Meeting Title: Brainforge x Paint Rollers Project Sync Date: 2025-08-25 Meeting participants: Robert Tseng, Casey Isaac


WEBVTT

1 00:05:53.510 00:05:54.330 Casey Isaac: Roberts.

2 00:05:56.530 00:05:57.559 Robert Tseng: Hey, Casey!

3 00:05:57.780 00:06:00.190 Casey Isaac: Sorry, man, for the delay.

4 00:06:00.830 00:06:01.740 Robert Tseng: Hey, all good.

5 00:06:02.950 00:06:04.119 Casey Isaac: Had a, …

6 00:06:05.730 00:06:11.459 Casey Isaac: Interesting morning, I had a flat tire on the way to work, so I’m in a coffee shop right now, waiting on my car.

7 00:06:12.030 00:06:13.260 Casey Isaac: Oh.

8 00:06:13.260 00:06:14.969 Robert Tseng: LA, LA problems.

9 00:06:15.320 00:06:24.190 Casey Isaac: It’s so annoying, man. I’m like, I haven’t driven in so long, I forgot what towing a car is like. There’s, like, no spare and stuff.

10 00:06:24.190 00:06:25.120 Robert Tseng: Oh!

11 00:06:25.350 00:06:29.490 Casey Isaac: I either, like, have a tone, and then… ….

12 00:06:29.490 00:06:30.170 Robert Tseng: Yeah.

13 00:06:30.290 00:06:31.480 Casey Isaac: So anyway, sorry I’m late.

14 00:06:31.790 00:06:40.070 Robert Tseng: No, all good. I was just drafting an email, I was gonna send it to you, because I said I was gonna send you some, some docs for you to look at, so I apologize if I didn’t get it to you earlier.

15 00:06:40.300 00:06:44.250 Casey Isaac: Don’t worry at all. How was the, … tennis.

16 00:06:44.510 00:06:55.209 Robert Tseng: Yeah, the open was good. Yeah, well, my friend lost, and … I did end up going again on Saturday with my wife, so I think that’s… that’s it for us this year, though.

17 00:06:55.850 00:06:57.030 Robert Tseng: Cool. Yeah. Yeah.

18 00:06:57.280 00:07:00.420 Casey Isaac: I’ve always wanted to… Check it out. ….

19 00:07:01.140 00:07:01.690 Robert Tseng: Yeah.

20 00:07:01.690 00:07:05.510 Casey Isaac: Looks like such a fun time, so, very chilling.

21 00:07:05.510 00:07:07.730 Robert Tseng: Tennis Disneyland, so… that’s what they say.

22 00:07:07.730 00:07:08.600 Casey Isaac: Yeah, totally.

23 00:07:09.240 00:07:10.120 Casey Isaac: Totally.

24 00:07:11.380 00:07:15.630 Casey Isaac: So, alright, so I don’t know what would be helpful for you.

25 00:07:16.560 00:07:17.510 Robert Tseng: Yeah, why don’t you just….

26 00:07:17.510 00:07:18.980 Casey Isaac: If you could just….

27 00:07:18.980 00:07:28.279 Robert Tseng: I know you told me about Sage 100, I was looking at their developer docs earlier today, actually, so, yeah, they have an API, I kind of understand, like, how they’re… I mean.

28 00:07:28.470 00:07:30.290 Robert Tseng: I think I understand how they’re…

29 00:07:30.610 00:07:36.470 Robert Tseng: their system works, and … I know, based on our conversation on Friday, you were

30 00:07:36.750 00:07:54.069 Robert Tseng: basically trying to pull data out of it, and then, like, you didn’t really like the formatting of some of the stuff, so, maybe we could streamline some workflows, like, I think that’d be cool. Like, it seems like there’s a lot of good opportunity to kind of make things more, automated.

31 00:07:55.190 00:08:03.000 Casey Isaac: For sure, yeah, and just to, I guess, to clarify on the Sage 100 thing, because I think there’s a cloud version that is more…

32 00:08:04.490 00:08:12.400 Casey Isaac: Open has APIs that… like, easier to connect into. I’m on, like, an old… Like, an on-premise version.

33 00:08:12.400 00:08:13.509 Robert Tseng: Right, the on-prem one, right.

34 00:08:13.510 00:08:18.149 Casey Isaac: Yeah, so… It’s not the easiest, I don’t think, to…

35 00:08:18.820 00:08:23.010 Casey Isaac: To integrate with, but that said, … I’m sure…

36 00:08:23.560 00:08:29.939 Casey Isaac: Yeah, we’ll work it out, but I can take you through… this, like… Let me show you…

37 00:08:30.980 00:08:34.189 Casey Isaac: Just an example of a report that…

38 00:08:36.610 00:08:38.349 Casey Isaac: Let’s take you through here, though.

39 00:08:38.780 00:08:41.430 Casey Isaac: And my workflow, signing in and stuff, and then….

40 00:08:41.799 00:08:42.839 Robert Tseng: Yeah. ….

41 00:08:43.980 00:08:48.429 Casey Isaac: So I’m accessing this server remotely, and…

42 00:08:50.710 00:08:52.880 Casey Isaac: I, I’ll pull, like, a…

43 00:08:53.520 00:09:04.760 Casey Isaac: I don’t know, one of the… Other reports, do some… What’s the… Maybe, like, … Sales order report.

44 00:09:05.460 00:09:06.480 Casey Isaac: …

45 00:09:09.580 00:09:12.399 Casey Isaac: The problem that I have is basically…

46 00:09:12.980 00:09:15.780 Casey Isaac: Here we go. Sorry, I’m still kind of getting used to this.

47 00:09:16.730 00:09:22.170 Casey Isaac: Not the most intuitive, but … okay, so let’s just be… So…

48 00:09:22.390 00:09:27.460 Casey Isaac: Here’s just, like, one report that I would pull… We’ll just do all customers.

49 00:09:27.960 00:09:37.530 Casey Isaac: … the prior year, and COGS… And… I’ll show you what happens.

50 00:09:40.880 00:09:45.459 Casey Isaac: I think, like, there’s a few things I’d love to talk through with you. I mean, first of all.

51 00:09:45.640 00:09:50.890 Casey Isaac: I had this vision of this, like, Being able to… I don’t know.

52 00:09:51.030 00:09:51.970 Casey Isaac: integrate.

53 00:09:52.110 00:09:55.280 Casey Isaac: you know, an LLM, and be able to ask my data questions.

54 00:09:55.810 00:10:02.050 Casey Isaac: Yeah. And just, like, I don’t know, pull up a bill of materials, or a certain, you know…

55 00:10:02.250 00:10:06.859 Casey Isaac: SKU that we sell, stuff like that. Totally. It just returns it easily.

56 00:10:07.000 00:10:09.220 Casey Isaac: That’s, like, what I would love.

57 00:10:09.450 00:10:11.479 Casey Isaac: But let me just show you how I’m…

58 00:10:11.890 00:10:18.090 Casey Isaac: some of the struggles I’m having, and I’m not very good in Excel, I really haven’t dealt with any kind of

59 00:10:18.640 00:10:20.840 Casey Isaac: Data, or in our analytics.

60 00:10:21.090 00:10:26.150 Casey Isaac: You know… In past roles, so…

61 00:10:26.280 00:10:33.240 Casey Isaac: This is what I have to do. I gotta save… District Court… And then…

62 00:10:34.250 00:10:36.300 Casey Isaac: I gotta email it to myself.

63 00:10:39.440 00:10:47.640 Casey Isaac: And I can, like, This is a kind of silly workflow that’s… it’s a little bit of a… Yes.

64 00:10:48.990 00:10:51.309 Casey Isaac: Band-Aid fix, just to give me access to…

65 00:10:51.650 00:10:54.060 Casey Isaac: Reporting for now, but this is what I have to do.

66 00:10:54.560 00:11:00.479 Casey Isaac: … And then I’ll send it to myself, and I’ll show you once I get it here.

67 00:11:01.000 00:11:01.780 Casey Isaac: That means….

68 00:11:02.140 00:11:02.750 Robert Tseng: Yeah.

69 00:11:02.950 00:11:04.149 Casey Isaac: Stop sharing.

70 00:11:04.520 00:11:07.229 Casey Isaac: I’ll pull up that report that I just set myself.

71 00:11:07.630 00:11:08.689 Casey Isaac: To show you now.

72 00:11:09.190 00:11:11.560 Casey Isaac: Why it’s so difficult for me to, like.

73 00:11:12.840 00:11:16.100 Casey Isaac: manipulate stuff. Like, I would… I would love to see…

74 00:11:16.410 00:11:17.989 Casey Isaac: So I’m gonna share once more.

75 00:11:18.350 00:11:26.090 Casey Isaac: is… I have that report now, okay, so… … Second one.

76 00:11:29.860 00:11:31.349 Casey Isaac: So here’s that report.

77 00:11:31.570 00:11:32.290 Casey Isaac: Okay.

78 00:11:33.330 00:11:37.469 Casey Isaac: I’d love to just be able to sort and, you know, kind of miss….

79 00:11:37.470 00:11:37.999 Robert Tseng: I only see….

80 00:11:38.000 00:11:38.470 Casey Isaac: Data.

81 00:11:38.470 00:11:40.510 Robert Tseng: Top part of your browser.

82 00:11:41.360 00:11:42.000 Casey Isaac: Weird.

83 00:11:44.000 00:11:47.140 Casey Isaac: … Let’s try that again.

84 00:11:50.110 00:11:50.980 Casey Isaac: How’s that?

85 00:11:51.190 00:11:52.220 Robert Tseng: Oh yeah, there we go.

86 00:11:52.610 00:11:53.360 Casey Isaac: Okay, cool.

87 00:11:53.900 00:11:54.980 Casey Isaac: So…

88 00:11:55.290 00:12:01.170 Casey Isaac: as you can see here, I… you know, it’s broken out by period and stuff, cool, but what I really like to do is just be able to…

89 00:12:02.020 00:12:11.820 Casey Isaac: Like, there’s no way… … No way to just kind of get aggregate year-to-date sales.

90 00:12:12.020 00:12:20.429 Casey Isaac: And to be able to, like, sort that by, you know, ascending, basically see, you know, who our top customers are year-to-date.

91 00:12:20.750 00:12:23.159 Casey Isaac: Same thing goes for any other report.

92 00:12:23.380 00:12:31.650 Casey Isaac: You know, if I’m trying to understand what our top-selling SKUs are, Can’t easily do that. …

93 00:12:32.230 00:12:34.559 Casey Isaac: you know, I’d like to be able to do that.

94 00:12:36.020 00:12:41.279 Casey Isaac: Year-to-date, by period, … I’d love to be able to…

95 00:12:41.940 00:12:50.370 Casey Isaac: Just kind of quickly and easily understand You know, who are… Pot customers are.

96 00:12:50.540 00:12:51.670 Casey Isaac: Stuff like that.

97 00:12:51.920 00:12:56.369 Casey Isaac: And then… You know, so that’s, like, the most basic.

98 00:12:57.200 00:12:59.139 Casey Isaac: At the most basic level, what I need.

99 00:12:59.240 00:13:09.259 Casey Isaac: But beyond that, yeah, like I said, I’d like to be able to just… like, I do a lot of just basically pulling reports, putting it into ChatGPT, and then asking the reports questions.

100 00:13:09.480 00:13:11.690 Robert Tseng: Yeah. I would love if I could just get….

101 00:13:11.690 00:13:15.610 Casey Isaac: all of my data from Sage, …

102 00:13:15.980 00:13:21.680 Casey Isaac: kind of integrated with an LLM to basically be able to ask it whatever question

103 00:13:22.010 00:13:28.040 Casey Isaac: You know, it’d be really cool if I could have, kind of, visualization, stuff like that.

104 00:13:28.240 00:13:33.039 Casey Isaac: … You know, be able to ask you to build graphs, or whatever.

105 00:13:33.270 00:13:42.480 Casey Isaac: So that, that’s more of, I would say, … A longer term, aspirational, …

106 00:13:42.680 00:13:49.079 Casey Isaac: But… but today, I just need… To be able to… generate clean reports. ….

107 00:13:49.080 00:13:49.750 Robert Tseng: Yeah.

108 00:13:49.750 00:13:51.709 Casey Isaac: So… Hopefully that helps.

109 00:13:51.920 00:13:52.440 Casey Isaac: Yeah.

110 00:13:52.440 00:13:52.810 Robert Tseng: Totally.

111 00:13:52.810 00:13:56.800 Casey Isaac: I can speak to, I can share some examples of the most common reports, if that’d be helpful.

112 00:13:56.960 00:13:58.990 Casey Isaac: that I’m pulling, day to day.

113 00:13:59.370 00:14:05.479 Robert Tseng: Yeah, why don’t you show me, like, something that you… that you pull, and then, I can also…

114 00:14:05.610 00:14:11.559 Robert Tseng: I can show you some things that we’ve built. We… what you’re describing is…

115 00:14:11.750 00:14:15.430 Robert Tseng: you know, I guess, as you’re pulling it up, my, my back, like.

116 00:14:15.980 00:14:27.860 Robert Tseng: I started this company… I guess I started consulting because I was, like, I’ve been doing data at multiple brands now, thought I could just kind of do it at a bigger scale. …

117 00:14:27.860 00:14:36.740 Robert Tseng: Yeah, I’ve been… I’ve been working on Brainforge. I merged with my business partner, like, a year ago, and so… but we’ve been working together on a couple clients before then.

118 00:14:36.740 00:14:44.880 Robert Tseng: So he was purely doing AI work, and I was doing purely data work. And then we worked on a couple clients together, and we realized that, like.

119 00:14:44.930 00:14:46.170 Robert Tseng: data…

120 00:14:46.240 00:14:53.090 Robert Tseng: the data work that I do kind of, like, leads into the AI… leads into AI work, and so what I mean by that is

121 00:14:53.180 00:15:07.749 Robert Tseng: Yeah, you know, I think people already kind of have an intuitive sense for, like, how data flows work, where you just take files or images and you throw it into an LLM, and then you’re able to… you’re able to chat with it with, with relatively little context.

122 00:15:07.760 00:15:14.810 Robert Tseng: Problem is there’s hallucinations and things like that, and if you use images, OCR is not super reliable, and so…

123 00:15:15.010 00:15:23.090 Robert Tseng: Actually, like, in order to get the consistency to a place where it’s actually usable, you do need to have, like, good data

124 00:15:23.470 00:15:31.179 Robert Tseng: hygiene, pretty much, which is just, like, making sure that all the data is clean in a single place, and in a structured way.

125 00:15:31.460 00:15:49.629 Robert Tseng: where that… the schemas are very easy to understand, and so an LLM can easily scan through it and be able to pick out the right pieces. So you’re just, like, eliminating… because even a human would have a hard time, kind of, like, figuring out, like, what were the right places to pull if you don’t have it set up. And so, yeah, we went from, like.

126 00:15:50.050 00:15:58.480 Robert Tseng: okay, well, maybe before AI work was just, yeah, just chuck into Jack’s GBT and kind of hope for the best, but then now it’s like.

127 00:15:58.610 00:16:16.559 Robert Tseng: I do… I do the data prep for… for clients, and they have everything kind of sitting in a warehouse, or some sort of, like, storage… storage place, it’s clean. And then we hook up, like, an LLM to it, and then the performance is way better. And, that starts to make, you know, AI a lot more usable in production.

128 00:16:16.560 00:16:26.520 Robert Tseng: And so we’ve kind of been able to do a lot of cool stuff with different clients, and then we also use… we build tools for ourselves internally. So, I could show you some of the stuff that we’ve built for our, for us, and…

129 00:16:26.520 00:16:31.499 Robert Tseng: Kind of how that is pretty… maybe transferable to kind of what you’re describing as well.

130 00:16:33.480 00:16:43.959 Casey Isaac: Awesome, yeah, I, and that was kind of what I was thinking. I don’t know if step one would be to kind of get this and just get all my data into a database that, you know, could be…

131 00:16:44.680 00:16:51.080 Casey Isaac: well, easier to integrate with other tools. How would… how would… I guess, at Brainforge, like, how do you guys…

132 00:16:51.490 00:16:55.299 Casey Isaac: Compared to, I don’t know, just using Snowflake, sort of…

133 00:16:55.580 00:16:59.939 Casey Isaac: Is it called, like, Cortex or something, or their LLM extension?

134 00:17:00.130 00:17:06.449 Casey Isaac: Yeah. I understand, I have no… I actually have no experience, really, even working with a snowflake or anything like that.

135 00:17:06.450 00:17:08.709 Robert Tseng: Sure. So… Yeah, we’re like the.

136 00:17:08.710 00:17:10.439 Casey Isaac: Yeah, I don’t dangerous in those both.

137 00:17:10.440 00:17:30.439 Robert Tseng: So, yeah, like, we… I mean, that is our preferred ecosystem. It is, like, the cheapest and fastest to get started with, and we have implemented Cortex with companies, and so, yeah, once you get into a place, they basically have already, like, hooked up an LLM, and you can just, like, ask questions to it directly. So, we do like Snowflake, if that’s kind of your preference.

138 00:17:30.550 00:17:33.350 Robert Tseng: yeah, I don’t really know. It depends on, kind of, like.

139 00:17:33.530 00:17:39.870 Robert Tseng: before I, like, select a database for you, or a data warehouse for you, I probably would want to better understand

140 00:17:40.080 00:17:58.209 Robert Tseng: You know, like, healthcare and finance companies, they can’t use, Snowflake because it’s not compliant. Snowflake won’t sign BAAs or anything, and so they… they have to go Microsoft, and they have to use Azure. Bigquery, you know, if your system is already mostly Google.

141 00:17:58.530 00:18:17.250 Robert Tseng: kind of for your admin and everything else, that it might make sense to stay in the Google ecosystem, because it’s a lot more flexible that way. So yeah, there’s, like, some questions that we don’t necessarily have to talk through on this call. I could just send you a list of questions that you can kind of just go through, and that’ll help me, like, recommend, like, which one to go with.

142 00:18:17.310 00:18:20.309 Robert Tseng: But yeah, I think that’s… those are just kind of…

143 00:18:20.460 00:18:24.099 Robert Tseng: Some of the considerations when you’re choosing one of these providers.

144 00:18:24.520 00:18:25.160 Casey Isaac: Cool.

145 00:18:25.290 00:18:29.429 Casey Isaac: Yeah, sounds good. We just make paint rollers, so, you know, it’s, like, pretty…

146 00:18:29.670 00:18:34.559 Casey Isaac: There’s no… as today, there’s no direct-to-consumer

147 00:18:34.840 00:18:43.990 Casey Isaac: element, everything is, so there’s no, kind of, consumer data or considerations there around protections. There’s, like, …

148 00:18:44.160 00:18:45.000 Casey Isaac: you know.

149 00:18:45.840 00:18:48.950 Casey Isaac: pretty standard-looking manufacturing company. I think.

150 00:18:49.220 00:18:50.520 Robert Tseng: Yeah. ….

151 00:18:50.840 00:18:57.789 Casey Isaac: you know, you mentioned somebody in the pool space or something that I think would probably be pretty similar.

152 00:18:58.270 00:18:58.780 Robert Tseng: Yeah.

153 00:18:58.960 00:18:59.550 Robert Tseng: Okay.

154 00:18:59.550 00:19:01.189 Casey Isaac: actual supply, I think you mentioned.

155 00:19:01.310 00:19:02.990 Robert Tseng: Yeah. We do….

156 00:19:03.290 00:19:04.699 Casey Isaac: Yeah, …

157 00:19:04.970 00:19:09.689 Casey Isaac: Whatever your questions… questions you have, though, if you want to send over, if you have, like, a template or something.

158 00:19:09.840 00:19:14.519 Casey Isaac: The… I’d love to understand a little bit more around your process, so just, like.

159 00:19:14.790 00:19:20.640 Casey Isaac: how we start working together, … Yeah. Do you have any kind of guidance on costs, or at least the structure?

160 00:19:20.870 00:19:26.320 Casey Isaac: Yeah. You know, what that looks like, you know, whether it’s kind of project-based.

161 00:19:26.760 00:19:32.089 Casey Isaac: Do you have, like, a subscription component? Why don’t… yeah, tell me just a little bit more about…

162 00:19:32.950 00:19:37.350 Casey Isaac: Are you, … Your process, be good.

163 00:19:38.220 00:19:51.709 Robert Tseng: Okay, yeah, so I guess how we work together… so there’s multiple directions. I think probably for this, it seems pretty straightforward. I mean, I think I would just want to, you know, if there are any other data sources, kind of figuring out how many data sources we need to integrate.

164 00:19:51.760 00:20:02.759 Robert Tseng: But assuming, let’s just say, it’s one ERP, kind of, like, pulling data out of it and making it usable, and we basically build out the full ETL, which is extracting it from

165 00:20:02.760 00:20:09.849 Robert Tseng: from the ERP, putting it into a storage, being able to transform it so that it’s ready for whatever reporting that you need.

166 00:20:09.850 00:20:27.049 Robert Tseng: and let’s say we load… we load it into, like, you know, a couple of your most important reports. That, I could easily, like, scope fixed. And, typically, like, our pricing when working with clients, we start at 5K. So it’s just, like, that, and we would just be able to run that through end-to-end.

167 00:20:27.140 00:20:28.339 Robert Tseng: If it’s, like.

168 00:20:28.340 00:20:51.830 Robert Tseng: oh, do you want to do more discovery, and you’re not exactly sure what reports you want to build in, then we can go hourly, and then we can… we can discuss, like, kind of hourly rates there. We have hourly with no minimum, and I can send you, like, kind of a rate card for what that looks like, or if we want to do retainer, then we could do that as well. But yeah, I think those are typically the starting points that we typically, go with.

169 00:20:51.990 00:20:54.579 Robert Tseng: And then, as far as, like.

170 00:20:54.690 00:21:04.880 Robert Tseng: timeline and how long this takes. Yeah, for fixed… for fixed scope, we would run it through within, like, within a month, so two to four weeks. …

171 00:21:05.120 00:21:06.470 Robert Tseng: Yeah, so…

172 00:21:06.650 00:21:17.369 Robert Tseng: I think there’s… the range is really just a matter of, like, how fast it is for us to get access to things. Less on the engineering side, I think that’s more predictable.

173 00:21:18.100 00:21:21.660 Robert Tseng: … And then… yeah, I think…

174 00:21:22.070 00:21:38.619 Robert Tseng: yeah, that’s kind of how we started with pool parts. Like you said, they’re a distributor. They also manufacture their own products as well. We basically stood up infrastructure for them, and I can share an architecture diagram with you, but it’s basically… we threw all their data into Snowflake, and then…

175 00:21:38.730 00:21:50.049 Robert Tseng: From there, they wanted some reporting. They decided not to go for one of the bigger known BI tools because, you know, they’re a pretty small team, both

176 00:21:50.190 00:22:04.690 Robert Tseng: both execs to, like, former bankers, they really liked Excel and pivot tables, so we just, plugged in a tool for them called Real. That was, like… and it works very well for, somebody who’s used to viewing data that way.

177 00:22:04.750 00:22:13.629 Robert Tseng: But yeah, I would say the BI, or the business intelligence part, figuring out what’s your preferred method of consuming reports is something that

178 00:22:13.640 00:22:24.920 Robert Tseng: you know, probably after one or two conversations, I would be able to recommend, like, what’s a, you know, what’s the tool for you, and you can… we can give you some demos and see what… what you prefer to interact with.

179 00:22:25.170 00:22:35.239 Robert Tseng: But yeah, that’s people, like, what I consider to be the… the last mile, that’s the easiest to stand up. Like, once the engineering work kind of kicks off, like, the…

180 00:22:35.320 00:22:50.719 Robert Tseng: the visualization and selecting what the reporting tool would be, can happen in parallel, and that is usually, like, just one or two weeks to get up. So, yeah, I think that’s kind of, from a high level, like, how

181 00:22:51.530 00:22:53.120 Robert Tseng: I could see this going.

182 00:22:55.330 00:22:56.140 Casey Isaac: Love it.

183 00:22:56.140 00:22:56.650 Robert Tseng: gap.

184 00:22:56.650 00:22:58.890 Casey Isaac: did, I think he…

185 00:23:00.190 00:23:12.649 Casey Isaac: Yeah, I would love to, just send over whatever questions you have, can get back to you immediately. This is a priority, I want to get it figured out sooner rather than later.

186 00:23:13.940 00:23:14.520 Robert Tseng: Yeah.

187 00:23:14.680 00:23:18.200 Casey Isaac: I, you know, we’re, we’re lean.

188 00:23:18.350 00:23:24.500 Casey Isaac: We don’t… Of any kind of budget for this, so, … But I, you know.

189 00:23:24.920 00:23:31.559 Casey Isaac: I’m hoping it’s not a super complex project, and the price tag is somewhere, you know, around that $5,000 minimum.

190 00:23:31.880 00:23:34.120 Casey Isaac: that you mentioned, …

191 00:23:34.880 00:23:42.060 Casey Isaac: The… I did have one other question for you, Robert. Do you do anything in the way… have you guys dabbled in automation at all, workflow automation?

192 00:23:42.400 00:23:43.900 Robert Tseng: Yeah. And….

193 00:23:44.600 00:23:47.850 Casey Isaac: That can kind of speak to…

194 00:23:48.120 00:23:52.860 Casey Isaac: what I’m thinking there, a little bit. Basically, we get a lot of orders by email.

195 00:23:53.080 00:23:57.479 Casey Isaac: Order comes through. It’s, like, in the email, or it’s in an attachment.

196 00:23:57.770 00:24:09.269 Casey Isaac: would love for, I don’t know, ChatGPT or Claude or whatever to analyze the email, say, hey, does this look like a sales order? Yes, no. If yes, and kind of route that to…

197 00:24:09.740 00:24:13.970 Casey Isaac: a folder, save it as a PDF or something.

198 00:24:14.240 00:24:17.890 Casey Isaac: Or write that order into a spreadsheet, and then have…

199 00:24:18.050 00:24:21.149 Casey Isaac: Basically, all of the email orders that come through.

200 00:24:21.320 00:24:28.009 Casey Isaac: in any given day, uploaded into our ERP system as orders, sales orders.

201 00:24:28.370 00:24:35.060 Casey Isaac: And, you know, obviously we don’t have to, like, connect UC stuff in there or something, but I’m wondering if that’s…

202 00:24:35.250 00:24:38.260 Casey Isaac: Sounds like something that’s kind of within your wheelhouse.

203 00:24:38.450 00:24:39.320 Casey Isaac: And….

204 00:24:39.320 00:24:39.880 Robert Tseng: Yeah, totally.

205 00:24:39.880 00:24:42.670 Casey Isaac: If so, I’d love to consider that, too.

206 00:24:43.640 00:24:55.130 Robert Tseng: Okay, yeah, yeah, as far as, like, message categorization or anything, like, we do a lot of that. I’ll just, I’ll just share something with you, so I’ll share my screen.

207 00:24:55.250 00:25:00.590 Robert Tseng: So, this is, like, an internal platform that we build for our teams, and so…

208 00:25:00.760 00:25:13.879 Robert Tseng: every… every Zoom that kind of we… every meeting that we get, there’s, gets pulled in. We have transcripts that get categorized by clients, or, like, and also, if, you know, we click into them, we…

209 00:25:13.880 00:25:20.980 Robert Tseng: We are able to automatically create tickets that go into our engineering, kind of, workflow management system. It’s called Linear.

210 00:25:20.980 00:25:27.990 Robert Tseng: We have email drafts already with different templates, so that if we want to do… if we want to trigger responses directly from here.

211 00:25:28.010 00:25:36.500 Robert Tseng: Or be able to, like, send a follow-up to somebody on a call, we just can do it from here. And then internally, we work in Slack, and so, kind of, similar thing. And so…

212 00:25:36.550 00:25:40.429 Robert Tseng: Yeah, we’ve done a lot with, like, being able to take

213 00:25:40.700 00:25:57.709 Robert Tseng: unstructured, like, national language, and, you know, whether it’s coming in from email, or meetings, or Slack, or wherever it is, and then being able to package it and move it around to different places. So, yeah, I mean, we… I basically live out on this every day, and this is something we’ve built for ourselves.

214 00:25:57.890 00:25:58.870 Robert Tseng: …

215 00:25:58.960 00:26:06.750 Robert Tseng: And we have a bunch of other, kind of, like, demos, things that we’ve built for clients. I think kind of what you were describing

216 00:26:06.770 00:26:19.179 Robert Tseng: especially if you’re trying to do, like, some sort of, like, document search on something that’s more complicated. I know we didn’t necessarily talk about, like, invoices or anything with a lot of line items. Like, we’ve built out, like, …

217 00:26:19.310 00:26:36.259 Robert Tseng: this is, like, a medical record analysis tool where, you know, this is, like, 50 to 100 pages. We’re able to come and extract the most relevant things that, like, for, like, a… like, a personal injury law firm or whatever, that they would want to know, so that they can qualify whether or not that’s a case that they would take on.

218 00:26:36.260 00:26:44.490 Robert Tseng: So that’s, like, an example of things that we’ve done. So, just being able to fine-tune these LLMs to be able to

219 00:26:44.720 00:26:58.659 Robert Tseng: kind of get the data that you want out of it consistently. We’ve done a lot of different things like that. So, yeah, happy to chat further about anything, like, workflow automation related, because that’s definitely in our warehouse.

220 00:26:58.960 00:27:00.889 Casey Isaac: Cool, yeah, that’s great, …

221 00:27:01.040 00:27:05.939 Casey Isaac: I’d love to… yeah, I can share more if you had questions on that, or if you have, like, a…

222 00:27:06.150 00:27:09.439 Casey Isaac: Yeah, yeah, and follow up, if you want to just include

223 00:27:10.100 00:27:14.870 Casey Isaac: Some of the qualification-type stuff around… Automations?

224 00:27:14.980 00:27:15.640 Casey Isaac: ….

225 00:27:15.640 00:27:16.160 Robert Tseng: Yeah.

226 00:27:16.470 00:27:18.169 Casey Isaac: I can get back to you on that, too.

227 00:27:18.860 00:27:21.869 Casey Isaac: Again, Yeah, you know, I think…

228 00:27:22.370 00:27:25.399 Casey Isaac: As far as systems we’re integrating into, it’s really…

229 00:27:26.630 00:27:35.170 Casey Isaac: It’s probably just, like, our email… And… our saved software.

230 00:27:35.580 00:27:38.689 Casey Isaac: We don’t have Slack today. I would like to set up Slack.

231 00:27:38.950 00:27:43.090 Casey Isaac: But, … That would probably be…

232 00:27:43.440 00:27:48.199 Casey Isaac: it. And then, you know, whatever kind of data warehouse or whatever.

233 00:27:48.790 00:27:49.410 Robert Tseng: Yeah.

234 00:27:49.410 00:27:50.569 Casey Isaac: that we choose.

235 00:27:51.020 00:27:54.070 Casey Isaac: So anyways, yeah, that’s… hopefully that…

236 00:27:55.340 00:27:59.390 Casey Isaac: gives you all you need, to get working on something, but I, ….

237 00:27:59.390 00:28:00.040 Robert Tseng: Yeah.

238 00:28:00.390 00:28:07.419 Casey Isaac: I definitely… yeah, I want to keep this going, because I’m, like, really struggling right now to understand

239 00:28:08.070 00:28:10.429 Casey Isaac: Much of anything, so… sweet.

240 00:28:10.430 00:28:11.140 Robert Tseng: Sure.

241 00:28:11.140 00:28:12.509 Casey Isaac: how things are set up.

242 00:28:13.180 00:28:18.689 Robert Tseng: Yeah, I can imagine you coming from tech, and then going back towards, like, I don’t know, like a…

243 00:28:19.640 00:28:34.439 Robert Tseng: like, an older kind of system. You have all these, like, oh, it should just already be this way, and it’s, like, kind of takes a while to kind of bring it up. I mean, those are honestly the best clients for us, because we love doing this kind of work, and it’s, like.

244 00:28:34.550 00:28:42.239 Robert Tseng: yeah, like, the… the ROI on it is, like, very visible, so we do like working with, kind of…

245 00:28:42.250 00:28:58.469 Robert Tseng: companies who are really just trying to, like… it’s basically like a digital transformation kind of, like, exercise, like, how do you kind of catch someone up to kind of this… this AI kind of native era, right? So, yeah. And it’s crazy, like, I’m in this stuff every day, and

246 00:28:58.670 00:29:04.400 Robert Tseng: you know, even if a system doesn’t have API anymore, if you just have a repetitive process to bring to get data.

247 00:29:04.400 00:29:04.990 Casey Isaac: Yeah.

248 00:29:04.990 00:29:08.120 Robert Tseng: it’s just so good, you can just… you can automate it. So, like.

249 00:29:08.120 00:29:20.530 Casey Isaac: That’s the beautiful thing about AI. Yeah, it’s like, you can… I’m imagining, like, I… if I stepped into this role a few years ago, I would be like, oh man, we gotta upgrade our ERP immediately.

250 00:29:20.530 00:29:21.879 Robert Tseng: Yeah. But… They don’t really.

251 00:29:21.880 00:29:30.660 Casey Isaac: It seems like you can now just kind of extend the life cycle of that piece of software with Yeti’s, you know.

252 00:29:31.150 00:29:33.429 Casey Isaac: These workflow automations, stuff like that, so….

253 00:29:33.430 00:29:34.050 Robert Tseng: Yeah.

254 00:29:34.580 00:29:36.000 Robert Tseng: Yeah. So…

255 00:29:36.450 00:29:47.040 Robert Tseng: Cool, I mean, if you have… I mean, if you have an eye for it and things that you want to improve, like, I… we’ve definitely kind of talked about how to… how to make it work. And yeah, I mean, it helps that, like.

256 00:29:47.320 00:29:56.729 Robert Tseng: you’re a younger owner compared to some of the other business owners we’ve talked to that are very, like, AI adverse, and that makes it a difficult conversation.

257 00:29:56.750 00:30:10.929 Robert Tseng: And, like, they really need to see the value before they, like, really try anything, and so we’re just stuck building demos all the time, but honestly, if you just have a workflow you want to, like, run it on, I would just recommend just doing it and just seeing, like.

258 00:30:10.970 00:30:23.809 Robert Tseng: Yeah, just… just… just let it… just let it… let it run in production and see… see how it helps, and I think you’ll… you’ll be… I’m sure you’ve already know what it can do, so, yeah, like, it’s…

259 00:30:24.030 00:30:25.520 Robert Tseng: I think it’s pretty cool.

260 00:30:26.270 00:30:27.560 Casey Isaac: Awesome, man. Yeah.

261 00:30:27.800 00:30:35.219 Casey Isaac: Well, dude, it’s great to see your face, and hopefully, I would love to grab coffee, too, sometime when I’m back in New York.

262 00:30:35.220 00:30:35.850 Robert Tseng: Yeah.

263 00:30:36.210 00:30:41.170 Casey Isaac: be there in September, actually, a couple of times, so….

264 00:30:41.380 00:30:41.840 Robert Tseng: Okay.

265 00:30:41.840 00:30:49.350 Casey Isaac: I… yeah, I’ll be there from the… let’s see, this 5th to this… the 14th, and then also…

266 00:30:49.980 00:30:58.069 Casey Isaac: sometime around the 23rd through, I don’t know, early October. So, I’m kind of back and forth over the next month, month and a half.

267 00:30:58.190 00:31:00.799 Casey Isaac: But yeah, we can, we can also just…

268 00:31:01.370 00:31:05.799 Casey Isaac: you know, do a Zoom to get this stuff figured out. I, again…

269 00:31:06.270 00:31:08.990 Casey Isaac: a lot of pain right now. This is definitely…

270 00:31:09.220 00:31:10.949 Casey Isaac: I gotta get it figured out, so…

271 00:31:11.210 00:31:13.309 Casey Isaac: Okay. It’s good to know.

272 00:31:13.540 00:31:18.319 Casey Isaac: Somebody who… it sounds like you… like you’re a perfect fit for what we’re looking for, so…

273 00:31:18.650 00:31:21.320 Casey Isaac: So anyways, yeah, let’s, let’s, …

274 00:31:21.890 00:31:24.870 Casey Isaac: Let’s get moving on it, just let me know what else you need.

275 00:31:25.190 00:31:32.270 Casey Isaac: And, … and then, yeah, but, you know, would love to meet you in person at some point back in the city, so….

276 00:31:32.710 00:31:47.279 Robert Tseng: Okay, yeah, that sounds good. Yeah, I’ll follow up with an email shortly after this call. I’ll send you, kind of, some… some follow-ups, and yeah, I might loop my business partner into the call, just so he can say hi as well, and yeah, we’ll just… we’ll just keep… keep talking.

277 00:31:47.850 00:31:50.219 Casey Isaac: That was great. Thanks so much, Robert. Good to see you, man.

278 00:31:50.490 00:31:50.970 Robert Tseng: Thanks, Casey.

279 00:31:51.700 00:31:52.530 Robert Tseng: You see. Bye.