570
Adam Carpenter: animals were in that herd? What herds those herds interacted with on, I mean, in this sort of ad infinitum way.

341
00:45:00.760 --> 00:45:06.369
Adam Carpenter: And I mean, you know, those guys like livestock has drones and things like that that they I mean, it's

342
00:45:06.770 --> 00:45:10.020
Adam Carpenter: it's it's. It is a remarkable process. It really.

343
00:45:10.020 --> 00:45:11.470
Mike Lampa: It truly is, is.

344
00:45:11.470 --> 00:45:18.770
Adam Carpenter: You wouldn't expect them to have drones, you know, like predator, like predator style drones like these big flying runway style drones. It's wild.

345
00:45:18.770 --> 00:45:19.310
Mike Lampa: Yeah.

346
00:45:20.230 --> 00:45:43.650
Mike Lampa: as always, an incredible experience with you and kudos to the entire State, from the governor down to all the different agency heads that I'm starting to get to interact with it. And you and the people in there developing these solutions. Just a great story. And I hope the message gets out in other States.

347
00:45:43.690 --> 00:45:47.539
Mike Lampa: Take a couple pages from your book and look into it, because

348
00:45:47.550 --> 00:45:52.840
Mike Lampa: it certainly will help our nation improve without a.

349
00:45:52.840 --> 00:46:01.220
Adam Carpenter: Yeah, absolutely. No, I I thank you, Mike. It's always a great conversation. And I I think and we have a few minutes left. I I want to leave with this one story here

350
00:46:01.450 --> 00:46:06.590
Adam Carpenter: that the governor likes to tell that really sort of brings it home. Why.

351
00:46:07.320 --> 00:46:13.410
Adam Carpenter: why doing this work is so important, not just how to do it and how to be good at it. But why this work is so important.

352
00:46:14.980 --> 00:46:18.899
Adam Carpenter: and when so, when he was on the campaign trail. People kept coming up to him and asking him.

353
00:46:19.660 --> 00:46:27.189
Adam Carpenter: You know, or this developers specifically kept coming up and asking them, Why can't I get a sewer permit in your state it's impossible, right?

354
00:46:27.810 --> 00:46:36.529
Adam Carpenter: And and he said, Well, I don't know, but you know, lets me and lets me, and I'll find out right? Typical typical answer. And he gets in office. And sure enough, he starts looking into this and

355
00:46:37.130 --> 00:46:43.940
Adam Carpenter: it turns out that environmental quality Deq. Is responsible for

356
00:46:43.990 --> 00:46:48.219
Adam Carpenter: sewer permitting in the State right? And they built a system about

357
00:46:48.530 --> 00:46:53.839
Adam Carpenter: at the time 12 years ago. So it'd be 15 years ago now or 16 years ago.

358
00:46:55.470 --> 00:46:58.489
Adam Carpenter: To automatically parse these sewer applications

359
00:46:58.650 --> 00:47:04.689
Adam Carpenter: right? It was looking for 5 things that need to be included in your application. Right? You know, think engineering, spec etc.

360
00:47:06.580 --> 00:47:07.680
Adam Carpenter: And

361
00:47:08.020 --> 00:47:19.820
Adam Carpenter: the system. One of the 5 things that it was programmed to look for was slightly incorrect. So it was looking one of the things it was looking for was the wrong thing didn't match what was listed on the website.

362
00:47:22.060 --> 00:47:28.730
Adam Carpenter: And so there's a there's a 90 day statutory response time we have to respond in 90 days.

363
00:47:28.820 --> 00:47:30.390
Adam Carpenter: Whatever the answer is.

364
00:47:30.520 --> 00:47:33.710
Adam Carpenter: So the system would take this application from anybody

365
00:47:33.870 --> 00:47:41.880
Adam Carpenter: immediately rejected, because one of the 5 things was not what it was looking for, even though it was incorrectly looking for the wrong thing.

366
00:47:42.870 --> 00:47:45.500
Adam Carpenter: It would then hold it for 89 days.

367
00:47:46.910 --> 00:47:55.579
Adam Carpenter: and then inform you that your application had been rejected because one of the 5 things you submitted wasn't right. And so you'd go to the website and go. Well, yes, it was. I submitted the right thing.

368
00:47:56.160 --> 00:48:07.139
Adam Carpenter: And so it turned out that unless you knew somebody in Deq you couldn't get a sewer. Permit the system automatically rejected 99% of sewer permit applications for 12 years

369
00:48:07.320 --> 00:48:08.790
Adam Carpenter: 12 years.

370
00:48:09.310 --> 00:48:14.989
Adam Carpenter: and it turns out about 17% of people knew somebody at Deq they could call, which means

371
00:48:15.250 --> 00:48:22.630
Adam Carpenter: about 18% of sewer permit. Applications were approved over the course of that 12 year time span 82%

372
00:48:22.960 --> 00:48:42.019
Adam Carpenter: were rejected. Now, the people that know somebody, those are sort of local developers, and that's great. I'm always rooted for for local small businesses, but it isn't helping home buyers that Dr. Horton and all these other larger builders simply can't play in the State, because they don't know what the rule set is, and it isn't worth it to them to find out right.

373
00:48:42.670 --> 00:48:51.550
Adam Carpenter: And and so I want you to think about what the housing market implications are of 82% of applicants being unable

374
00:48:51.600 --> 00:49:08.730
Adam Carpenter: to be allowed by the government to build a house on the sewer system because you can't. Everybody's like, we'll just build septic. Okay? But you can't do that in a sewer area, right? So like, if you're building a subdivision in neighborhood. You have to build it on the city sewer. You can't put a septic in the middle of town.

375
00:49:08.830 --> 00:49:13.920
Mike Lampa: Right. And so so you're required to be on the sewer. But you're not allowed to be on the sewer.

376
00:49:13.960 --> 00:49:25.260
Adam Carpenter: And then you look into the State, and you see the State has a huge housing problem. The biggest State's residents. Biggest complaint is that nobody can afford a home housing. Prices are absolutely absurd in the State.

377
00:49:25.340 --> 00:49:26.860
Adam Carpenter: and

378
00:49:26.950 --> 00:49:31.469
Adam Carpenter: and the joke when you move here, is that all they ever build? Are these sort of 5 and 10 acre

379
00:49:31.893 --> 00:49:46.920
Adam Carpenter: you know, mansions on, out, out in the country, and you go well, that's all they can build. That's what you can build on a septic tank. You can't build an apartment complex on a septic tank, I mean, technically, I know you can. But nobody wants to

380
00:49:47.674 --> 00:49:52.539
Adam Carpenter: and you can't build a neighborhood in the city or a subdivision. Even

381
00:49:52.580 --> 00:49:56.410
Adam Carpenter: so, what all they're left with is

382
00:49:57.330 --> 00:50:18.309
Adam Carpenter: these big houses in the middle of nowhere which don't get me wrong. I'm grateful they exist, I'm in one. But but the best thing for the housing market would be for that to exist alongside a bunch of new builds within the city and expanding the city limits with suburbs, and that was essentially legally disallowed for 12 years. In 82% of cases. And so

383
00:50:18.310 --> 00:50:29.739
Adam Carpenter: you know, the the implications of a single configuration setting in a single system in a single agency are easily in the tens of billions of dollars

384
00:50:29.810 --> 00:50:33.120
Adam Carpenter: for real, everyday average people.

385
00:50:33.710 --> 00:50:36.530
Adam Carpenter: Bye, and that's why we do what we do.

386
00:50:36.750 --> 00:50:37.590
Mike Lampa: Yeah, right? Exactly.

387
00:50:37.590 --> 00:50:39.839
Adam Carpenter: Fixing that system took us a month.

388
00:50:41.250 --> 00:50:45.590
Adam Carpenter: You know. And so it, this, this work

389
00:50:46.060 --> 00:50:55.639
Adam Carpenter: has an impact, a huge outsized impact. And so I I leave with that because it's I think it's just a phenomenal story of how very, very little

390
00:50:55.910 --> 00:51:01.640
Adam Carpenter: little bit of inefficiency in government has an absolutely enormous effect.

391
00:51:01.890 --> 00:51:03.500
Mike Lampa: Damn awesome.

392
00:51:04.120 --> 00:51:07.000
Mike Lampa: Adam Carpenter. Thank you so much, sir.

393
00:51:07.120 --> 00:51:07.650
Mike Lampa: Yeah.

394
00:51:07.650 --> 00:51:09.530
Adam Carpenter: Absolutely. Thank you. Thank you for having me.

395
00:51:09.890 --> 00:51:16.929
Kalia Garrido: Yeah, that was a great story. You know. You said you you should make yourself somebody that people want to work with, and people want to work with you. I think you're that guy.

396
00:51:17.400 --> 00:51:19.349
Adam Carpenter: I try to be. I've tried very New York

397
00:51:19.350 --> 00:51:25.670
Adam Carpenter: to to sort of become that, you know. It doesn't have as as I think, Jay Leno once said, it doesn't have to be difficult.

398
00:51:25.910 --> 00:51:27.010
Mike Lampa: That's right.

399
00:51:27.060 --> 00:51:30.253
Kalia Garrido: I like it. Thank you for sharing your expertise with us today.

400
00:51:30.520 --> 00:51:30.900
Kalia Garrido: And they

401
00:51:30.900 --> 00:51:45.090
Kalia Garrido: thank you for everybody who's listening in you can follow us on Youtube to find the recording of this and a bunch of other sessions. We've got another session coming up on Thursday. If anybody wants to join and put the link in the chat, and we wish you all a wonderful day.

402
00:51:45.440 --> 00:51:46.760
Adam Carpenter: Yeah, thank, you.

403
00:51:47.020 --> 00:51:47.840
Kalia Garrido: Thanks. Everyone.

404
00:51:47.840 --> 00:51:49.870
Mike Lampa: The holiday season, too. Yeah.

405
00:51:49.870 --> 00:51:52.150
Kalia Garrido: That's it. Yeah, bye-bye.

406
00:51:52.310 --> 00:51:53.990
Mike Lampa: Alright, thank you. Bye-bye.