WEBVTT

1
00:00:10.440 --> 00:00:21.930
Brady Hugins: Alright, make sure I'm showing up here… There we are.

2
00:00:25.800 --> 00:00:33.069
Brady Hugins: Hey everyone, this is Brady Hugans, with Day 6 for the Mirror Mirror webinar series.

3
00:00:33.710 --> 00:00:40.679
Brady Hugins: Appreciate you all being here, whether you're live or joining us, via recording.

4
00:00:40.910 --> 00:00:41.920
Brady Hugins: So…

5
00:00:43.070 --> 00:00:52.510
Brady Hugins: This is a 30-45 minute presentation, usually. We do sometimes go over if needed, but it should be pretty quick.

6
00:00:53.130 --> 00:00:55.439
Brady Hugins: And go over some of the core…

7
00:00:56.300 --> 00:01:02.810
Brady Hugins: systems that we have. So, integration. Integration is…

8
00:01:03.230 --> 00:01:08.830
Brady Hugins: an important part of the ecosystem style that we develop, it's because

9
00:01:09.020 --> 00:01:20.119
Brady Hugins: It's about bringing everything together. It's the reason why we have core project work, we have core unified data, profile. We do have a central

10
00:01:20.230 --> 00:01:23.269
Brady Hugins: Tech stack that can work across many brands.

11
00:01:23.640 --> 00:01:34.690
Brady Hugins: But all of that is needed through integration and integration steps. So it's what holds the whole stack together, and what gives you the leverage in order to go across multiple brands.

12
00:01:35.460 --> 00:01:49.199
Brady Hugins: And you're doing this all on the back end, in the sense that, you have your own database, you have your own workflows, your own automation, some… often your own LLM tools. Sometimes they can be…

13
00:01:50.610 --> 00:02:00.489
Brady Hugins: AI LLMs that are remote, that you're working with, or some of the frontier models that are being released with Anthropic and whatnot?

14
00:02:01.030 --> 00:02:03.259
Brady Hugins: Or are they chemical models?

15
00:02:03.520 --> 00:02:07.339
Brady Hugins: Maybe simpler, depending on what kind of hardware you're running.

16
00:02:08.520 --> 00:02:11.210
Brady Hugins: So, appreciate you being here.

17
00:02:11.770 --> 00:02:15.680
Brady Hugins: So this is day 6 of a 7…

18
00:02:16.730 --> 00:02:22.679
Brady Hugins: day series now. I think people can probably get the nuts and bolts of everything in…

19
00:02:23.330 --> 00:02:39.639
Brady Hugins: two to three days, but, for general work cycle, you know, having it broken up into 7 different chapters is an easy way for integrating everything together. We've covered along the contact profile, so this… think of it as assembled, like.

20
00:02:40.030 --> 00:02:44.240
Brady Hugins: In phases, or different… Different segments.

21
00:02:45.830 --> 00:02:55.180
Brady Hugins: the plan, like, what do you plan for your own technology model? Like, what are you renting? What are you owning? What does it make sense for you? How do you build?

22
00:02:55.690 --> 00:03:03.949
Brady Hugins: Also, rhythms. So, what are common rhythms and checks that you have, either on a daily, weekly, monthly basis? What does that cycling look like?

23
00:03:04.280 --> 00:03:12.030
Brady Hugins: It could be checking in on air checking, it could be financial reporting, it could be… Database cleanups.

24
00:03:13.430 --> 00:03:21.570
Brady Hugins: Chapter 4 is engines and Automations. These are the… what I would consider the locomotives of the modern age.

25
00:03:21.910 --> 00:03:41.079
Brady Hugins: They can be incredibly complex, simple, work very quickly, or they can be very expensive, but there's what's driving the engines. There's what's moving things forward. Setting up these automations in… so that more intelligent work or more manual work is being done in these ways is what's…

26
00:03:41.300 --> 00:03:59.669
Brady Hugins: causing this AI evolution. Now, we're in the midst of it. I think we're at the beginning parts of the cycle, but we'll continue to develop, and things will get faster and more thorough and contextual over time.

27
00:04:00.640 --> 00:04:11.879
Brady Hugins: Chapter 5 is about delivering, and making sure that we're delivering on a… Performative basis to our…

28
00:04:12.900 --> 00:04:18.149
Brady Hugins: customers so that the last mile is covered. So, today we're working on integration.

29
00:04:22.540 --> 00:04:26.560
Brady Hugins: So… It becomes one system in this way.

30
00:04:29.480 --> 00:04:32.080
Brady Hugins: As we integrate the different tools together.

31
00:04:32.640 --> 00:04:35.959
Brady Hugins: So… When you have one system.

32
00:04:37.300 --> 00:04:55.110
Brady Hugins: you connect in various different parts of your brands. You may have different rules and brand structures that is different for each of your brands, or products, or different processes. And then you have core systems. So, these are designed much like enterprise-level

33
00:04:55.290 --> 00:05:06.149
Brady Hugins: Bureaucracies and functionalities are designed, but just for a small set, so that that functionality can be distributed through the scripting and coding that you have set up.

34
00:05:06.610 --> 00:05:13.659
Brady Hugins: It's a discipline, so it can read the same sources each time, it can be hard-coded or dynamically coded based on utility.

35
00:05:16.080 --> 00:05:20.930
Brady Hugins: But one of the things that you can rely upon is that the system is, in fact, working.

36
00:05:21.640 --> 00:05:38.070
Brady Hugins: the same way. Computers are reliable in the sense that they are very smart, they can… they can do a lot of amazing things, but they're very stupid in the sense that they do the same things over and over again, unless they have guidance and controls.

37
00:05:40.920 --> 00:05:52.440
Brady Hugins: So, one of the terminologies that's often used is drift. So, when two systems are holding in communion and union, where there's data flowing.

38
00:05:52.590 --> 00:06:00.300
Brady Hugins: And common data and intelligence, That's when there… there's no drift.

39
00:06:00.590 --> 00:06:04.810
Brady Hugins: But when they do inevitably work

40
00:06:05.370 --> 00:06:17.890
Brady Hugins: Over time, there's almost always some sort of drift, and you can measure that. And measuring drift in the systems is part of the integration component and meta-analysis that you're looking at.

41
00:06:19.040 --> 00:06:29.760
Brady Hugins: So that when you solve an email issue, maybe in one brand, it may come up and then be solved automatically in another brand, because that rule set fits for both.

42
00:06:30.280 --> 00:06:37.799
Brady Hugins: And when you do that, Dozens, or hundreds, thousands, millions of times, theoretically.

43
00:06:38.740 --> 00:06:46.009
Brady Hugins: You have a very involved system that can be very complex or simple based on what is preferred or needed.

44
00:06:47.830 --> 00:06:50.170
Brady Hugins: So, one of the key points, in fact.

45
00:06:52.610 --> 00:07:00.160
Brady Hugins: excuse me, is having a single source of truth. Single source of truth is our core database, or…

46
00:07:00.530 --> 00:07:10.020
Brady Hugins: Databases, where we shape the data, keep the data, contacts, orders, events, and having canonical versions of things.

47
00:07:10.500 --> 00:07:11.590
Brady Hugins: So…

48
00:07:12.070 --> 00:07:23.990
Brady Hugins: When we look at things, we can view into the data and know that that data is accurate, or that we have one central place that we say that this is the source of truth and work with that data.

49
00:07:24.100 --> 00:07:26.180
Brady Hugins: And compare it to other sources.

50
00:07:26.590 --> 00:07:35.320
Brady Hugins: So you have… the owner, and then the projection. So, the owner is the canonical source.

51
00:07:36.030 --> 00:07:37.740
Brady Hugins: It's the factual source.

52
00:07:39.150 --> 00:07:46.519
Brady Hugins: When I was working in enterprise-level software, this was a very important component to look at for databases, for sources of truth.

53
00:07:47.010 --> 00:07:50.240
Brady Hugins: To make sure that we were looking at the accurate information.

54
00:07:52.150 --> 00:07:57.749
Brady Hugins: And then projections from that information. In these cases, it could be dashboards, pages.

55
00:07:58.580 --> 00:08:08.190
Brady Hugins: Email merges… So… They're, a reflection of the data and the activity, so that there's… there's…

56
00:08:08.580 --> 00:08:18.489
Brady Hugins: system, or there's coding and process in the source of truth, and then that's being projected out, so it… it's created with, like, builders, web builders.

57
00:08:18.880 --> 00:08:24.479
Brady Hugins: You can build with… You can have web pages that build hundreds of pages very quickly.

58
00:08:25.290 --> 00:08:30.080
Brady Hugins: If you do it in the right methodology, with the right data in the right spots.

59
00:08:36.909 --> 00:08:48.080
Brady Hugins: So, this one has a note about bugs. The moment systems both think they own a field, you have a bug that's been shipping for weeks before anyone notices, and this is true.

60
00:08:48.670 --> 00:08:51.250
Brady Hugins: In workflows that are silent errors.

61
00:08:51.710 --> 00:09:02.550
Brady Hugins: Which means that there's a success that's being flagged, but it's a non-success from the actual effect. So there's some sort of coding error in there. So it's testing…

62
00:09:03.330 --> 00:09:07.819
Brady Hugins: It's called smoke testing. Whenever you go through a form.

63
00:09:08.050 --> 00:09:12.519
Brady Hugins: And you put in your testing information, it's called a smoke test.

64
00:09:16.280 --> 00:09:17.410
Brady Hugins: So…

65
00:09:18.670 --> 00:09:28.609
Brady Hugins: When you're able to do this, you're able to see React change things very quickly, update things very quickly. As an example, I've been able to, like, build out and…

66
00:09:29.140 --> 00:09:35.450
Brady Hugins: put a playlist for an event, or event series that I run. It's complicated.

67
00:09:35.560 --> 00:09:40.350
Brady Hugins: But it's an interesting use of the data. I have inputs where people

68
00:09:41.410 --> 00:09:50.750
Brady Hugins: give me suggestions for music feedback for an event, and then I curate them and put them in a playlist, depending on grouping and types.

69
00:09:52.540 --> 00:09:53.640
Brady Hugins: So…

70
00:09:54.160 --> 00:10:04.819
Brady Hugins: This is talking about observability in regards to failures. The quiet ones are the most dangerous ones, because you don't… if you don't know something's failing.

71
00:10:05.690 --> 00:10:14.299
Brady Hugins: Then it's… can be… A false sense of reality, or a false sense of what is actually going on.

72
00:10:16.860 --> 00:10:20.739
Brady Hugins: So… Part of the smoke detection

73
00:10:22.570 --> 00:10:25.270
Brady Hugins: Is part of verifying that process.

74
00:10:27.130 --> 00:10:28.620
Brady Hugins: So…

75
00:10:29.530 --> 00:10:37.399
Brady Hugins: This is a true story. We have a social publisher that I've been working on for a while. There's a couple different workflows.

76
00:10:38.510 --> 00:10:43.879
Brady Hugins: And it said it was successfully publishing for 37 days, every 4 hours.

77
00:10:44.400 --> 00:10:47.269
Brady Hugins: And it was a filter, silently, skipping…

78
00:10:47.760 --> 00:10:53.150
Brady Hugins: The approved post, so… there wasn't any, like, erring or anything.

79
00:10:55.270 --> 00:11:05.429
Brady Hugins: So, what happens is, if you don't… if you say that there's a workflow, but you're… that's working successfully, but if you're not measuring the success and verifying.

80
00:11:05.960 --> 00:11:10.799
Brady Hugins: Then you can have a false… Success for that workflow.

81
00:11:11.650 --> 00:11:13.090
Brady Hugins: So, it's about…

82
00:11:13.660 --> 00:11:21.520
Brady Hugins: checking and double-checking. An email is called SendLogs. I have daily digests that are emailed to me every day with any error checks.

83
00:11:22.360 --> 00:11:24.439
Brady Hugins: We can have self-healing.

84
00:11:24.730 --> 00:11:30.069
Brady Hugins: Workflows, which is a very interesting concept, where it checks for the error and then fixes it itself.

85
00:11:33.270 --> 00:11:41.049
Brady Hugins: Auditing. Auditing becomes a very important, regular process with an ecosystem, and seeing the outputs of those audits.

86
00:11:43.470 --> 00:11:49.380
Brady Hugins: So, webhooks. So, webhooks are a way of…

87
00:11:49.500 --> 00:11:52.990
Brady Hugins: Pinging different systems so that they're…

88
00:11:53.510 --> 00:12:01.650
Brady Hugins: interacting with each other, so it could be, like, a Stripe webhook where you have a payment come in.

89
00:12:02.300 --> 00:12:08.070
Brady Hugins: It could be a form webhook, where you have a form information come in.

90
00:12:09.640 --> 00:12:11.469
Brady Hugins: So, it's an event.

91
00:12:12.710 --> 00:12:15.470
Brady Hugins: And usually, these are the most important

92
00:12:15.870 --> 00:12:22.040
Brady Hugins: Parts of the website, because they measure some of the highest impact activities.

93
00:12:24.240 --> 00:12:26.980
Brady Hugins: So, it could be a tap on an AMC card.

94
00:12:28.110 --> 00:12:31.580
Brady Hugins: There could be… in the… in these issues.

95
00:12:33.580 --> 00:12:36.709
Brady Hugins: There can be errors that come up.

96
00:12:37.150 --> 00:12:46.750
Brady Hugins: The reception is down, the payload isn't formatted the right way, so that means the data is informed, or it's not staged in the right way at the right time.

97
00:12:48.580 --> 00:12:56.029
Brady Hugins: So there's a couple terms to go into with… Verification and resiliency.

98
00:12:56.410 --> 00:12:59.110
Brady Hugins: Indemp potency?

99
00:12:59.410 --> 00:13:07.209
Brady Hugins: It's basically making sure the… There's results happening the same way each time.

100
00:13:07.780 --> 00:13:13.070
Brady Hugins: So, there's a standardization of process, and that's occurring at the same time.

101
00:13:13.400 --> 00:13:15.319
Brady Hugins: retry failure, so…

102
00:13:15.470 --> 00:13:26.030
Brady Hugins: Failures happen. Like, it's part of the process that either there's internet goes down, lightning strikes a tree and collapses a…

103
00:13:26.730 --> 00:13:33.209
Brady Hugins: electrical pole, and all of a sudden, there's people without power. So a retry is needed.

104
00:13:33.590 --> 00:13:40.159
Brady Hugins: In order for it to be a part of the process when things fail, so it can't be assumed that it's working all the time.

105
00:13:40.850 --> 00:13:42.400
Brady Hugins: Dead letter cues.

106
00:13:42.930 --> 00:13:48.839
Brady Hugins: So this is… Checking… it's…

107
00:13:50.370 --> 00:13:59.050
Brady Hugins: the ones that fail still land in the dead letter Q, and then that way you get a record of what's happening when there's a

108
00:13:59.350 --> 00:14:04.159
Brady Hugins: A false or… or… not a false, a failure to connect.

109
00:14:05.080 --> 00:14:10.550
Brady Hugins: And then you could check on those the next day, or, do error checks with those.

110
00:14:11.040 --> 00:14:18.569
Brady Hugins: And then replay. Sometimes you can replay the issue.

111
00:14:23.910 --> 00:14:29.639
Brady Hugins: And… fix the issue, and then replay afterwards, and make sure that all goes through.

112
00:14:30.090 --> 00:14:32.659
Brady Hugins: Again, the end steps of a smoke test.

113
00:14:34.420 --> 00:14:41.859
Brady Hugins: And payments is a really common one, that you want to have this sort of style and system set up.

114
00:14:42.510 --> 00:14:47.679
Brady Hugins: And you can have this style and system set for each product in the ecosystem as well.

115
00:14:51.000 --> 00:14:59.619
Brady Hugins: Alright, so… This is a snapshot of my ecosystem. This is in Nation.

116
00:15:00.410 --> 00:15:01.959
Brady Hugins: Which, I have…

117
00:15:02.300 --> 00:15:14.289
Brady Hugins: several hundred workflows in there. I'm just gonna go through a standard form handler. So I have one contact spine, so that means if someone's marked for different brands or different products, there's just one contact spine.

118
00:15:15.980 --> 00:15:24.309
Brady Hugins: And then, we're gonna go over the standard form handler. So this is… Now you can have…

119
00:15:25.760 --> 00:15:27.839
Brady Hugins: Oh, let's zoom out just a little bit.

120
00:15:31.660 --> 00:15:39.389
Brady Hugins: So, oftentimes, you have a form, or many forms, on your website, Or wherever you're collecting information.

121
00:15:41.660 --> 00:15:45.600
Brady Hugins: And… That needs to be validated.

122
00:15:45.770 --> 00:15:53.530
Brady Hugins: And then it needs to be looped into the preferred, or the right existing contact, and then the

123
00:15:54.450 --> 00:15:56.700
Brady Hugins: Information needs to be merged.

124
00:15:57.600 --> 00:16:01.370
Brady Hugins: And then upload it into the unified contact profile.

125
00:16:02.540 --> 00:16:09.320
Brady Hugins: Here, this one has a… Confirmation email being sent as well, and the process being sent as well.

126
00:16:11.640 --> 00:16:16.130
Brady Hugins: Very simple, in some ways, very complex in others.

127
00:16:16.620 --> 00:16:24.480
Brady Hugins: Some of the data can be very complicated in some of these forms, with many different… Steps, or…

128
00:16:25.160 --> 00:16:30.759
Brady Hugins: Customization, so the merging process may be a little complex.

129
00:16:31.880 --> 00:16:37.230
Brady Hugins: We have the upsert into the unified contact and coordination with the latest information.

130
00:16:37.710 --> 00:16:54.249
Brady Hugins: So even though it's a simple process of handling a form, it may have some complexity behind it. Or it could be a very simple form, like an email form. Someone signs up and says, hey, I want to join your email letter list because I'd really love to go to your events.

131
00:16:55.780 --> 00:16:59.050
Brady Hugins: They sign up their name, their email.

132
00:17:00.070 --> 00:17:12.819
Brady Hugins: That's validated as useful and practical inputs. Looked up if there's a contact, or if it's a new contact. Merging the information into the unified contact profile.

133
00:17:13.550 --> 00:17:15.540
Brady Hugins: Sending confirmation email.

134
00:17:16.530 --> 00:17:28.329
Brady Hugins: So, this is a good example of what is standard. Now, you can copy this and then paste into different forms as well, different products.

135
00:17:28.520 --> 00:17:38.739
Brady Hugins: Or you can have a code builder go and build them, because Nation has MCP access, so a lot of these code builders can come in there and build these out for you.

136
00:17:40.480 --> 00:17:45.930
Brady Hugins: So, yeah, that's an example.

137
00:17:49.950 --> 00:17:53.189
Brady Hugins: So, even though this is just one form.

138
00:17:53.820 --> 00:18:00.900
Brady Hugins: You can have it be a form across multiple brands, multiple websites, Even…

139
00:18:01.120 --> 00:18:07.819
Brady Hugins: Agentically, as an agency, you can offer sign-up forms for people.

140
00:18:08.640 --> 00:18:14.169
Brady Hugins: That's one of the best things to do, is actually collect sales and marketing information for

141
00:18:14.520 --> 00:18:21.419
Brady Hugins: clients and customers. Almost everyone wants a new, happy customer that wants to meet with them and talk with them about their products.

142
00:18:24.450 --> 00:18:26.320
Brady Hugins: So,

143
00:18:28.140 --> 00:18:38.609
Brady Hugins: Yeah, so everything downstream reads the spine, so you can imagine the ingestion process going to your central database, and then a…

144
00:18:38.610 --> 00:18:50.320
Brady Hugins: a process where it reads and then puts out to different pages. So, you can have confidential information, private information, different categories of where to keep things.

145
00:18:50.690 --> 00:18:55.720
Brady Hugins: And yet the flow can be very quick still, even though it's… might be…

146
00:18:55.880 --> 00:19:03.880
Brady Hugins: in private and appropriate areas, with the proper key setup. Almost all of these have keys…

147
00:19:04.350 --> 00:19:11.519
Brady Hugins: That are maybe 50, 60 plus characters, some of them 125 characters, depending on the setup you have.

148
00:19:14.760 --> 00:19:15.590
Brady Hugins: Okay.

149
00:19:20.620 --> 00:19:30.379
Brady Hugins: Integration drift, so… There are certain rules or preferences that you want to keep in mind as you build.

150
00:19:32.690 --> 00:19:37.519
Brady Hugins: And… For me, it's having discipline around databasing.

151
00:19:37.870 --> 00:19:45.560
Brady Hugins: Keeping information in the proper databasing, keeping things, information,

152
00:19:46.220 --> 00:19:54.970
Brady Hugins: I'm aware, very aware of the database, I'm very aware of, like, where the data is, how much I have of whatnot. So it helps me…

153
00:19:56.020 --> 00:20:00.909
Brady Hugins: Understand the… Home of the data.

154
00:20:02.280 --> 00:20:07.399
Brady Hugins: The technical term is a canonical home. When we say source of truth, it's canon.

155
00:20:07.860 --> 00:20:12.300
Brady Hugins: You know, when they say, this is canon, Star Wars canon, Star Trek canon.

156
00:20:12.450 --> 00:20:19.549
Brady Hugins: You know, what is the appropriate source of truth? That's what they mean. So it's the canonical home?

157
00:20:20.010 --> 00:20:34.349
Brady Hugins: Audits. As I said before, audits are an important part of the process that often can occur on a daily basis or an ongoing basis when you check on things before you make any changes.

158
00:20:35.980 --> 00:20:51.209
Brady Hugins: deprecation, so these are things that you might have changed, in the past, or deprecated, which means that you're taking the technology off and not using it anymore, and having a registry for when that happens. So, you see how the discipline works?

159
00:20:52.090 --> 00:20:56.050
Brady Hugins: Throughout the ecosystem, and how… There's very little.

160
00:20:56.250 --> 00:20:57.790
Brady Hugins: Left chance.

161
00:20:58.160 --> 00:21:08.359
Brady Hugins: when you're working with some of these information systems, that even if AI drifts over in some ways, or assumes some ways, the audits take them back to…

162
00:21:09.020 --> 00:21:11.030
Brady Hugins: A standardized process.

163
00:21:11.320 --> 00:21:13.869
Brady Hugins: And AI does it very, very quickly.

164
00:21:16.880 --> 00:21:22.910
Brady Hugins: So, we have Franklin here, and he can… he's welcome to answer, or we can…

165
00:21:23.120 --> 00:21:26.979
Brady Hugins: You know, think this and speak this out to the end of the universe.

166
00:21:29.940 --> 00:21:38.000
Brady Hugins: So, oftentimes, when we look at our business, we might have multiple places where our information lives.

167
00:21:39.280 --> 00:21:40.839
Brady Hugins: And there may be…

168
00:21:42.040 --> 00:21:50.659
Brady Hugins: more accurate sources of that information other than others. You can think of, like, a will. Like, when you update a will or a trust, you have, like.

169
00:21:50.860 --> 00:22:06.909
Brady Hugins: the standard set that's submitted with the court, that's the canonical version, standardized version, and then everyone else's versioning after that. But because of the way the court process works, it keeps standard documentation of time, of information, of who…

170
00:22:07.090 --> 00:22:08.240
Brady Hugins: It affects.

171
00:22:12.630 --> 00:22:19.410
Brady Hugins: So, what happens when they disagree? So there's some sort of ruling. So, in that case, it would be the latest is the truest.

172
00:22:19.860 --> 00:22:26.049
Brady Hugins: Sometimes you can get new information that may not be the most true, or it's…

173
00:22:27.770 --> 00:22:33.570
Brady Hugins: mysterious. You know, you can… you can think that there's… might be…

174
00:22:33.670 --> 00:22:37.390
Brady Hugins: A pr- a user that might be a certain person, but it's not.

175
00:22:37.870 --> 00:22:45.550
Brady Hugins: So, they're… There may be some…

176
00:22:45.760 --> 00:22:51.780
Brady Hugins: need for connection between different systems in order for information to be more clear.

177
00:22:54.660 --> 00:22:59.840
Brady Hugins: If a job ran successfully but did nothing for a month, what'd you know? How?

178
00:23:01.710 --> 00:23:09.969
Brady Hugins: That's an important part, question for every business owner, right? A lot of people run their business based on cash flow.

179
00:23:10.450 --> 00:23:12.740
Brady Hugins: And they say, oh, if the cash is coming in.

180
00:23:12.960 --> 00:23:24.900
Brady Hugins: Everything's going alright, and to a certain degree, that's true, but to a certain degree, that's not. There's a lot more to success than just cash and income and financial statements.

181
00:23:25.080 --> 00:23:28.929
Brady Hugins: Yeah, impacted a business in the long way, those very things.

182
00:23:31.050 --> 00:23:44.800
Brady Hugins: So, if something was failing, and you didn't know about it and didn't get a report on it, it's important for us to check in and perhaps write a workflow for those, or make sure that we're interacting with

183
00:23:45.010 --> 00:23:47.300
Brady Hugins: Our systems, so that we're…

184
00:23:49.220 --> 00:23:56.120
Brady Hugins: Successfully understanding the… what's going on in the workflow, in our processes and business.

185
00:23:58.820 --> 00:24:06.570
Brady Hugins: Payments is a good example of, like, if you had a product that was selling and you didn't have any information about payments.

186
00:24:07.030 --> 00:24:10.819
Brady Hugins: For customer information, like, Delivery information.

187
00:24:14.330 --> 00:24:18.810
Franklin Sheffield: Do you… do you think that often…

188
00:24:18.930 --> 00:24:28.530
Franklin Sheffield: When a model is running quote-unquote successfully, but a company is not getting their desired end results?

189
00:24:28.720 --> 00:24:30.430
Franklin Sheffield: Do you think that…

190
00:24:31.120 --> 00:24:43.630
Franklin Sheffield: Like, what they probably should look at first, or should be high on the priority list, would be something along the lines of, like, broadening the goal, or adjusting the goal to be more inclusive of

191
00:24:44.170 --> 00:24:45.600
Franklin Sheffield: other ideas?

192
00:24:48.350 --> 00:24:51.970
Brady Hugins: It's… that's a good question. I think that…

193
00:24:52.780 --> 00:24:58.669
Brady Hugins: When standard businesses often rely on, like, looking at past information.

194
00:24:58.920 --> 00:25:03.270
Brady Hugins: But often, when a business has an issue or a problem.

195
00:25:03.920 --> 00:25:08.659
Brady Hugins: They'd need to look at old information in new ways, so they'd need to consider new factors.

196
00:25:10.130 --> 00:25:15.180
Brady Hugins: So what you're saying is true, and implies that it's true, is that…

197
00:25:16.210 --> 00:25:18.060
Brady Hugins: You know, how do you,

198
00:25:18.980 --> 00:25:31.640
Brady Hugins: if a business is functioning successfully, you know, how do you help change it? And oftentimes, it's through the motivation of either the owners or the values of the owners.

199
00:25:32.090 --> 00:25:34.810
Brady Hugins: That's almost easiest to align with.

200
00:25:38.720 --> 00:25:42.370
Brady Hugins: and the issues. And now…

201
00:25:43.210 --> 00:25:50.460
Brady Hugins: There's almost always… there's certainly a lot of issues going on with business, with change, with…

202
00:25:51.520 --> 00:25:54.340
Brady Hugins: the impact AI, how expensive some…

203
00:25:54.480 --> 00:25:57.750
Brady Hugins: tools are… I mean, I've looked at some of these

204
00:25:58.390 --> 00:26:01.420
Brady Hugins: Builder tools, and they're very expensive, so they're…

205
00:26:02.420 --> 00:26:07.010
Brady Hugins: These businesses need to adapt, and they need to change, and…

206
00:26:07.260 --> 00:26:20.180
Brady Hugins: adapt their companies accordingly. Do you have any specific examples of this kind of successful, company that you have in mind that is having a hard time adapting?

207
00:26:22.600 --> 00:26:30.050
Franklin Sheffield: Yeah, and I think I can do you one even better by offering, like, another kind of situational, maybe, challenge.

208
00:26:30.160 --> 00:26:35.790
Franklin Sheffield: So, like, imagine you have a company that's doing a Shopify online.

209
00:26:36.800 --> 00:26:42.630
Franklin Sheffield: like, dropshipping business. Everything's automated, marketing's automated, right, and they're making sales.

210
00:26:42.750 --> 00:26:45.850
Franklin Sheffield: But their sales are dropping over time.

211
00:26:45.990 --> 00:26:50.530
Franklin Sheffield: But the system is outputting success, because…

212
00:26:50.810 --> 00:26:57.949
Franklin Sheffield: You know, they… they're making money, and it, you know, like, it's all being automated, and it's generating its reports, and there haven't been any errors.

213
00:26:58.310 --> 00:27:11.279
Franklin Sheffield: And so now the client is thinking, well, okay, I want to reimagine the gold to be like, I want to make my customers happy, because maybe they've read some comments and they're identifying that people aren't necessarily happy with the product or something.

214
00:27:11.280 --> 00:27:18.890
Franklin Sheffield: So, like, how do they integrate that into a data system? That, like, vague idea of, like, I want to make people happier?

215
00:27:20.550 --> 00:27:26.060
Brady Hugins: That's… I would consider a somewhat common problem how you do that.

216
00:27:27.230 --> 00:27:37.599
Brady Hugins: With Shopify specifically, Shopify has a… and I'm saying this from a developer perspective, because this is a developer course, has an API,

217
00:27:38.240 --> 00:27:46.240
Brady Hugins: And where you can start interacting it with its database of information. Now, with Shopify, that's probably web analytics.

218
00:27:46.370 --> 00:27:52.980
Brady Hugins: Payment analytics, and various sorts of emails and whatnot going through the delivery of those products.

219
00:27:53.590 --> 00:28:11.770
Brady Hugins: So, you can start taking in that information into a central source of truth, and doing analysis on it independently with your own tools, or with whatever tools you'd like to work with, using Shopify as, like, an arm or leg of the business, rather than thinking of it as, like, the whole body of it.

220
00:28:13.400 --> 00:28:15.200
Brady Hugins: And that would create…

221
00:28:15.660 --> 00:28:24.649
Brady Hugins: Well, you would then be able to glean some of the information more independently from Shopify, because Shopify is a closed box where you can only do so much.

222
00:28:24.970 --> 00:28:34.729
Brady Hugins: I've worked with Shopify a little bit, and I've leaned away from it because of some back-end architecture issues, and because I split some of the payment issues out with the web, but…

223
00:28:35.080 --> 00:28:43.729
Brady Hugins: some people love the UI, and they want to interact with Shopify, so I'm like, do what you love, and do what's easiest for you, as long as it works for the general business, so…

224
00:28:44.320 --> 00:28:46.210
Brady Hugins: I'm really pretty neutral about it.

225
00:28:46.210 --> 00:29:00.120
Franklin Sheffield: Honestly, I don't love Shopify, but as a digital marketing professional, sometimes I have clients that want to be hands-on, and so I'm like, okay, now I need a CMS system that has, like, user-friendly qualities, but also gives me a bit of depth.

226
00:29:00.150 --> 00:29:06.609
Franklin Sheffield: So Shopify hits that sweet spot for me, where it's like, they can get in there and manually make a something.

227
00:29:06.820 --> 00:29:10.799
Franklin Sheffield: And not break everything, but I can also be in the code in the backend.

228
00:29:11.770 --> 00:29:12.430
Brady Hugins: Right.

229
00:29:12.670 --> 00:29:15.710
Brady Hugins: So… with Shopify.

230
00:29:16.310 --> 00:29:22.629
Brady Hugins: setting up an API, and you can do this with almost all modern tools that have these databases.

231
00:29:22.880 --> 00:29:33.119
Brady Hugins: you can have a dynamic API. Sometimes it's read-only, where you can only read what's going on in the Shopify area, and then you can do your own analysis on top of that.

232
00:29:33.500 --> 00:29:43.190
Brady Hugins: But, some of them, you can have interactions where you can have dynamically updated content. So, for instance, like we were talking about canonical source of truth.

233
00:29:43.400 --> 00:29:45.620
Brady Hugins: An idea of that would be, like.

234
00:29:45.750 --> 00:30:00.610
Brady Hugins: product pictures, right? Let's say you take some pictures of a product, and, like, a whole series of, like, 10 products, and you keep them on your central database. Now, you can have it fed out to Shopify automatically through the workflow and API,

235
00:30:00.960 --> 00:30:14.420
Brady Hugins: But you can also say, hey, Kay, we did a great job on Shopify, why don't we go and take that central database that we have, and let's open up an additional Amazon Marketplace shop.

236
00:30:14.540 --> 00:30:25.960
Brady Hugins: Let's say we get 10-15% more sales from this additional shop over here, but it's all being served from the same central canonical source of truth with database.

237
00:30:28.110 --> 00:30:33.629
Franklin Sheffield: That, that's essentially what I do for my clients. I call that the franchise era, but…

238
00:30:34.400 --> 00:30:41.910
Franklin Sheffield: I am the database right now. Like, I'm the guy who, like, keeps everything tiled, and, like, I've got terabytes of stuff at this point.

239
00:30:42.610 --> 00:30:50.949
Brady Hugins: So, you would benefit from setting up your own infrastructure and ecosystem like this as you're going and collecting on new

240
00:30:51.450 --> 00:30:59.960
Brady Hugins: Customers and… and, work as an agency, because there's two… there's two types of clients you're gonna run into.

241
00:31:00.750 --> 00:31:14.389
Brady Hugins: One, they're gonna want you to do it because they don't know how to do it, and they don't want to do it. So they're gonna need you to build your own tools so that they can hop on your back and have the result that you can create.

242
00:31:14.870 --> 00:31:25.249
Brady Hugins: And then, you're going to have the clients that want to build their own system, so they want to do what you're doing as far as having a centralized database, and they want to build it up for the long run.

243
00:31:25.510 --> 00:31:32.219
Brady Hugins: This series is built around the second person that's willing to build it up for a long run, but at the same time.

244
00:31:32.360 --> 00:31:42.279
Brady Hugins: It… this… we're… we're looking… at most people, they don't want to do the development work. They want someone else to do it. That's the reality of the situation.

245
00:31:42.400 --> 00:31:51.170
Brady Hugins: It's probably only 10-20%, maybe up to 30, 40 in select cases that are willing to do this kind of development work, I've found.

246
00:31:51.940 --> 00:32:01.379
Brady Hugins: It may be smaller, I don't know, but, we have to assume that the customers are going to ask you, Franklin, and say, hey, how do we do this, and what do we do about it?

247
00:32:08.580 --> 00:32:09.980
Brady Hugins: Any more questions?

248
00:32:13.470 --> 00:32:20.480
Franklin Sheffield: I did have one, about data redundancy, so this is more from the world of IT.

249
00:32:20.710 --> 00:32:29.380
Franklin Sheffield: So, like, from my understanding, like, there's a lot of commonalities between, like, a server environment and an AI agent database environment.

250
00:32:29.710 --> 00:32:37.249
Franklin Sheffield: One of the problems that you have in a server environment is, like, when the power goes out, you could lose everything.

251
00:32:37.250 --> 00:32:55.699
Franklin Sheffield: And so, they use a lot of strategies, but one of the strategies they use, and specifically, I could see being a nightmare for an AI situation. So, like, if an AI agent replaces the function of a hub or a server, and their job now becomes keep one-third of this

252
00:32:55.770 --> 00:33:01.419
Franklin Sheffield: data and send two-thirds of it to… or, like, keep half of this, make a copy, and send it

253
00:33:01.620 --> 00:33:09.440
Franklin Sheffield: So you create this, like, intricate redundancy system to protect the integrity of your data. Like, I feel like that could…

254
00:33:09.610 --> 00:33:28.950
Franklin Sheffield: probably be very complex, if not almost impossible to get right and do well in an environment where all of your servers are also trying to use that fragmented data to create, A, what's the most common and relevant and current data, but B, like, products with it.

255
00:33:31.150 --> 00:33:40.090
Brady Hugins: Yeah, that's a great call-out from understanding how AI is working on a systemic way, and how it looks like a server environment.

256
00:33:41.410 --> 00:33:57.769
Brady Hugins: What I found to help out with that is moving into a direction that doesn't seem like it's a pro… doesn't seem like it makes sense, but it does when you look at it, is actually cut the workflows down and simplify them more.

257
00:33:58.460 --> 00:34:01.099
Brady Hugins: Make them more modular-based.

258
00:34:01.210 --> 00:34:04.980
Brady Hugins: And more useful across multiple tools.

259
00:34:05.140 --> 00:34:09.860
Brady Hugins: Like, for instance, like, I have a PDF creator.

260
00:34:10.420 --> 00:34:12.219
Brady Hugins: That is a program

261
00:34:12.469 --> 00:34:24.289
Brady Hugins: That goes and scans across the branding database, or the branding information that it has, so it outputs the PDF in the appropriate format along the rules.

262
00:34:24.290 --> 00:34:35.519
Brady Hugins: Instead of saying, hey, I'm gonna make a builder for this one, a builder for that one, a builder for this one, I just say, I'm gonna make one builder, and it's gonna use that one for that brand, that one for that brand, that one for that brand.

263
00:34:36.030 --> 00:34:40.839
Brady Hugins: So, like you're saying, like, it's overly complex, I'm like, well…

264
00:34:41.560 --> 00:34:46.539
Brady Hugins: It is one, like, kind of like to set it up initially.

265
00:34:47.020 --> 00:34:53.980
Brady Hugins: Because of, like, the rules, but once those are set up, you don't really have to do…

266
00:34:54.159 --> 00:35:05.580
Brady Hugins: the work anymore. It always functions that way, and you can adapt and edit based on those. And you can go right into the markdowns and change individual things. Because it's a natural language model.

267
00:35:05.830 --> 00:35:08.410
Brady Hugins: You're working with things that are more directive.

268
00:35:08.570 --> 00:35:18.410
Brady Hugins: And you could say, like, go look over here and do this, because that's what the coding says, where you can go in there, change the coding that the AI is looking at, or whatever tool is looking at it.

269
00:35:19.030 --> 00:35:24.719
Brady Hugins: So, it's… Pretty fluid, what you can do.

270
00:35:25.370 --> 00:35:31.670
Brady Hugins: But modularization, I think, is the most effective because, as an example, I have these…

271
00:35:32.210 --> 00:35:45.029
Brady Hugins: Venus love letters that are… kind of from the outside, looks pretty simple. They're like a 5-page PDF output. It's personalized for astrology, but it's only one planet, but I have to do the entire astrology chart.

272
00:35:45.320 --> 00:36:02.320
Brady Hugins: look up their geocoding, with the active API database, and then output it towards a branded PDF within, like, 5 minutes, 3 minutes or less, because people expect PDFs to… to, like, arrive in their mailbox within a minute or less.

273
00:36:02.530 --> 00:36:10.770
Brady Hugins: So, that's a little challenging, to have something, like, hit off, like, every, like, minute, so you have to have a trigger that hits.

274
00:36:11.500 --> 00:36:12.500
Brady Hugins: And…

275
00:36:14.180 --> 00:36:22.209
Brady Hugins: So, what happened is that the whole workflow was pretty complex, and so I had to break it down into certain components.

276
00:36:22.410 --> 00:36:30.839
Brady Hugins: And then those components made my whole PDF generator process with just the Venus Love Letters as a component of it.

277
00:36:31.230 --> 00:36:37.630
Brady Hugins: So, it's like systemizing and modularizing it, but what I found also is that

278
00:36:38.210 --> 00:36:53.610
Brady Hugins: doing the work and trying it is actually the most helpful thing you can do to figure it out, is actually just to attempt to do it. Because you don't always know, kind of, what you're gonna run into, or what,

279
00:36:55.340 --> 00:37:01.310
Brady Hugins: bottlenecks. I've noticed that automation is all about For bottlenecking.

280
00:37:01.790 --> 00:37:05.120
Brady Hugins: And then smoothing out bottlenecking is a lot of that.

281
00:37:11.280 --> 00:37:12.230
Brady Hugins: Okay.

282
00:37:14.210 --> 00:37:17.590
Brady Hugins: Good question. Thank you, Franklin.

283
00:37:17.830 --> 00:37:22.810
Brady Hugins: So, next Friday, we are going to have a Chapter 7.

284
00:37:23.020 --> 00:37:29.340
Brady Hugins: Owning the system forever, so what that looks like, from a rental to an ownership model.

285
00:37:30.210 --> 00:37:33.210
Brady Hugins: This is part of a 7-day series course.

286
00:37:33.390 --> 00:37:41.390
Brady Hugins: $297, and it gets you the… it provides the ability to build solo in a self-paced way.

287
00:37:41.700 --> 00:37:50.690
Brady Hugins: We'll be updating this yearly, but also having some of these classes regularly as we adapt and evolve with some of the systems.

288
00:37:51.390 --> 00:38:04.790
Brady Hugins: We do have a live cohort that's available as well. This will be on a monthly basis, and then we'll have a guided series, in the future. Those will be on Fridays as well, real-life accountability.

289
00:38:05.600 --> 00:38:20.930
Brady Hugins: And then a live office hours, once a month, in accompaniment with some of our data sovereignty membership. Our data sovereignty membership is an alternative to the course, the sense that you'll receive the entire course, but just for $33 a month.

290
00:38:22.020 --> 00:38:24.430
Brady Hugins: So, we appreciate you being here.

291
00:38:25.420 --> 00:38:31.100
Brady Hugins: That concludes for today, so… happy Friday, everyone. Thank you.

292
00:38:32.890 --> 00:38:33.750
Franklin Sheffield: fake view.

293
00:38:34.450 --> 00:38:35.380
Brady Hugins: Alright.

294
00:38:37.700 --> 00:38:39.570
Brady Hugins: Gonna stop recording…