2025-11-14_ai-default-abc_standup

Meeting Title: AI-Default-ABC Standup Date: 2025-11-14 Meeting participants: Awaish Kumar, Amber Lin, Uttam Kumaran, Zoran Selinger, Samuel Roberts, Rico Rejoso, Demilade Agboola, Henry Zhao, Robert Tseng, Casie Aviles, Gabriel Lam, Mustafa Raja

WEBVTT

1 00:01:32.530 ⇒ 00:01:33.550 Uttam Kumaran: Hello.

2 00:01:35.730 ⇒ 00:01:36.290 Amber Lin: Hi there.

3 00:01:36.290 ⇒ 00:01:36.890 Zoran Selinger: Hello.

4 00:01:37.280 ⇒ 00:01:38.020 Awaish Kumar: Hello.

5 00:01:39.080 ⇒ 00:01:39.920 Uttam Kumaran: A…

6 00:02:05.270 ⇒ 00:02:06.120 Uttam Kumaran: Blue.

7 00:02:45.290 ⇒ 00:02:55.220 Uttam Kumaran: Okay, maybe we’ll get started. So, Amber, I spoke with Robert to start with Insomnia. I think we can probably just, like, pause for a bit. I think we have.

8 00:02:56.390 ⇒ 00:02:59.819 Uttam Kumaran: Easily 2 weeks’ worth of insights for him to just…

9 00:03:00.250 ⇒ 00:03:03.720 Amber Lin: Oh, that would be… okay, that’s awesome. Then I…

10 00:03:03.720 ⇒ 00:03:04.140 Uttam Kumaran: I may…

11 00:03:04.140 ⇒ 00:03:05.140 Amber Lin: time.

12 00:03:05.350 ⇒ 00:03:06.960 Uttam Kumaran: Yes, I would really love.

13 00:03:06.960 ⇒ 00:03:07.330 Amber Lin: No way.

14 00:03:07.330 ⇒ 00:03:09.230 Uttam Kumaran: Yeah.

15 00:03:09.360 ⇒ 00:03:11.749 Uttam Kumaran: So we have, I think, a…

16 00:03:12.390 ⇒ 00:03:20.499 Uttam Kumaran: Well, I think, yeah, we can decide on where that goes, but I think we’re sort of at a good place there. I think my only… my only question

17 00:03:20.830 ⇒ 00:03:25.470 Uttam Kumaran: for today on insomnias, if… do we want to try to send any of this?

18 00:03:25.760 ⇒ 00:03:28.110 Uttam Kumaran: Over today, or do you think it’s.

19 00:03:28.110 ⇒ 00:03:29.429 Amber Lin: Work for Optics, Robert.

20 00:03:29.430 ⇒ 00:03:29.850 Uttam Kumaran: or…

21 00:03:30.250 ⇒ 00:03:30.870 Uttam Kumaran: Yeah.

22 00:03:31.600 ⇒ 00:03:38.880 Amber Lin: Well, when I talked to Robert, Robert said he’ll want to send something by end of day today, so that was my plan to try to polish things up.

23 00:03:40.800 ⇒ 00:03:48.599 Robert Tseng: Yeah, I mean, I know you sent a couple more things, I haven’t reviewed it yet, but, you know, I gave you some direction off of our chat yesterday, so…

24 00:03:48.700 ⇒ 00:04:02.720 Robert Tseng: I think we’re close to being able to send something. We don’t need to send all of it, I think just, like, one or two things to get in front of Birdie, I think, would be important, and then you, you know, kind of, like, how you’re gonna prepare for the call with Matt next week, right?

25 00:04:05.930 ⇒ 00:04:06.450 Amber Lin: Yep.

26 00:04:06.690 ⇒ 00:04:07.320 Uttam Kumaran: Cool.

27 00:04:08.750 ⇒ 00:04:21.799 Uttam Kumaran: Okay, great, so I feel good about insomnia. I mean, I’ll just knock out a couple of the easy ones. So, for Hydra, Demolati, I know it’s late, the meeting later. For some reason, she’s like.

28 00:04:22.010 ⇒ 00:04:30.299 Uttam Kumaran: always wants to meet, like, Friday, end of day. So, I just want to confirm if, like, you’re gonna be there, or,

29 00:04:31.020 ⇒ 00:04:32.240 Uttam Kumaran: You’re gonna be off.

30 00:04:36.300 ⇒ 00:04:38.289 Demilade Agboola: I might probably be off, to be honest.

31 00:04:38.630 ⇒ 00:04:39.250 Uttam Kumaran: Okay.

32 00:04:39.980 ⇒ 00:04:44.820 Uttam Kumaran: So I will, I’m gonna present… I’ll present your…

33 00:04:45.660 ⇒ 00:05:00.480 Uttam Kumaran: your PR. Maybe if you… if there’s anything, any nuances, if you can shoot me a loom or just Slack me. I looked at it a little bit yesterday, but, like, nothing in depth, so… but I can present on it, or if there’s any questions you want me to ask her.

34 00:05:00.590 ⇒ 00:05:03.579 Uttam Kumaran: You could also ask her directly in the channel,

35 00:05:03.960 ⇒ 00:05:05.650 Uttam Kumaran: to take a look at it, I think…

36 00:05:05.920 ⇒ 00:05:08.300 Uttam Kumaran: Now that you’re starting to see if you can…

37 00:05:09.250 ⇒ 00:05:11.420 Uttam Kumaran: Start to ping them, that’ll… that’ll be helpful.

38 00:05:12.280 ⇒ 00:05:16.799 Demilade Agboola: Okay, sure. Well, can you also add me as optional to the meeting? Because I can actually sit on my calendar.

39 00:05:17.880 ⇒ 00:05:18.750 Uttam Kumaran: Yeah, you are.

40 00:05:20.400 ⇒ 00:05:20.959 Demilade Agboola: I, I…

41 00:05:20.960 ⇒ 00:05:23.920 Uttam Kumaran: Oh, you, you… Oh.

42 00:05:24.650 ⇒ 00:05:25.420 Demilade Agboola: Yeah.

43 00:05:26.080 ⇒ 00:05:26.760 Uttam Kumaran: Really?

44 00:05:27.840 ⇒ 00:05:28.990 Demilade Agboola: No.

45 00:05:30.060 ⇒ 00:05:30.870 Uttam Kumaran: But…

46 00:05:35.020 ⇒ 00:05:38.110 Uttam Kumaran: Wait, dude, yeah, I see it on your calendar.

47 00:05:39.490 ⇒ 00:05:40.730 Demilade Agboola: Am I tripping?

48 00:05:41.210 ⇒ 00:05:48.799 Demilade Agboola: Oh, yeah, I can’t… oh, my bad. So it’s… it’s so late, it’s the next day for me, and so I’m… when I’m… I can’t see…

49 00:05:48.800 ⇒ 00:05:50.260 Uttam Kumaran: Sorry, sorry.

50 00:05:51.560 ⇒ 00:05:54.460 Uttam Kumaran: Yeah, dude, it’s 5PM my time, bro, like…

51 00:05:54.700 ⇒ 00:05:59.009 Demilade Agboola: Yeah, I can’t… it’s the next day. Long day. Yeah.

52 00:06:00.890 ⇒ 00:06:14.999 Uttam Kumaran: It’s gonna be a long day. Okay. So, alright, cool. Alright, so yeah, if you… whatever, if you just wanna keep pushing or on… or send a… send a note on questions, you can even send it directly in the channel. They’re… they’re, like, these guys are always online, so…

53 00:06:15.310 ⇒ 00:06:17.630 Uttam Kumaran: Cool.

54 00:06:17.890 ⇒ 00:06:26.709 Uttam Kumaran: For, README, yeah, maybe I’ll just, like, ping…

55 00:06:27.750 ⇒ 00:06:40.410 Uttam Kumaran: a way to join, but I basically sent a message, Robert, that, like, we can’t do shit in there, so, like, I can’t combine the amplitude data in Mongo, like, doesn’t work. I… I didn’t, like, have it… I didn’t… I didn’t have enough…

56 00:06:40.700 ⇒ 00:06:52.009 Uttam Kumaran: Oh yeah, wait, sorry, you’re here. I didn’t see you. I didn’t have enough, like, context to go defend that against two other people that have no idea what they’re talking about in yesterday’s meeting.

57 00:06:52.170 ⇒ 00:06:56.949 Uttam Kumaran: But, like… I do now, so… it’s…

58 00:06:57.160 ⇒ 00:06:59.949 Uttam Kumaran: I clearly outlined to him, like, what…

59 00:07:00.150 ⇒ 00:07:16.020 Uttam Kumaran: like, we were trying to accomplish, and I will let him figure it out, or we can escalate. But I can’t pull any data out of Mongo, and he was like, you can run aggregates there, but we’re not trying to run aggregates. We’re trying to join the data.

60 00:07:16.150 ⇒ 00:07:21.689 Uttam Kumaran: Right? So, like, I don’t know how to, like, proceed at this point.

61 00:07:21.690 ⇒ 00:07:30.140 Robert Tseng: Well, what if we just run the aggregates, and, like, I could just… we could do an aggregation and amp… I just need to… I think they just wanna…

62 00:07:30.140 ⇒ 00:07:33.610 Uttam Kumaran: I think what you’re saying is right. You just won the total counts?

63 00:07:33.880 ⇒ 00:07:40.569 Robert Tseng: Yeah, like, whatever, like, bear… if… yeah, like, I think that’s better than nothing, I guess.

64 00:07:40.570 ⇒ 00:07:43.740 Uttam Kumaran: So, so what, so what counts do you want? You want cl… you want…

65 00:07:44.060 ⇒ 00:07:47.330 Uttam Kumaran: all plans, like, number, like, yeah. And then…

66 00:07:47.330 ⇒ 00:07:55.289 Robert Tseng: Yeah, I mean, I’m assuming you can’t do any time-based aggregations. If you can, that’d be great. We could just look at the past month, and then I could just compare that to what’s in amplitude.

67 00:07:55.780 ⇒ 00:07:56.280 Awaish Kumar: Who can…

68 00:07:56.280 ⇒ 00:08:03.610 Robert Tseng: past month, all projects created, past month, all… I mean, the four steps in the funnel, I can send it to you again, but… yeah.

69 00:08:03.910 ⇒ 00:08:10.620 Awaish Kumar: You can… can… we can run aggregations, but the thing is, we can’t see all the data… full data set there. It only shows.

70 00:08:10.620 ⇒ 00:08:12.739 Robert Tseng: You don’t… I just need the numbers, it’s fine.

71 00:08:13.560 ⇒ 00:08:15.550 Awaish Kumar: Okay, yeah, it leads… Yeah.

72 00:08:15.550 ⇒ 00:08:26.159 Robert Tseng: Like, I just need to know, are the number of… if they have, like, 500 Stripe payments in there in the past month, and Amplitude is showing 50, then, like.

73 00:08:26.340 ⇒ 00:08:28.720 Robert Tseng: Yeah, like, that would… that would be…

74 00:08:29.120 ⇒ 00:08:32.109 Robert Tseng: you know, that’s something that I could share with them.

75 00:08:33.440 ⇒ 00:08:33.990 Awaish Kumar: Okay.

76 00:08:33.990 ⇒ 00:08:51.409 Robert Tseng: So, okay, let me… let me just go pull up the amplitude report. If you could do time-based aggregations, just pull up for the past month, or, like, since… since start of… ideal… if you could do monthly since September, like, that would be great. But, if not, then I would just say since…

77 00:08:51.720 ⇒ 00:08:53.669 Robert Tseng: Like, yeah, since September.

78 00:08:53.950 ⇒ 00:09:00.060 Robert Tseng: We’re not… sorry, sorry, monthly since September, or just, like, the past month?

79 00:09:01.310 ⇒ 00:09:04.960 Uttam Kumaran: And we can’t export anything, so this will literally be, like, we have to… can we… we just put it.

80 00:09:04.960 ⇒ 00:09:11.570 Robert Tseng: Yeah, literally just give me the screenshot or whatever of the number, and I will just match it to… I’ll just compare it to Amplitude.

81 00:09:11.570 ⇒ 00:09:16.899 Uttam Kumaran: And so you want, like, new users, new paid, like, new paid users, new subscriptions…

82 00:09:17.580 ⇒ 00:09:18.250 Robert Tseng: Yeah.

83 00:09:21.370 ⇒ 00:09:23.880 Uttam Kumaran: I wish you feel comfortable with that? Grabbing that?

84 00:09:24.810 ⇒ 00:09:31.250 Awaish Kumar: Yeah, like, I can do some of it, like, use, like, the subscriptions created, or…

85 00:09:31.710 ⇒ 00:09:32.060 Uttam Kumaran: Yeah.

86 00:09:32.710 ⇒ 00:09:39.019 Uttam Kumaran: Some of it, if you, if you, if you can do month, plan type, Create a date.

87 00:09:40.010 ⇒ 00:09:41.450 Uttam Kumaran: That’s one thing.

88 00:09:42.650 ⇒ 00:09:43.450 Uttam Kumaran: Yeah.

89 00:09:44.680 ⇒ 00:09:49.149 Awaish Kumar: See what you can get, and then, I mean, I’m… we’re just gonna do… yeah, it’s just…

90 00:09:49.150 ⇒ 00:09:50.350 Uttam Kumaran: Fucking stupid.

91 00:09:52.480 ⇒ 00:09:57.469 Awaish Kumar: Yeah, we just need, like, they have quite a lot of database users and passwords there, like.

92 00:09:57.620 ⇒ 00:10:00.699 Awaish Kumar: The users, like, if they can create one for us.

93 00:10:01.460 ⇒ 00:10:04.110 Uttam Kumaran: Oh, I know. Believe me. I know. I know.

94 00:10:04.480 ⇒ 00:10:08.420 Uttam Kumaran: I’m trying… I… yeah, that’s the only reason we’re here. It’s not because I didn’t ask.

95 00:10:08.620 ⇒ 00:10:13.000 Uttam Kumaran: The guy was like, we don’t want to give access to data. I was like.

96 00:10:13.440 ⇒ 00:10:17.199 Uttam Kumaran: what kind of analytics you want us to do, then? But, okay.

97 00:10:17.950 ⇒ 00:10:24.529 Uttam Kumaran: Alright, so… if you can look at new users, subscriptions, buy some dates, pull aggregates, that’s fine.

98 00:10:27.150 ⇒ 00:10:27.740 Awaish Kumar: Okay.

99 00:10:37.240 ⇒ 00:10:41.030 Uttam Kumaran: Okay, great. So that’s README…

100 00:10:42.360 ⇒ 00:10:46.449 Uttam Kumaran: For Ellie, yeah, Zora, I think just, just, just email them.

101 00:10:47.270 ⇒ 00:10:48.800 Zoran Selinger: I just did, I just did.

102 00:10:48.800 ⇒ 00:10:49.960 Uttam Kumaran: Cool. Perfect.

103 00:10:50.310 ⇒ 00:10:51.290 Zoran Selinger: Thank you.

104 00:10:54.140 ⇒ 00:10:56.740 Uttam Kumaran: Anything else on Ellie, Robert?

105 00:10:57.440 ⇒ 00:11:06.879 Robert Tseng: No, I mean, she’s still trying to get approval of it. I mean, yeah, I think things are just gonna stall. So, I have been slacking her every day, so she’s still responsible.

106 00:11:06.880 ⇒ 00:11:07.310 Uttam Kumaran: Oh, I thought.

107 00:11:07.310 ⇒ 00:11:12.200 Robert Tseng: Whatever. Yeah. She just can’t be in our channel anymore, because, like, their free plan ran out, so…

108 00:11:12.380 ⇒ 00:11:15.540 Robert Tseng: I think… Yeah, email’s fine.

109 00:11:15.540 ⇒ 00:11:15.940 Uttam Kumaran: Bye, Harry.

110 00:11:15.940 ⇒ 00:11:16.670 Robert Tseng: Medium, t-shirt.

111 00:11:16.670 ⇒ 00:11:17.410 Uttam Kumaran: Okay.

112 00:11:17.900 ⇒ 00:11:24.309 Robert Tseng: I can either invite her, or I was saying, like, if you need us to be on Teams, because that’s where they are, like, if we renew, then we’ll move to Teams, that’s fine.

113 00:11:26.150 ⇒ 00:11:26.880 Robert Tseng: Yeah.

114 00:11:26.880 ⇒ 00:11:39.910 Uttam Kumaran: Great. Next question I have is on ABC. So, I presented this yesterday, we kind of got approval. I don’t… kind of now thinking about our…

115 00:11:40.780 ⇒ 00:11:42.389 Uttam Kumaran: Two calls today.

116 00:11:42.530 ⇒ 00:11:47.890 Uttam Kumaran: like, I’m wondering how we should, like, if we should not just, like, send this like this, like…

117 00:11:48.280 ⇒ 00:11:49.390 Uttam Kumaran: How do we…

118 00:11:50.570 ⇒ 00:11:57.400 Uttam Kumaran: like, what should we do? Like, I was just gonna send this to Matt, CFO, and, like, one of the GMs, and then… and then was gonna…

119 00:11:58.640 ⇒ 00:12:05.790 Uttam Kumaran: circulate it wider, but should we put together a deck, or, like, should we just add on more stuff about our process here? Like, what do you think?

120 00:12:06.670 ⇒ 00:12:11.850 Robert Tseng: Yeah, I mean, at this point, I feel like… I mean, have they been reading our stuff? If not, then, like, maybe.

121 00:12:11.850 ⇒ 00:12:12.200 Uttam Kumaran: No.

122 00:12:12.200 ⇒ 00:12:14.430 Robert Tseng: put together a Gantt chart for them.

123 00:12:14.660 ⇒ 00:12:22.470 Robert Tseng: Maybe they don’t even need that, because they’re not really thinking 2 or 3 steps ahead, and the deck will be fine. You know, I would say the deck is probably lower lift for us, so we’ll just do the deck.

124 00:12:22.800 ⇒ 00:12:28.910 Robert Tseng: send it to them. Yeah, I can… I can help review that deck. It’ll be something similar to what we sent at Element.

125 00:12:30.350 ⇒ 00:12:38.320 Robert Tseng: And then, yeah, if they… if they have more questions about, kind of, like, timeline or whatever, then we can… we can build out the Gantt chart.

126 00:12:40.000 ⇒ 00:12:40.610 Uttam Kumaran: Okay.

127 00:12:40.710 ⇒ 00:12:43.409 Uttam Kumaran: Amber, do you wanna… do you wanna take that?

128 00:12:43.540 ⇒ 00:12:51.670 Uttam Kumaran: The content here is… is good, I just… Yeah, we’re learning about, like, Slowing down to speed up.

129 00:12:51.870 ⇒ 00:12:53.549 Uttam Kumaran: So, I just wanted to,

130 00:12:53.900 ⇒ 00:12:59.019 Uttam Kumaran: I just want to make sure that that’s captured here. Because also, these guys are very…

131 00:12:59.280 ⇒ 00:13:05.720 Uttam Kumaran: legacy. So, we are… we will inundate them with big words pretty fast.

132 00:13:05.820 ⇒ 00:13:10.920 Uttam Kumaran: So I really want to highlight this…

133 00:13:12.630 ⇒ 00:13:17.220 Uttam Kumaran: I think they’re… we should have this, but they’re gonna breeze past this.

134 00:13:17.570 ⇒ 00:13:30.049 Uttam Kumaran: they’re like, this and, like, this is, like, what I want to highlight, is, like, how we’re gonna go through systematically, and what they should expect, and on what timeline. They’re gonna sign, they’ll sign it, so…

135 00:13:32.360 ⇒ 00:13:37.279 Uttam Kumaran: Okay. So, Amber, if you can take that, and I would like to get that to Matt.

136 00:13:38.450 ⇒ 00:13:41.780 Uttam Kumaran: Literally as soon as we can. Cool, okay.

137 00:13:43.710 ⇒ 00:13:56.860 Uttam Kumaran: Cool. For default, we’re kind of continuing course. I would say Amber’s been doing a great job on really guiding their decision on pricing. I think…

138 00:13:57.180 ⇒ 00:13:59.039 Uttam Kumaran: We may be at the point

139 00:14:00.060 ⇒ 00:14:07.299 Uttam Kumaran: where they’re at… they’re making some decisions. I don’t know, Robert, have you done a lot in, like, B2B and, like, SaaS pricing stuff?

140 00:14:09.010 ⇒ 00:14:16.720 Robert Tseng: I mean, I did a little bit of that. I mean, this was, like, early… I mean, pricing is just, like…

141 00:14:17.420 ⇒ 00:14:22.210 Robert Tseng: You just… you just experiment a bunch, yeah. I think I set, like, a cool…

142 00:14:22.410 ⇒ 00:14:29.739 Robert Tseng: a recording from, like, a friend of mine who’s, like, doing pricing stuff, but yeah, I mean, I don’t know how early they are, but it’s totally not very, like.

143 00:14:30.390 ⇒ 00:14:34.350 Robert Tseng: Scientific, you just… people just price whatever the heck they want when they’re early on.

144 00:14:34.350 ⇒ 00:14:34.750 Uttam Kumaran: I think we’re.

145 00:14:34.750 ⇒ 00:14:35.720 Robert Tseng: You just need to run enough time.

146 00:14:35.720 ⇒ 00:14:36.619 Uttam Kumaran: Like, a lot

147 00:14:36.950 ⇒ 00:14:44.149 Uttam Kumaran: Yeah, we’ve had a lot of conversations, and I think we have enough information. I kind of want to go one step further to help them

148 00:14:44.340 ⇒ 00:14:46.659 Uttam Kumaran: Build, like, a pricing dock.

149 00:14:46.940 ⇒ 00:14:48.699 Robert Tseng: Wait, sorry, who’s this for?

150 00:14:48.700 ⇒ 00:14:49.740 Uttam Kumaran: It’s her default.

151 00:14:50.020 ⇒ 00:14:58.310 Robert Tseng: Oh, for default. Oh, okay, sure. Yeah, I mean, yeah, okay, I mean, there’s Series A pricing doc, sure. Yeah, I mean, let’s, let’s, let’s do one. Let’s do, let’s do a pricing thing for them.

152 00:14:59.080 ⇒ 00:15:04.309 Uttam Kumaran: Yeah, so we’ve had a bunch of conversations with them. I think we’ve arrived at some conclusions.

153 00:15:04.560 ⇒ 00:15:10.880 Uttam Kumaran: I think Caitlin can go circulate it and, like, build advocacy for it. She’s given us all the details. I kind of just want to, like.

154 00:15:11.080 ⇒ 00:15:11.870 Uttam Kumaran: Give her some food.

155 00:15:11.870 ⇒ 00:15:19.580 Robert Tseng: Who on our team knows enough about this that we can just catch me up on where we’re at? Like, I think I can… I know staffing models, like I… Amber.

156 00:15:19.580 ⇒ 00:15:20.380 Uttam Kumaran: Amber, yeah.

157 00:15:20.380 ⇒ 00:15:20.730 Amber Lin: Okay.

158 00:15:20.730 ⇒ 00:15:25.580 Robert Tseng: Okay, then Amber, we’ll chat. Let’s just kind of see what you’ve got, and I can give you some…

159 00:15:25.910 ⇒ 00:15:29.049 Robert Tseng: Or I might take it over, or I’ll let you know. Yeah.

160 00:15:29.050 ⇒ 00:15:31.140 Amber Lin: Yeah, sounds good. Grab anytime, you’re free.

161 00:15:31.560 ⇒ 00:15:32.190 Robert Tseng: Okay.

162 00:15:32.970 ⇒ 00:15:39.399 Uttam Kumaran: For urban stems, yeah, I think, Oasis, I’m just gonna continue to…

163 00:15:40.010 ⇒ 00:15:50.040 Uttam Kumaran: press on you to make sure things are stable, but I feel like things rained last night, and things are stable. I think next week we can talk about more of, like, on-call and stuff like that, but overall, I feel…

164 00:15:50.710 ⇒ 00:15:51.750 Uttam Kumaran: Pretty good.

165 00:15:54.020 ⇒ 00:16:02.859 Uttam Kumaran: I… need to read your sizing doc, Robert, on forecasting, but I think one piece there is, Oisha, I might get your help.

166 00:16:02.990 ⇒ 00:16:07.110 Uttam Kumaran: Today, if you can take a look at,

167 00:16:09.090 ⇒ 00:16:14.910 Uttam Kumaran: If you can take a look at this document,

168 00:16:21.650 ⇒ 00:16:34.840 Uttam Kumaran: I will tag you here… And I’m gonna just share… Basically, we’re trying… oops.

169 00:16:35.880 ⇒ 00:16:48.210 Uttam Kumaran: Basically, Robert put together this stock on… like, a project around… driver-based forecast.

170 00:16:48.500 ⇒ 00:16:55.029 Uttam Kumaran: I think there are some further questions we have here.

171 00:16:55.130 ⇒ 00:17:03.350 Uttam Kumaran: I guess, Robert, which parts… is it really these, these, like, is it like this, mainly?

172 00:17:04.660 ⇒ 00:17:06.849 Robert Tseng: Like, next steps, or what?

173 00:17:06.859 ⇒ 00:17:07.199 Uttam Kumaran: Yeah.

174 00:17:07.200 ⇒ 00:17:10.950 Robert Tseng: I mean, maybe if I could just walk through this doc, like, if you scroll up a little bit…

175 00:17:11.380 ⇒ 00:17:17.720 Robert Tseng: I should actually call it 90% accuracy, never said it’s actually the wrong… Brooke this.

176 00:17:18.560 ⇒ 00:17:19.760 Uttam Kumaran: Okay, hold on.

177 00:17:19.760 ⇒ 00:17:21.899 Robert Tseng: That’s fine, I’ll edit it, that’s fine.

178 00:17:22.200 ⇒ 00:17:22.579 Uttam Kumaran: Okay.

179 00:17:22.589 ⇒ 00:17:29.089 Robert Tseng: I just thought that was funny, I didn’t catch that yesterday.

180 00:17:29.289 ⇒ 00:17:33.989 Uttam Kumaran: In some industries, dude, 10% accuracy could be good. I don’t know, it’s all relative.

181 00:17:34.450 ⇒ 00:17:37.430 Robert Tseng: I meant, like, within 10%, so…

182 00:17:37.840 ⇒ 00:17:39.960 Robert Tseng: The way it reads, this was just wrong.

183 00:17:40.200 ⇒ 00:17:51.669 Robert Tseng: Okay, well, I mean, right now, they’re off by, like, 25 to 30%, right? So they’re overestimating by 25-30%. And I think… so, you know, the summary is what it is, like, I just kind of described what the state of affairs is.

184 00:17:51.970 ⇒ 00:18:10.190 Robert Tseng: I think, you know, I backed into the, like, $3 t o$ 5 million impact based on the current side of their business. You know, if you make some assumptions about the waste that they would recover, if you forecast more accurately, it has a direct margin impact. So that’s kind of, like, basically what I was saying, making about that.

185 00:18:10.270 ⇒ 00:18:24.540 Robert Tseng: And then, if you scroll down, yeah, like, this, that, this mapping, like, I didn’t… I wasn’t able to look at the SQL itself, but, like, I did look at every budget forecast-related model in the gables, that, after Dave and Lottie got me set up.

186 00:18:24.550 ⇒ 00:18:30.979 Robert Tseng: And I kind of, like… I mean, I could have done a better job in Figma, but I just let GPT build this flow for me.

187 00:18:31.470 ⇒ 00:18:42.829 Robert Tseng: So basically, like, these are… it’s… it’s just like a budget… it’s just like a budget allocation process. It’s not a true forecast. You know, it’s… it kind of operates similar to…

188 00:18:42.850 ⇒ 00:19:00.500 Robert Tseng: whatever. It’s just like an overrun budget that kind of gets broken up into different things. So, budget, and then they budget by component, they budget, and then they… and then they do some… some aggregations. So, I kind of… I laid out my findings there on, like, what’s actually in these models, and, like, why this is not a true forecast.

189 00:19:01.180 ⇒ 00:19:18.009 Robert Tseng: is completely static. If any assumptions change, like, this doesn’t get updated. They’re not tracking any sort of, like, variance from, like, a target over time, so it’s… it’s like someone just built it, and then set it and forget it, or whatever. So, it’s just, like, not… yeah, there’s no…

190 00:19:18.080 ⇒ 00:19:27.030 Robert Tseng: Yeah, anyway, so it’s purely off of, like, whatever budget expectations that they set up front. It looks like they probably do this, like, once or twice a year.

191 00:19:27.480 ⇒ 00:19:31.420 Robert Tseng: And to me, it’s more of a budgeting exercise than a forecast.

192 00:19:31.700 ⇒ 00:19:41.239 Robert Tseng: So then I kind of, like, laid out, like, the next phases of, like, what we need to do. What does driver-based, like, forecasting even mean? Like, what are the things that I would look at?

193 00:19:41.270 ⇒ 00:19:59.710 Robert Tseng: On both… really on the demand side, looking at their sales history, looking at other external signals, things as much as, like… and, yeah, being able to, to, to fold that into… into, like, a few different models for, for a forecast. So,

194 00:19:59.750 ⇒ 00:20:02.329 Robert Tseng: Yeah, that’s basically what I’m describing here.

195 00:20:02.380 ⇒ 00:20:08.679 Robert Tseng: I mean, I can walk through it, and I can do a voiceover on a loom to make it more clear, but these are all kind of the different…

196 00:20:08.980 ⇒ 00:20:16.069 Robert Tseng: parts to… to, like, how… how you would approach this. And then I wrote, like, a business case summary at the end, or, like.

197 00:20:16.370 ⇒ 00:20:23.290 Robert Tseng: That kind of is if I needed to send, like, a long-form doc to someone. Like, this explains, like, why,

198 00:20:23.500 ⇒ 00:20:25.459 Robert Tseng: We need to change this approach.

199 00:20:25.460 ⇒ 00:20:27.879 Uttam Kumaran: I would like to send this to Zach today.

200 00:20:28.340 ⇒ 00:20:30.290 Uttam Kumaran: Yeah. It’s like this. But, like…

201 00:20:30.290 ⇒ 00:20:30.640 Robert Tseng: Okay.

202 00:20:30.640 ⇒ 00:20:32.810 Uttam Kumaran: Just be like, hey, we’re gonna… we… this is, like.

203 00:20:33.260 ⇒ 00:20:36.760 Uttam Kumaran: what we were… this is, like, what we’re gonna present on Tuesday. I think a way.

204 00:20:36.760 ⇒ 00:20:37.100 Robert Tseng: Okay.

205 00:20:37.100 ⇒ 00:20:40.829 Uttam Kumaran: If you can just get Like, a fig jam in here?

206 00:20:41.020 ⇒ 00:20:44.129 Uttam Kumaran: Because I don’t mind, we can make edits, he’s not gonna, like… but…

207 00:20:44.940 ⇒ 00:20:48.060 Uttam Kumaran: I wanna just… this is great, I just wanna get it out.

208 00:20:49.200 ⇒ 00:20:58.509 Uttam Kumaran: Because we’re not, like… yeah, so… is there anything else we want to add here? I just would like… we can just make a little bit of a better diagram, but is there anything else we want to add here?

209 00:20:59.380 ⇒ 00:21:14.460 Robert Tseng: I mean, like, all the content is there in terms of visuals, like, I mean, I don’t know, everything would just be hypothetical, but I’m sure they can’t do any sort of, like, pacing right now. Like, Eden asked for this as well, which… which is why I think this is…

210 00:21:14.650 ⇒ 00:21:18.600 Robert Tseng: Like, something… we need this capability, it’s like,

211 00:21:19.220 ⇒ 00:21:28.529 Robert Tseng: Yeah, like, they… they don’t know… in their forecast, every… at… at the most granular level, they have, like, a weekly… they have a weekly number.

212 00:21:31.330 ⇒ 00:21:45.549 Robert Tseng: Well, I mean, I don’t know, like, I guess I could screenshot what, like, some of… I could reshare some of the screenshots of their existing reports, and then, like, try to mock up, like, what the next one should look like, but, like, I don’t know, like, that might be too much for what it is right now.

213 00:21:49.150 ⇒ 00:21:53.380 Uttam Kumaran: Yeah, maybe if you have the existing report screenshots.

214 00:21:53.730 ⇒ 00:21:54.469 Robert Tseng: Yeah, I have that.

215 00:21:54.470 ⇒ 00:21:54.890 Uttam Kumaran: Let’s put it.

216 00:21:54.890 ⇒ 00:21:57.870 Robert Tseng: I have a screenshotted already, yeah. Okay.

217 00:21:57.870 ⇒ 00:22:01.400 Uttam Kumaran: So let’s, let’s put… let’s… let’s put that in here, and then that’s fine.

218 00:22:02.750 ⇒ 00:22:09.099 Robert Tseng: Okay. Yeah, I’ll go through and just kind of clean up the structure a bit. I just kind of put it together late last night, so… No, I think it’s…

219 00:22:09.100 ⇒ 00:22:12.620 Uttam Kumaran: great, like, I think we should just get this out, like, it’s,

220 00:22:13.080 ⇒ 00:22:16.610 Uttam Kumaran: I want to tee up, like, I’m teeing up a couple things to him.

221 00:22:16.760 ⇒ 00:22:21.630 Uttam Kumaran: Which is the fact that, one, he’s like, hey, can we move you guys to, like, support? I’m like, dude, there’s no way…

222 00:22:21.810 ⇒ 00:22:25.129 Uttam Kumaran: there’s no way. I was like, look, in my honest opinion.

223 00:22:25.280 ⇒ 00:22:44.539 Uttam Kumaran: there’s no way you could continue to do this without, like, pretty significant support. Like, there’s no way… Emily is not senior enough to manage. Second, I think I want… I want to tailor this towards, like, what is the opportunity, and their lead forecasting person, Perry, is on the outs, so they’re booked. Like, I don’t know what they’re gonna do.

224 00:22:45.060 ⇒ 00:22:45.560 Robert Tseng: Okay.

225 00:22:45.560 ⇒ 00:22:48.040 Uttam Kumaran: So, I think he’s just playing hard to get, like, yeah.

226 00:22:48.040 ⇒ 00:22:48.650 Robert Tseng: Yeah.

227 00:22:49.800 ⇒ 00:23:01.160 Demilade Agboola: Also, like, I was wondering if we could put examples of where, like, 25-30% disparity comes in? Like, it was… your forecast was this amount, it ended up being this amount, and…

228 00:23:03.050 ⇒ 00:23:03.740 Robert Tseng: Yeah.

229 00:23:04.350 ⇒ 00:23:05.799 Robert Tseng: Okay, I could do that.

230 00:23:20.070 ⇒ 00:23:22.330 Uttam Kumaran: Okay, great, solid. Cool.

231 00:23:23.450 ⇒ 00:23:32.959 Robert Tseng: My only worry is that Zach’s not the right person for that, because it’s more of, like, a COO or, like, someone like that who would sign off on that project, you know?

232 00:23:33.080 ⇒ 00:23:33.860 Robert Tseng: So…

233 00:23:34.830 ⇒ 00:23:35.590 Uttam Kumaran: Yeah.

234 00:23:36.890 ⇒ 00:23:40.940 Uttam Kumaran: Yeah, I mean, correct. I mean, I can send it to him and…

235 00:23:41.220 ⇒ 00:23:45.669 Uttam Kumaran: be like, this is something we’d like to discuss, is there anyone else that’s, like.

236 00:23:45.810 ⇒ 00:23:50.089 Uttam Kumaran: I know Perry leaving, who else is involved in this? Can you bring them? Yeah.

237 00:23:50.090 ⇒ 00:23:52.819 Robert Tseng: Maybe that’s how you should frame the message, yeah.

238 00:23:53.760 ⇒ 00:23:54.330 Uttam Kumaran: Okay.

239 00:23:54.430 ⇒ 00:23:55.130 Uttam Kumaran: Alright.

240 00:23:57.520 ⇒ 00:23:58.489 Uttam Kumaran: So let me…

241 00:24:07.380 ⇒ 00:24:14.419 Uttam Kumaran: Okay, blah, blah, blah, blah, blah, blah, blah, blah, blah…

242 00:24:15.080 ⇒ 00:24:17.880 Uttam Kumaran: Let’s talk about Honey Stinger.

243 00:24:18.000 ⇒ 00:24:31.750 Uttam Kumaran: So, for Honey Stinger, yeah, we… right now, maybe I’ll give you… I’ll give you my take, and then I’ll give some recommendations, and we can brainstorm. Right now, all of their Amazon and Walmart is all,

244 00:24:32.480 ⇒ 00:24:46.950 Uttam Kumaran: like, it’s all 1P, like, so they don’t have the customer data sets, like, for pool parts, they were fulfilling a lot on Amazon, so we were able to get through Amazon Seller Central. These guys are just a vendor, and so we’re not able to get

245 00:24:47.510 ⇒ 00:24:52.689 Uttam Kumaran: customer data. I guess, Sam, what did we… what do we have in there?

246 00:24:52.890 ⇒ 00:24:55.950 Uttam Kumaran: That you saw, that is coming from Amazon and Walmart.

247 00:24:56.670 ⇒ 00:24:58.040 Samuel Roberts: Yeah, it looked like…

248 00:24:58.370 ⇒ 00:25:06.709 Samuel Roberts: maybe some inventory stuff, vendor-wise. I can share the… what stinked, but all the order stuff was on the seller. Yeah, there’s nothing there about…

249 00:25:07.340 ⇒ 00:25:11.120 Samuel Roberts: Even transactions or anything that I can tell.

250 00:25:12.570 ⇒ 00:25:22.080 Samuel Roberts: Okay. I don’t know if there’s… there were some reports and things, I didn’t dig too deep into what all the reports were, so there might be some info there, or access to info, but…

251 00:25:22.510 ⇒ 00:25:27.140 Samuel Roberts: I don’t really know what, what’s gonna be helpful there?

252 00:25:27.990 ⇒ 00:25:28.700 Samuel Roberts: Okay.

253 00:25:28.700 ⇒ 00:25:39.990 Uttam Kumaran: So there’s… so there’s kind of, like, two pivots that, like, I’m kind of thinking overall. One is, like, yeah, we put together this Klaviyo analysis. I still don’t think that’s, like, that was not our primary objective.

254 00:25:40.160 ⇒ 00:25:47.430 Uttam Kumaran: like, I think our primary objective was this omnichannel reporting, and so…

255 00:25:47.960 ⇒ 00:25:54.979 Uttam Kumaran: kind of a couple of ways we can go. One is, like, we could just profile access to what we do have.

256 00:25:55.090 ⇒ 00:26:01.020 Uttam Kumaran: And Sam, like, output there would just be, like, one or two slides on, like, what is in…

257 00:26:01.160 ⇒ 00:26:10.230 Uttam Kumaran: like, what do we have from those vendors? Similarly, they gave us access to Retail Connect, so for the meeting today, we could profile, and we could say, like.

258 00:26:10.570 ⇒ 00:26:28.259 Uttam Kumaran: okay, after discovery, like, we actually don’t have a bunch of this stuff, so what can we do with what we do have? Second is, the AI kind of Klaviyo piece. I suggested yesterday that we, like, bring it all via Fivetran to get the message body. Did that end up happening, or is that, like, something… okay, that ended up coming through?

259 00:26:30.430 ⇒ 00:26:31.200 Uttam Kumaran: Do you know?

260 00:26:31.200 ⇒ 00:26:35.020 Samuel Roberts: I don’t think so, I think we can do that. I didn’t realize that was.

261 00:26:35.290 ⇒ 00:26:36.700 Uttam Kumaran: Okay, so, like…

262 00:26:36.700 ⇒ 00:26:37.710 Samuel Roberts: Yeah, I can add that.

263 00:26:37.710 ⇒ 00:26:45.700 Uttam Kumaran: Yeah, we can get the message body working and do that AI message body thing, but again, like, no promises on, like, what the output’s gonna be.

264 00:26:45.700 ⇒ 00:26:46.140 Samuel Roberts: Right.

265 00:26:46.140 ⇒ 00:26:48.939 Uttam Kumaran: I… in terms of, like, the analysis that Henry did.

266 00:26:49.150 ⇒ 00:26:53.790 Uttam Kumaran: again, like, I think that’s kind of stepping too much into just Klaviyo stuff.

267 00:26:54.100 ⇒ 00:27:01.860 Uttam Kumaran: So… we’re kind of at a point where, like, unless I have… unless we have, like.

268 00:27:02.220 ⇒ 00:27:07.759 Uttam Kumaran: Unless we get some kind of interesting data from RetailConnect, sort of omni-channel stitching.

269 00:27:08.120 ⇒ 00:27:10.830 Uttam Kumaran: Is, like, doesn’t seem, like, super clear right now.

270 00:27:11.300 ⇒ 00:27:14.810 Uttam Kumaran: So… Next slide.

271 00:27:15.770 ⇒ 00:27:16.769 Uttam Kumaran: Where we’re at.

272 00:27:18.770 ⇒ 00:27:22.139 Henry Zhao: You have a lot of interesting data in Klaviyo, though.

273 00:27:23.260 ⇒ 00:27:27.250 Uttam Kumaran: But we didn’t… this is where we didn’t get brought in to do email analysis.

274 00:27:27.650 ⇒ 00:27:28.480 Henry Zhao: Yeah, okay.

275 00:27:29.370 ⇒ 00:27:30.870 Uttam Kumaran: That’s why. It’s like…

276 00:27:39.620 ⇒ 00:27:44.940 Robert Tseng: Retail Connect gives us what?

277 00:27:46.840 ⇒ 00:27:50.859 Uttam Kumaran: Yeah, let’s even… maybe let’s just explore this together. I don’t know…

278 00:27:52.120 ⇒ 00:27:53.750 Uttam Kumaran: Let me just see what’s in there.

279 00:28:09.440 ⇒ 00:28:18.359 Henry Zhao: Robert, I think we talked about this in the offers and services meeting, right? Like, there’s a lot of companies have this challenge of connecting online with offline, so…

280 00:28:18.360 ⇒ 00:28:19.000 Robert Tseng: Yeah.

281 00:28:19.770 ⇒ 00:28:21.190 Henry Zhao: Definitely something we should solve for.

282 00:28:22.930 ⇒ 00:28:25.259 Uttam Kumaran: Alright, let’s just… maybe let’s just do a little…

283 00:28:26.420 ⇒ 00:28:28.489 Uttam Kumaran: Let’s see if I can get into this.

284 00:28:32.700 ⇒ 00:28:35.299 Uttam Kumaran: How do I even, like… what did he say?

285 00:28:48.770 ⇒ 00:28:50.389 Uttam Kumaran: Supplier 1.

286 00:28:53.450 ⇒ 00:28:54.930 Uttam Kumaran: Oh, retail link.

287 00:29:00.060 ⇒ 00:29:06.819 Uttam Kumaran: And Sam, while we’re doing this, can you… can you just, describe what you’re seeing in Mother Duck for Amazon, Walmart?

288 00:29:07.210 ⇒ 00:29:10.720 Uttam Kumaran: That way, we can just, like, knock… we can just, like, think through that.

289 00:29:14.310 ⇒ 00:29:15.070 Samuel Roberts: True.

290 00:29:16.370 ⇒ 00:29:17.380 Uttam Kumaran: Oh, shit.

291 00:29:28.130 ⇒ 00:29:35.860 Samuel Roberts: The Walmart never synced anything, because it was the wrong… it’s a different… it’s just the ads, and that was not a connection that we had. I think it was the…

292 00:29:36.280 ⇒ 00:29:38.429 Samuel Roberts: The private one.

293 00:29:39.220 ⇒ 00:29:40.190 Uttam Kumaran: Okay.

294 00:29:40.190 ⇒ 00:29:41.420 Samuel Roberts: On 5chan.

295 00:29:41.850 ⇒ 00:29:43.700 Samuel Roberts: Yeah, the marketplace is a…

296 00:29:45.080 ⇒ 00:29:49.029 Samuel Roberts: Or there was a waitlist for the DSP, which is the one we have access to, but that’s not…

297 00:29:49.700 ⇒ 00:29:51.839 Samuel Roberts: gonna give us orders anyway, I don’t think.

298 00:29:58.670 ⇒ 00:30:02.819 Uttam Kumaran: Amazon, I’m looking now, we’ve got… let me get there.

299 00:30:02.870 ⇒ 00:30:07.790 Samuel Roberts: Up here, so we’ve got all this vendor… Sorry.

300 00:30:08.300 ⇒ 00:30:14.139 Robert Tseng: No, like, the main thing that Byron, his, like, hope is to be able to say, like.

301 00:30:14.260 ⇒ 00:30:20.980 Robert Tseng: A customer came in through one of their, like, other channels, then they ended up

302 00:30:21.470 ⇒ 00:30:26.730 Robert Tseng: coming to Honey… or maybe, like, an ad brought them to HoneySinger.com, and then…

303 00:30:26.940 ⇒ 00:30:41.439 Robert Tseng: they ended up seeing that Amazon has a better price, and so they’re, like, repurchasing, like, on Amazon. Like, I think he just wants to be able to tell, like, he wants to be able to tell that story. I know we can’t stitch… we’re having a hard time stitching the same

304 00:30:41.590 ⇒ 00:30:46.499 Robert Tseng: customer across Amazon and Shopify, because Amazon is one… is 1P fulfilled.

305 00:30:47.290 ⇒ 00:30:54.850 Robert Tseng: But, like, is there another channel where we could do that in? Like… Or… or something, and.

306 00:30:55.040 ⇒ 00:30:59.410 Uttam Kumaran: Well, like, you mentioned we have orders data, right, Sam? Like, do we have timestamps?

307 00:31:00.000 ⇒ 00:31:00.670 Samuel Roberts: I’m trying to, yeah.

308 00:31:00.670 ⇒ 00:31:01.420 Robert Tseng: Yeah, I see.

309 00:31:03.550 ⇒ 00:31:06.810 Robert Tseng: We can go off of timestamps.

310 00:31:14.250 ⇒ 00:31:16.610 Samuel Roberts: Water amount means…

311 00:31:17.880 ⇒ 00:31:20.000 Robert Tseng: This is all the stuff that you get out of Amazon.

312 00:31:20.570 ⇒ 00:31:21.420 Uttam Kumaran: Yeah.

313 00:31:23.120 ⇒ 00:31:27.470 Uttam Kumaran: I mean, so we’re seeing orders… we have… Yeah.

314 00:31:28.470 ⇒ 00:31:30.820 Uttam Kumaran: Repeat purchase reports…

315 00:31:31.870 ⇒ 00:31:32.680 Samuel Roberts: Okay.

316 00:31:45.410 ⇒ 00:31:48.070 Uttam Kumaran: Yeah, these are orders, but .

317 00:31:50.210 ⇒ 00:31:52.710 Samuel Roberts: But as a procurement order, does that mean it’s…

318 00:31:53.090 ⇒ 00:31:55.350 Uttam Kumaran: Yeah, this is the… Those, yeah.

319 00:31:55.350 ⇒ 00:31:56.220 Samuel Roberts: Yeah.

320 00:31:56.680 ⇒ 00:31:57.510 Uttam Kumaran: Yeah.

321 00:31:58.300 ⇒ 00:32:02.490 Uttam Kumaran: That’s the thing, so they’re buying bulk. Oh… Yeah, like…

322 00:32:05.090 ⇒ 00:32:08.869 Samuel Roberts: I think we’re just getting the vendor side shipping, too.

323 00:32:09.170 ⇒ 00:32:10.210 Samuel Roberts: Amazon.

324 00:32:11.080 ⇒ 00:32:12.610 Samuel Roberts: I don’t know if this is…

325 00:32:14.270 ⇒ 00:32:17.279 Uttam Kumaran: This is a retail… oh, this is a retail report, right? Look.

326 00:32:17.420 ⇒ 00:32:19.149 Uttam Kumaran: Retail report daily.

327 00:32:19.320 ⇒ 00:32:25.420 Uttam Kumaran: Start date, customer returns, order revenue, ordered units, shipping COGS,

328 00:32:26.200 ⇒ 00:32:29.130 Uttam Kumaran: So what… yeah, what is this? This is… this seems like order data.

329 00:32:30.260 ⇒ 00:32:34.429 Samuel Roberts: It seems like… is it summarized data? Is that what it is? By day?

330 00:32:35.790 ⇒ 00:32:44.329 Uttam Kumaran: Yeah, it’s, daily… Order, like, the amount of stuff that was ordered, Looks like… yeah.

331 00:33:01.990 ⇒ 00:33:10.059 Robert Tseng: So these are all just, like, aggregated models that are just off of, like, day, week, month. They don’t actually give us, like, transaction or order level data.

332 00:33:11.160 ⇒ 00:33:12.260 Samuel Roberts: I don’t think so.

333 00:33:13.250 ⇒ 00:33:14.720 Samuel Roberts: Yeah, ship units…

334 00:33:16.150 ⇒ 00:33:21.040 Uttam Kumaran: Yeah, but there is stuff about, like… looks like there is some stuff about traffic.

335 00:33:21.370 ⇒ 00:33:22.700 Uttam Kumaran: to the page.

336 00:33:22.820 ⇒ 00:33:27.540 Uttam Kumaran: There is some stuff about sourcing. I will have to look at the definitions of these.

337 00:33:27.780 ⇒ 00:33:30.939 Uttam Kumaran: There is some stuff about…

338 00:33:32.160 ⇒ 00:33:36.959 Uttam Kumaran: I mean, I don’t know, dude, we could totally put some shit together about all this Amazon shit.

339 00:33:38.900 ⇒ 00:33:39.770 Uttam Kumaran: Right.

340 00:33:40.380 ⇒ 00:33:41.040 Robert Tseng: Yeah.

341 00:33:43.340 ⇒ 00:33:48.369 Uttam Kumaran: Oh, yeah, okay, so I… I don’t know, I feel I’m, like, a lot less worried right now, now that I see it.

342 00:33:48.370 ⇒ 00:33:51.739 Samuel Roberts: Yeah, I think it’s just not gonna be the customer stuff, right? It’s just gonna be…

343 00:33:51.740 ⇒ 00:33:53.449 Robert Tseng: Order, quantity details?

344 00:33:54.430 ⇒ 00:33:55.360 Robert Tseng: What is that?

345 00:33:55.360 ⇒ 00:33:55.720 Uttam Kumaran: Where?

346 00:33:55.720 ⇒ 00:33:57.030 Robert Tseng: quantities.

347 00:33:57.490 ⇒ 00:34:00.630 Robert Tseng: order quantity details is… Oh, okay, yeah.

348 00:34:04.860 ⇒ 00:34:07.169 Robert Tseng: Okay, well, I mean, we get order numbers there.

349 00:34:08.219 ⇒ 00:34:15.369 Robert Tseng: These are purchase orders, oh, I see. This is more on the… okay, this is what they’re sending…

350 00:34:15.580 ⇒ 00:34:16.330 Samuel Roberts: Yeah.

351 00:34:16.330 ⇒ 00:34:17.330 Robert Tseng: Amazon. Okay.

352 00:34:17.330 ⇒ 00:34:18.479 Samuel Roberts: I think so.

353 00:34:35.179 ⇒ 00:34:40.809 Uttam Kumaran: Okay, but I guess this is my point, like, we should have gone through all this stuff, like, I just, I’m like, alright, whatever.

354 00:34:43.839 ⇒ 00:34:49.729 Uttam Kumaran: So, okay, I mean, there’s more than enough stuff here to put a story about Amazon as a channel.

355 00:34:50.079 ⇒ 00:34:53.359 Uttam Kumaran: Like, do you think that’s relevant for him, Robert, or at least we can…

356 00:34:53.579 ⇒ 00:34:56.339 Uttam Kumaran: show comparisons between Amazon and Shopify.

357 00:34:57.050 ⇒ 00:34:57.710 Robert Tseng: Yeah.

358 00:34:57.820 ⇒ 00:35:01.929 Robert Tseng: Differences in customer behavior between these two platforms. Yeah.

359 00:35:03.040 ⇒ 00:35:04.730 Uttam Kumaran: And then what’s… where is,

360 00:35:05.360 ⇒ 00:35:08.159 Uttam Kumaran: Where’s the Walmarts? Is there anything from Walmart in here?

361 00:35:08.360 ⇒ 00:35:14.269 Samuel Roberts: No, there’s nothing coming in from Walmart, because the only thing we had access to is that private ESP connection on FiveTran.

362 00:35:14.970 ⇒ 00:35:19.730 Uttam Kumaran: But does that have any of the vent… like, even equivalent vendor stuff from Amazon?

363 00:35:22.740 ⇒ 00:35:27.729 Samuel Roberts: I couldn’t set it up because there’s a waitlist on Vibean, and I joined that waitlist, but I can’t…

364 00:35:28.470 ⇒ 00:35:28.980 Samuel Roberts: Sync in.

365 00:35:28.980 ⇒ 00:35:31.960 Uttam Kumaran: But we should have called… but we should call someone from FiveTrain.

366 00:35:32.370 ⇒ 00:35:37.370 Uttam Kumaran: Can you… yeah, I mean, I have… we have, like, a bunch of contacts there. I can call them.

367 00:35:37.370 ⇒ 00:35:38.730 Henry Zhao: Polyatomic do it?

368 00:35:39.080 ⇒ 00:35:40.340 Henry Zhao: Compile timing to it?

369 00:35:41.540 ⇒ 00:35:47.290 Uttam Kumaran: No, they don’t… I mean, you’re… it’s gonna take a… it’s gonna take, like, a week, so…

370 00:35:47.680 ⇒ 00:35:55.880 Uttam Kumaran: Can you, can you submit a ticket, Sam? And then send a ticket in this channel, and then I will connect you with someone from Fivetran.

371 00:35:56.110 ⇒ 00:35:56.729 Uttam Kumaran: Cool, for sure.

372 00:35:56.730 ⇒ 00:35:59.689 Samuel Roberts: Yeah, okay, I think I did that…

373 00:35:59.880 ⇒ 00:36:02.400 Samuel Roberts: Alright, I did the waitlist, I’ll see if I can set a…

374 00:36:02.640 ⇒ 00:36:07.199 Uttam Kumaran: Yeah, we just have to, like, skip… that’s just all for the… For the casuals.

375 00:36:08.020 ⇒ 00:36:08.720 Samuel Roberts: Okay.

376 00:36:09.490 ⇒ 00:36:11.119 Samuel Roberts: Yeah, I’ll make a ticket.

377 00:36:12.690 ⇒ 00:36:13.680 Uttam Kumaran: Okay, so…

378 00:36:13.680 ⇒ 00:36:14.730 Robert Tseng: Is the mother duck actually…

379 00:36:14.730 ⇒ 00:36:15.250 Uttam Kumaran: Dude.

380 00:36:15.760 ⇒ 00:36:17.160 Robert Tseng: It’s in Honey Stinger?

381 00:36:18.710 ⇒ 00:36:21.140 Robert Tseng: Oh, there’s an access key? I… okay.

382 00:36:27.600 ⇒ 00:36:31.290 Uttam Kumaran: You’re already in here, but oh, it’s… you could actually just, like, I think…

383 00:36:31.290 ⇒ 00:36:32.279 Robert Tseng: Oh, or just water issues.

384 00:36:32.280 ⇒ 00:36:33.410 Uttam Kumaran: Run this.

385 00:36:34.390 ⇒ 00:36:35.850 Uttam Kumaran: Yeah, so…

386 00:36:36.550 ⇒ 00:36:37.260 Robert Tseng: Okay.

387 00:36:41.360 ⇒ 00:36:47.650 Robert Tseng: Yeah, I mean, I’ll poke around in Amazon and come up with a story that, like, what we can do in there, like, if that helps us buy it another week.

388 00:36:51.210 ⇒ 00:36:53.189 Uttam Kumaran: Yeah, I just don’t know,

389 00:36:54.010 ⇒ 00:36:58.639 Uttam Kumaran: I mean, at this point, I think he… he also just doesn’t know what he has access to.

390 00:36:58.790 ⇒ 00:37:02.869 Uttam Kumaran: So I think one is, like, we could put a slide on, like, just what we profiled.

391 00:37:03.260 ⇒ 00:37:10.709 Uttam Kumaran: I mean, I would love… Yeah, so on Amazon, on Shopify, like.

392 00:37:11.470 ⇒ 00:37:14.859 Uttam Kumaran: Yeah, we have all the usual stuff,

393 00:37:15.130 ⇒ 00:37:22.379 Uttam Kumaran: We could… we could look at, like, let’s see.

394 00:37:22.380 ⇒ 00:37:37.780 Robert Tseng: So, like, I guess it’s okay if we don’t do the stitching. So, like, I mean, I just want to say there’s, like, a couple creative ways. If we really understand customer behavior across these two platforms, what they can do… what he can do is, like, he can… he can test, like, how much, like,

395 00:37:38.240 ⇒ 00:37:56.639 Robert Tseng: he could… he can run some experiments, like, here, promotion for Shopify, or promotion on, for whoever he has emails for, I guess. They could run a campaign, like, promotion on Shopify matching Amazon prices, or whatever. And, like, then he can see, like.

396 00:37:56.680 ⇒ 00:38:15.079 Robert Tseng: like, a week, he could track over the course of the week of that campaign if, like, the customer order, like, kind of shifts, and, like, more Amazon customers, maybe Amazon volume kind of comes down a little bit, and they shift over to Shopify, and then we’d be able to make some assumptions there about, like.

397 00:38:15.080 ⇒ 00:38:30.229 Robert Tseng: okay, when you’re actually at the same price, like, I mean, it’s probably not… that’s probably not true, but if it’s, like, if when you push Shopify pricing to the same price, or slightly lower than Amazon, you drive, you know, X percentage of customers to Shopify.

398 00:38:30.230 ⇒ 00:38:37.089 Robert Tseng: It’s better for the business because it’s still a higher margin profile, you lose 30% just by doing Amazon 1P.

399 00:38:37.090 ⇒ 00:38:45.859 Robert Tseng: And that’s a lever that he doesn’t… that he’s ever seen before, right? So, like, I think there are still interesting, like, cross-platform things that we could do if we just are able to

400 00:38:45.940 ⇒ 00:39:00.270 Robert Tseng: profile very clearly, like, what does a typical Shopify customer do, what are they doing on Amazon? We could make the… and yeah, like, even if it’s not, like, a clear stitching, like, we can still do a good analysis like that.

401 00:39:01.850 ⇒ 00:39:02.450 Uttam Kumaran: Okay.

402 00:39:02.800 ⇒ 00:39:08.240 Uttam Kumaran: Yeah, then, given, like, the rush today, I may just, like.

403 00:39:08.640 ⇒ 00:39:16.690 Uttam Kumaran: Like, one is, like, we’re not getting nearly enough stuff from Shopify, from,

404 00:39:16.970 ⇒ 00:39:24.429 Uttam Kumaran: polyatomic. Like, I don’t have any search terms or related info, so I’m… Sam, can we sync through Fivetran?

405 00:39:25.100 ⇒ 00:39:25.750 Samuel Roberts: Yeah, that’s.

406 00:39:25.750 ⇒ 00:39:30.969 Uttam Kumaran: Can you go ahead and… can you go ahead and start that? Yep. I don’t know how long that’s gonna take.

407 00:39:31.150 ⇒ 00:39:36.459 Uttam Kumaran: And then… Yeah, maybe,

408 00:39:38.000 ⇒ 00:39:46.429 Uttam Kumaran: I guess it’s gotta be, like, me, Henry, or Amber, like, we should just, like, stay on and try to… because we just have to get something for this by… by 2.

409 00:39:48.590 ⇒ 00:39:50.850 Uttam Kumaran: So… I don’t…

410 00:39:51.720 ⇒ 00:39:55.350 Uttam Kumaran: Yeah, I think if we can all stay on and we can talk through that, that would be…

411 00:39:56.260 ⇒ 00:39:59.399 Uttam Kumaran: That would be great, and we could just kind of work on that until then.

412 00:40:02.460 ⇒ 00:40:05.680 Samuel Roberts: I just made another request for the beta access.

413 00:40:05.680 ⇒ 00:40:06.110 Uttam Kumaran: Okay, cool.

414 00:40:06.110 ⇒ 00:40:10.899 Samuel Roberts: for that Walmart thing, but I don’t get any confirmation back, so I don’t have anything to share.

415 00:40:11.080 ⇒ 00:40:14.929 Uttam Kumaran: No, no, no, support a submit… support a submit support request.

416 00:40:16.280 ⇒ 00:40:20.249 Samuel Roberts: Oh, see, all I was seeing there was feature requests for support, but let me… let me dig in.

417 00:40:20.250 ⇒ 00:40:24.049 Uttam Kumaran: No, no, no, submit, like, submit one of the Zendesk requests, and then send that to me.

418 00:40:24.050 ⇒ 00:40:24.660 Samuel Roberts: Got it.

419 00:40:24.960 ⇒ 00:40:28.950 Uttam Kumaran: I will, I’ll… I’ll ping some people.

420 00:40:29.940 ⇒ 00:40:32.000 Uttam Kumaran: Just put it as P1 or whatever.

421 00:40:33.370 ⇒ 00:40:34.380 Samuel Roberts: Okay.

422 00:40:35.170 ⇒ 00:40:41.810 Uttam Kumaran: Okay, cool. Then maybe, yeah, I mean, maybe what’s helpful…

423 00:40:42.110 ⇒ 00:40:47.110 Uttam Kumaran: we have about, what, meetings at 2 Central, so…

424 00:40:47.670 ⇒ 00:40:56.869 Uttam Kumaran: It would be great, maybe, for the next, like… I mean, I’m… I’m free for another hour or so, but we can talk… we can just talk Eden briefly, if, like, the…

425 00:40:57.340 ⇒ 00:41:03.309 Uttam Kumaran: maybe me, you, Sam, Henry Amber, we just tried to put something together.

426 00:41:03.960 ⇒ 00:41:06.019 Uttam Kumaran: What do you think, Robert? .

427 00:41:07.660 ⇒ 00:41:10.889 Robert Tseng: I mean, I… I feel like I can… I can… I can jump.

428 00:41:10.890 ⇒ 00:41:11.270 Uttam Kumaran: We’re doing.

429 00:41:11.270 ⇒ 00:41:16.709 Robert Tseng: If I spend, like, an hour or two hours, like, I can get… I can get Byron… I can get docking points for Byron.

430 00:41:17.610 ⇒ 00:41:19.930 Uttam Kumaran: Do you have time between now and then?

431 00:41:20.600 ⇒ 00:41:22.469 Robert Tseng: Yeah, yeah, I’ll do it. I mean, this one’s…

432 00:41:22.470 ⇒ 00:41:33.009 Uttam Kumaran: I would… I would suggest, like, you… you… I mean, just… you four take it then. Like, that way, if you start to get nuggets, you can have them help with producing the deck.

433 00:41:33.600 ⇒ 00:41:35.130 Uttam Kumaran: Or… or whatever.

434 00:41:35.130 ⇒ 00:41:46.369 Robert Tseng: Yeah, I mean, I saw things like, Henry, you gotta respond to Rebecca, it’s been, like, 2 weeks, and she has no idea what’s going on for Medica. So, like, there’s, like, stuff on Eden, like, I don’t want you to drop. Okay. So I… like, I’d rather just…

435 00:41:46.370 ⇒ 00:41:46.750 Uttam Kumaran: Okay.

436 00:41:46.750 ⇒ 00:41:51.850 Robert Tseng: I mean, Amber can… Amber can join if he wants, but, like, maybe me, Sam, and Amber. I don’t think Henry should.

437 00:41:51.850 ⇒ 00:41:52.510 Uttam Kumaran: Okay.

438 00:41:52.940 ⇒ 00:41:53.290 Henry Zhao: Okay.

439 00:41:53.630 ⇒ 00:41:54.230 Uttam Kumaran: Okay.

440 00:41:54.520 ⇒ 00:42:01.380 Henry Zhao: What should I say to Rebecca, though? Should I just run with the API docs that we already have, or do I need to talk to…

441 00:42:01.500 ⇒ 00:42:06.640 Henry Zhao: Michelle about the on-site… like, was she firm in that we need to go to Wisconsin, or…

442 00:42:07.700 ⇒ 00:42:21.919 Robert Tseng: Well, I don’t know, Davey, I need you to go and, like, insert yourself, like, I just called her, and she was, like, fine talking to me. So, like, I don’t know, like, you… if you need to talk to Michelle, you can talk to Michelle. I don’t know where Rebecca’s coming from. You did the virtual onsite with them already, so, like.

443 00:42:22.080 ⇒ 00:42:27.570 Robert Tseng: I don’t know, I didn’t really have that… I don’t… I don’t think I nearly have as much context as you do.

444 00:42:27.570 ⇒ 00:42:29.860 Henry Zhao: Should I introduce myself to Michelle, saying, like.

445 00:42:30.260 ⇒ 00:42:37.500 Henry Zhao: I talked to Robert, and he said you wanted us to do an onsite, or just be like, can I already get the data from this API?

446 00:42:38.450 ⇒ 00:42:39.010 Robert Tseng: Hi.

447 00:42:39.620 ⇒ 00:42:57.799 Robert Tseng: Yeah, well, I mean, I think Rebecca just needs a timeline of, like, well, what is she going to be able to get out of it, and when? Like, I think that’s… she’s asked, like, kind of multiple times, so your stakeholder is her, like, Michelle will just… like, I don’t think that’s not… I think your… the goal is just to get Rebecca some clarity on, like.

448 00:42:57.990 ⇒ 00:43:08.519 Robert Tseng: what are the next steps for the Pharmedica thing, right? Kind of like how Zoron is building out the whole roadmap for the marketing analytics side, you kind of need to do that on the farm ops side.

449 00:43:08.690 ⇒ 00:43:18.850 Robert Tseng: I know there’s a kickoff call with Brad and everything, too, so… I mean, that whole, like, pharmacy analytics stuff is, like, that’s yours to own, so.

450 00:43:18.850 ⇒ 00:43:19.250 Henry Zhao: Yeah, yeah.

451 00:43:19.250 ⇒ 00:43:19.840 Robert Tseng: Yeah.

452 00:43:20.640 ⇒ 00:43:27.819 Robert Tseng: I’m not gonna be on that call with Eden, I’m gonna be on a different call, so, like, I… I don’t know what you’re gonna say to Brad and team today, but yeah.

453 00:43:29.790 ⇒ 00:43:35.639 Henry Zhao: Yeah, Brad hasn’t been responding to me, so, like, I’ve been trying to meet with him, like, all week to talk about this, but, yeah.

454 00:43:35.950 ⇒ 00:43:54.549 Robert Tseng: Well, I mean, this is the kickoff call, or this is something, like, I don’t know, like, you gotta… yeah, we gotta… we gotta get or get organized around them. Like, to Danny… Danny, this is… the pharmacy stuff is Danny’s highest priority, so if, you know, you don’t want Rebecca or Brad kind of bubbling up to him that, like, oh, we don’t know what the heck we’re doing, so, like.

455 00:43:54.670 ⇒ 00:44:04.319 Robert Tseng: I don’t know, I think there… there is a conversation that needs to be had, it’s gonna happen today. I think… I mean, do you… can you facilitate that conversation? Do you feel like you have enough…

456 00:44:04.320 ⇒ 00:44:04.980 Henry Zhao: Thanks again.

457 00:44:04.980 ⇒ 00:44:08.989 Robert Tseng: Have sex with… okay. Definitely. But yeah, I guess he’ll see what happens.

458 00:44:08.990 ⇒ 00:44:15.999 Henry Zhao: Yeah, I’ve been in contact with them that, like, pharmacy was held off because we had all this attribution and capitalist stuff that we had to work on, so they know that.

459 00:44:16.000 ⇒ 00:44:20.080 Robert Tseng: I understand that. Okay, but yeah, now it’s like, that’s not an excuse, we have the… we have the roadmap.

460 00:44:20.080 ⇒ 00:44:20.809 Henry Zhao: Yeah, now we know.

461 00:44:20.810 ⇒ 00:44:23.010 Robert Tseng: I should be putting the remote, yeah. Okay.

462 00:44:23.010 ⇒ 00:44:33.460 Henry Zhao: But Awasht, I need to be prepared for this call in one aspect, where… remember, Awasht, you said that you were gonna be able to build a Dagster script to automate the SLA report for…

463 00:44:33.620 ⇒ 00:44:37.240 Henry Zhao: Katie, what is the explanation for why we ended up not doing that?

464 00:44:40.230 ⇒ 00:44:43.190 Henry Zhao: Because they’re gonna ask about that, and I wanted to be prepared.

465 00:44:46.570 ⇒ 00:44:48.590 Awaish Kumar: I’m sorry, I was rude.

466 00:44:48.960 ⇒ 00:44:55.600 Awaish Kumar: So we… basically, we didn’t need a macro for that, right? That could have been done in a…

467 00:44:57.680 ⇒ 00:45:04.770 Awaish Kumar: in a DVD model itself, if we had the data, So, like, the thing…

468 00:45:04.770 ⇒ 00:45:05.630 Henry Zhao: Don’t have the data.

469 00:45:06.070 ⇒ 00:45:08.840 Awaish Kumar: Like, yeah, the fields like pharmacy status.

470 00:45:08.960 ⇒ 00:45:14.610 Awaish Kumar: that you can see in the Basque UI, but it’s not possible to bring it in

471 00:45:14.910 ⇒ 00:45:17.729 Awaish Kumar: Through the webhooks in our warehouse.

472 00:45:17.840 ⇒ 00:45:25.740 Awaish Kumar: So, like, that status, pharmacy status, is not there. Current status is there, which we already have in order somebody

473 00:45:26.020 ⇒ 00:45:29.200 Awaish Kumar: And like, you can…

474 00:45:29.200 ⇒ 00:45:29.550 Henry Zhao: But that…

475 00:45:29.850 ⇒ 00:45:32.720 Awaish Kumar: To build the UI, right?

476 00:45:32.870 ⇒ 00:45:33.750 Awaish Kumar: I don’t know, like.

477 00:45:33.750 ⇒ 00:45:37.540 Henry Zhao: But that’s why we wanted the DAXR script, is because of those columns that are not there. Does that make sense?

478 00:45:38.240 ⇒ 00:45:42.960 Awaish Kumar: Yeah, but if the data is not there, like, what Texas group has to do, like.

479 00:45:43.950 ⇒ 00:45:55.130 Awaish Kumar: the data, we are not getting it from BASC, and there’s no other way to get data from BASC, except using webhooks, and there’s no webbook which gives us the pharmacy status.

480 00:45:57.030 ⇒ 00:46:02.440 Henry Zhao: Well, what I don’t understand about this, like, vial size and all this stuff is why is it so hard for BAS to implement it? Is it not just…

481 00:46:02.560 ⇒ 00:46:05.699 Henry Zhao: Adding columns to something?

482 00:46:06.590 ⇒ 00:46:11.709 Henry Zhao: Like, what is so tricky? What is this taking Basque so long to just add it, to give it this data?

483 00:46:12.170 ⇒ 00:46:17.970 Awaish Kumar: Yeah, I… I think they haven’t, maybe, don’t have it in, you know.

484 00:46:18.780 ⇒ 00:46:22.780 Awaish Kumar: like, model in a way, their model, maybe…

485 00:46:22.780 ⇒ 00:46:34.329 Robert Tseng: exists somewhere else, right? Is there a way that we can bring it in from a separate, like, I… I mean, when I go in the UI, like, I can go and select file size, like, or whatever, set it on the… on the product.

486 00:46:35.250 ⇒ 00:46:39.270 Awaish Kumar: Isn’t it automatically being selected by…

487 00:46:39.380 ⇒ 00:46:43.019 Awaish Kumar: VAR score is somewhat manually selecting vial sizes.

488 00:46:43.420 ⇒ 00:46:51.739 Robert Tseng: It’s not, like, manually every order, but it’s pre-configured that in the products, that they’re coming out in a particular vial size, in whatever quantity, and

489 00:46:52.060 ⇒ 00:47:07.139 Robert Tseng: So, is this just, like, a product mappings problem, where, like, the product mapping sheet that Demlade has, like, it’s just missing file size? Maybe there’s a product list that they have in a Google Sheet or somewhere that we can pull in file size from? Like, why do we have to wait for

490 00:47:07.140 ⇒ 00:47:17.930 Robert Tseng: like, it’s gonna be similar to, like, the whole COGS thing, right? Where it is gonna be based off an assumption, but it’s, like, file size is not changing that often, like, it should be fine, right?

491 00:47:18.440 ⇒ 00:47:23.450 Demilade Agboola: It does change the offer, because for the same product… Okay. Yeah.

492 00:47:23.550 ⇒ 00:47:32.359 Demilade Agboola: So, for the same products, you can have the 0.4, 0.8, 1.2, 5, 2.5, whatever.

493 00:47:32.360 ⇒ 00:47:51.789 Robert Tseng: Okay, I remember that. Then what… isn’t… isn’t the name of the variant ID… aren’t they adjusted already by vial size? So, like, can’t we just… can we assume off of one of the naming conventions of the product itself on what the vial size is? Don’t we create a different variant ID for every… every product size… every… every different… every time there’s a new size?

494 00:47:53.050 ⇒ 00:47:59.179 Demilade Agboola: So we have… it’s not necessarily by sizes, though. That’s the problem.

495 00:47:59.400 ⇒ 00:48:09.580 Demilade Agboola: So, like, and that’s… that was the issue with COGS, because even when you have the, like, the product, it ended up being a thing of intuition, of being able to say, hey, this is the…

496 00:48:09.640 ⇒ 00:48:22.429 Demilade Agboola: 2.5, this is the 1.2, and he relies so much on human knowledge to be able to be like, this is the size, and therefore that would… that would be the cogs. And so being able to, for the… being able to have, like, the valve size.

497 00:48:23.090 ⇒ 00:48:40.499 Demilade Agboola: just put there in a way that we can associate it directly to the orders will, like, number one, just make some of those processes easier for, like, Henry, but number two, for things like COGS, will just be easy to be able to broadcast the price across the product and the value size.

498 00:48:43.480 ⇒ 00:48:52.829 Robert Tseng: Okay, so is it the prob… the problem is that from the order data itself, we cannot tell what the vial size should be, and so we’re just, like, getting something wrong because

499 00:48:53.210 ⇒ 00:48:57.809 Robert Tseng: like, I don’t know, our COGS assumptions are off, and, like, the quantities are… is that… is that what this is? Like…

500 00:48:58.030 ⇒ 00:49:15.140 Henry Zhao: But wait, we’re also spending too much time on vial size, the issue is not just vial size, it’s like, everything that they’ve requested, there’s, like, one or two columns that I need that I don’t have from BASC, and I’m like, is this good enough? Like, for example, I have shipping date, I have sent pharmacy date, but they’re like, I need the date that it was filled by the pharmacy.

501 00:49:15.770 ⇒ 00:49:22.899 Henry Zhao: Or there was, like, sometimes pharmacies get transferred over, and I don’t have data on when… on if a medicine was transferred to a different pharmacy.

502 00:49:23.320 ⇒ 00:49:24.510 Robert Tseng: Yeah, I guess where I’m trying to…

503 00:49:24.510 ⇒ 00:49:26.309 Henry Zhao: Drive, or they like…

504 00:49:26.690 ⇒ 00:49:35.139 Robert Tseng: yeah, I’m just… make the prox… like, you… you come up with the proxy. Just say, this is… this is the best that we can do. And if they need to adjust it, we can adjust it later, right? Like, this is…

505 00:49:35.660 ⇒ 00:49:36.520 Robert Tseng: Yeah.

506 00:49:36.520 ⇒ 00:49:40.209 Henry Zhao: Today, Judd. Today, Judd requested some date that, like.

507 00:49:40.400 ⇒ 00:49:52.709 Henry Zhao: it was off for 20 people, and I was like, is this good enough? I can proxy it with this. He said, okay, so I pulled it, it was done. Like, today in the meeting, I’m just gonna say, we need to communicate better, this is what’s missing, here’s the plan to get that resolved, and we’re gonna move forward that way.

508 00:49:53.220 ⇒ 00:49:59.130 Henry Zhao: But I just needed the day… I just needed context from Awash on what he did, so I could prepare for this call fully.

509 00:50:01.520 ⇒ 00:50:15.309 Robert Tseng: Okay, yeah, we’re always gonna run through this pattern where they’re gonna ask us for things that we cannot… we don’t control the product, we can’t create the new fields. So, as much as we can create the proxy and push it, like, I think that’s our first reaction.

510 00:50:15.380 ⇒ 00:50:23.590 Robert Tseng: And then, like, yeah, we can continue to ask for, like, the higher fidelity data point, but we’re not going to get it immediately. That’s out of our control. So, I feel like…

511 00:50:23.590 ⇒ 00:50:23.980 Henry Zhao: Yeah, just another.

512 00:50:23.980 ⇒ 00:50:25.030 Robert Tseng: That’s how we should manage it.

513 00:50:25.030 ⇒ 00:50:25.590 Henry Zhao: pardon?

514 00:50:26.020 ⇒ 00:50:30.859 Henry Zhao: Yeah, just another context to give you, like, Katie asked for something that she spends is very high pride and very urgent.

515 00:50:31.030 ⇒ 00:50:35.320 Henry Zhao: To automate, basically, a report that she pulls from BASC and then just filters 3 fields.

516 00:50:35.530 ⇒ 00:50:37.070 Henry Zhao: Like, it was gonna be a lot of.

517 00:50:37.070 ⇒ 00:50:38.070 Robert Tseng: Yeah, you don’t need to automate that.

518 00:50:38.070 ⇒ 00:50:47.539 Henry Zhao: Because we don’t have those columns. So, yeah, if they push back, like, just understand that context. It’s like, it’s not that high pride. Like, she has the data, she’s able to do it by just filtering 3 columns.

519 00:50:48.680 ⇒ 00:50:52.589 Henry Zhao: And we had talked before that, like, our job is not to be, like, a support role, to, like.

520 00:50:52.880 ⇒ 00:50:55.430 Henry Zhao: do trivial things, right? We need to also prioritize, so…

521 00:50:55.430 ⇒ 00:50:56.360 Robert Tseng: Yeah.

522 00:50:57.080 ⇒ 00:50:59.800 Henry Zhao: Okay, just wanted to make sure that everyone was aligned and aware of that.

523 00:51:00.240 ⇒ 00:51:01.910 Henry Zhao: They haven’t said anything about it, so…

524 00:51:02.100 ⇒ 00:51:05.499 Henry Zhao: I don’t… behind our back, I don’t want them to be like, oh, this… they never automated this.

525 00:51:06.830 ⇒ 00:51:07.670 Robert Tseng: Okay.

526 00:51:11.370 ⇒ 00:51:14.760 Henry Zhao: How do I feel, do we want to do… Yeah, go ahead.

527 00:51:14.930 ⇒ 00:51:17.869 Uttam Kumaran: Do you want to… when’s that meeting? Do you want to do a little pre-meeting?

528 00:51:18.630 ⇒ 00:51:22.130 Uttam Kumaran: I would like that. I have zero context, so I could be a good guinea pig.

529 00:51:23.300 ⇒ 00:51:24.029 Uttam Kumaran: When does that mean?

530 00:51:24.030 ⇒ 00:51:27.079 Henry Zhao: Yeah, the meeting… the meeting’s at 2 Eastern today.

531 00:51:29.030 ⇒ 00:51:29.890 Henry Zhao: So, right?

532 00:51:29.890 ⇒ 00:51:31.160 Uttam Kumaran: Oh, nice, okay, yeah.

533 00:51:31.270 ⇒ 00:51:37.240 Uttam Kumaran: Let’s, let’s just do, like, towards the end of the retro, I feel like we’re not gonna use the whole hour, so…

534 00:51:37.760 ⇒ 00:51:39.600 Henry Zhao: Okay. We’ll just hop on, and then…

535 00:51:40.880 ⇒ 00:51:41.600 Uttam Kumaran: Yeah.

536 00:51:41.600 ⇒ 00:51:46.800 Henry Zhao: I’m doing the… I’m leading the retro beginning today, so I can make mine a little bit faster, so we have time for this also.

537 00:51:47.540 ⇒ 00:51:49.080 Uttam Kumaran: Okay. Yeah.

538 00:51:53.510 ⇒ 00:51:58.070 Uttam Kumaran: All I know is this is big meeting pharmacy Analytics, so I just want to make sure you’re set up.

539 00:52:00.790 ⇒ 00:52:02.559 Uttam Kumaran: And, do you want to take Amber?

540 00:52:02.560 ⇒ 00:52:11.040 Henry Zhao: Yeah, last week, Brad, who was, like, leading this project, was also traveling, so… so Brad was traveling, Rebecca was traveling, we were working on attribution, so…

541 00:52:11.650 ⇒ 00:52:16.080 Henry Zhao: Yeah, it’s not like it was stalled for 2 weeks, it was stalled for, like, reason. Yeah.

542 00:52:16.880 ⇒ 00:52:17.650 Uttam Kumaran: Okay, cool.

543 00:52:21.330 ⇒ 00:52:22.070 Amber Lin: Cool.

544 00:52:22.230 ⇒ 00:52:27.910 Amber Lin: Robert, Sam, are we using this room? Do you want to go to another meeting room?

545 00:52:29.130 ⇒ 00:52:32.140 Robert Tseng: Yeah, I mean, we’ve got a different group, all of the ones.

546 00:52:32.980 ⇒ 00:52:33.350 Amber Lin: Okay.

547 00:52:35.520 ⇒ 00:52:41.680 Uttam Kumaran: Okay. Thanks, guys. Sam, do you wanna… do you wanna stay on here for… quickly for that?

548 00:52:41.680 ⇒ 00:52:44.049 Samuel Roberts: Is this a little while on me, yeah. Okay.

549 00:52:44.620 ⇒ 00:52:49.730 Uttam Kumaran: Up to you. I… if you think the team can grind that. I guess, stay here for me 5 minutes, Sam.

550 00:52:49.730 ⇒ 00:52:54.440 Samuel Roberts: Yeah, yeah, once I got those things set up in 5chan, I’m not sure how useful I’ll be.

551 00:52:54.670 ⇒ 00:52:56.750 Samuel Roberts: Over there anyway, but… Okay.

552 00:52:57.170 ⇒ 00:52:57.760 Uttam Kumaran: Okay.

553 00:52:59.180 ⇒ 00:52:59.840 Samuel Roberts: And this would be good.

554 00:53:00.260 ⇒ 00:53:01.250 Samuel Roberts: now, so…

555 00:53:02.250 ⇒ 00:53:09.369 Uttam Kumaran: Yeah, let’s jump… let’s jump to, AI team stuff. Maybe Gabe, if you… if you want to lead, I think maybe…

556 00:53:09.540 ⇒ 00:53:17.310 Uttam Kumaran: In the interest of just, like, kind of priority, maybe let’s talk a little bit about the…

557 00:53:17.420 ⇒ 00:53:24.839 Uttam Kumaran: Lilo opportunity. I’m just wondering if you guys had a chance to look at that, and

558 00:53:25.760 ⇒ 00:53:28.739 Uttam Kumaran: You know, can… can get some… can get some thoughts.

559 00:53:28.870 ⇒ 00:53:34.440 Uttam Kumaran: On, like, kind of the scope that they sent. And I can talk a little bit about, like, what next steps are, but yeah.

560 00:53:37.320 ⇒ 00:53:44.080 Gabriel Lam: Okay, yeah, I can just quickly go over the updates with the case study assistant,

561 00:53:44.880 ⇒ 00:54:03.919 Gabriel Lam: We are meeting, I believe, the initial roadmap we set out. The MVP is ready and testing. There’s a couple minor features that we are resolving today. One of the goals is to compare the copy that we have here compared to copies that Hannah would have, through her own current existing process.

562 00:54:04.300 ⇒ 00:54:08.020 Gabriel Lam: We’re hoping to get But this…

563 00:54:08.390 ⇒ 00:54:10.620 Gabriel Lam: I think they’re fairly low-hanging fruit.

564 00:54:10.750 ⇒ 00:54:18.270 Gabriel Lam: Mainly to do with, you know, Slack notifications and… the UI, but the prompt…

565 00:54:18.440 ⇒ 00:54:23.289 Gabriel Lam: We think is at a good spot to cover the… both cases of, sort of, you know.

566 00:54:23.390 ⇒ 00:54:25.620 Uttam Kumaran: You say everything in one go, versus…

567 00:54:25.620 ⇒ 00:54:29.480 Gabriel Lam: when I listen to Hannah’s meetings, they’re a little more prescriptive. I think

568 00:54:29.660 ⇒ 00:54:36.450 Gabriel Lam: we’ve balanced both types of interviews. There’s a couple quality-of-life things that we thought about in the future.

569 00:54:36.710 ⇒ 00:54:44.239 Gabriel Lam: But we wanted to get more testing, through Hanna, and also using this project as an example to see if there’s…

570 00:54:44.650 ⇒ 00:54:47.719 Gabriel Lam: Last-minute things that we want to refine and get out.

571 00:54:48.120 ⇒ 00:54:52.999 Gabriel Lam: So that’s sort of a general overview of the case study assistant.

572 00:54:53.730 ⇒ 00:55:05.710 Gabriel Lam: For Lilo, I’ve only had a brief chance to look at the granola notes and the loom. The other guys can hop in here as well. My general thought was the…

573 00:55:05.840 ⇒ 00:55:07.709 Gabriel Lam: The loom looks sort of like…

574 00:55:07.850 ⇒ 00:55:11.230 Gabriel Lam: basically an Excel spreadsheet that is…

575 00:55:11.630 ⇒ 00:55:19.929 Gabriel Lam: made nice and put on a web interface. But I think the granola goes into a little more detail about what else they’re asking for.

576 00:55:20.040 ⇒ 00:55:25.049 Gabriel Lam: Specifically to do with the stand-ups, as well as,

577 00:55:25.300 ⇒ 00:55:31.300 Gabriel Lam: something that they wanted… something I noticed was they wanted… to have…

578 00:55:31.630 ⇒ 00:55:35.530 Gabriel Lam: Iteration cycles and people to be able to ship things quickly.

579 00:55:35.790 ⇒ 00:55:40.419 Gabriel Lam: And so I think that there might be some level of training, or some level of…

580 00:55:40.530 ⇒ 00:55:43.040 Gabriel Lam: Guidance that needs to be there to be like, this is…

581 00:55:43.270 ⇒ 00:55:49.119 Gabriel Lam: These are the workflows that we have right now that allow us to… to really accelerate and have

582 00:55:49.530 ⇒ 00:55:51.659 Gabriel Lam: High velocity when it comes to delivery.

583 00:55:51.940 ⇒ 00:55:55.259 Gabriel Lam: Beyond just giving them a platform.

584 00:55:56.600 ⇒ 00:55:57.970 Gabriel Lam: But that’s just… that’s just me.

585 00:56:00.020 ⇒ 00:56:04.020 Uttam Kumaran: Cool. Yeah, let’s go around the horn. Maybe, Mustafa, you want to go next.

586 00:56:07.710 ⇒ 00:56:18.259 Mustafa Raja: Yeah, I saw, I saw part of the Loom, and, I saw the forecasting sheet that they have, and they, I believe, want to, sort of…

587 00:56:18.460 ⇒ 00:56:27.919 Mustafa Raja: have a custom solution, for that in their own platform. I believe they have the agents generated in the OpenAI canvas.

588 00:56:28.490 ⇒ 00:56:34.029 Mustafa Raja: So, that part, that part sort of looks, a little simpler.

589 00:56:34.170 ⇒ 00:56:40.280 Mustafa Raja: And then I also generated summary of, granola, so…

590 00:56:40.860 ⇒ 00:56:46.960 Mustafa Raja: Yeah, looks, looks like, something that…

591 00:56:47.210 ⇒ 00:56:48.930 Mustafa Raja: We should be able to cater to.

592 00:56:50.160 ⇒ 00:56:50.750 Uttam Kumaran: Okay.

593 00:56:58.110 ⇒ 00:56:59.679 Uttam Kumaran: Casey, you wanna go?

594 00:56:59.680 ⇒ 00:57:04.799 Casie Aviles: Yeah. So yeah, it looks like we’re going to build, like, a custom…

595 00:57:05.210 ⇒ 00:57:08.099 Casie Aviles: Front-end for them, forecasting, right?

596 00:57:08.250 ⇒ 00:57:14.259 Casie Aviles: What I’m curious about is, do we take on, like, also the prediction?

597 00:57:14.560 ⇒ 00:57:20.040 Casie Aviles: Or is that something that they’ve already done? Or do we want to improve on that?

598 00:57:20.670 ⇒ 00:57:21.390 Casie Aviles: Nope.

599 00:57:21.570 ⇒ 00:57:26.560 Casie Aviles: Looks like they have, like, we forecast based on, like, the existing data, right?

600 00:57:28.490 ⇒ 00:57:34.689 Uttam Kumaran: What prediction? Oh… Yeah, I mean, I think they would have the formulas for that, basically.

601 00:57:34.690 ⇒ 00:57:46.089 Samuel Roberts: Yeah, I think it’s pretty much just making that spreadsheet into a UI that I assume, like, other people can use beyond just their team. Is that the idea here? That they want to provide this as a platform to

602 00:57:46.290 ⇒ 00:57:48.010 Samuel Roberts: clients to mess with? Is that…

603 00:57:48.680 ⇒ 00:57:50.470 Uttam Kumaran: No, it’s actually just internal.

604 00:57:50.680 ⇒ 00:57:54.989 Samuel Roberts: It’s just internal, okay, I wasn’t sure if this was something that they were providing out or not. Okay.

605 00:57:55.270 ⇒ 00:57:55.860 Samuel Roberts: Good.

606 00:57:56.100 ⇒ 00:57:56.700 Uttam Kumaran: Yeah, it’s just all…

607 00:57:56.700 ⇒ 00:57:57.330 Samuel Roberts: I’m trying to watch.

608 00:57:57.940 ⇒ 00:58:02.220 Samuel Roberts: Perfect, okay, yeah, but it’s something I want to harden a little bit beyond the spreadsheet, it seems, then.

609 00:58:04.170 ⇒ 00:58:06.589 Casie Aviles: Looks like there’s also, like, a data…

610 00:58:06.890 ⇒ 00:58:12.769 Casie Aviles: Engineering piece there, like, we need to… Get data from certain sources.

611 00:58:13.010 ⇒ 00:58:16.459 Casie Aviles: Kind of similar to what we did with Insania, I guess.

612 00:58:17.620 ⇒ 00:58:20.579 Casie Aviles: Yeah, so I think that’s something we can do.

613 00:58:21.480 ⇒ 00:58:23.739 Casie Aviles: Like, with Polyatonic, I guess, yeah.

614 00:58:25.940 ⇒ 00:58:28.310 Samuel Roberts: Is that something we’d have to sync with the other…

615 00:58:28.930 ⇒ 00:58:33.170 Samuel Roberts: Tables, or would that be something that they would just input into the new…

616 00:58:35.150 ⇒ 00:58:37.859 Samuel Roberts: The new forms for new clients, or something.

617 00:58:39.420 ⇒ 00:58:43.899 Samuel Roberts: Like, they’re looking to fully replace that spreadsheet system with this, is that the idea?

618 00:58:46.110 ⇒ 00:58:52.029 Samuel Roberts: Because I think then it might… I’m not sure how much data syncing we’d have to worry about.

619 00:58:52.240 ⇒ 00:58:52.920 Casie Aviles: Hmm.

620 00:58:53.170 ⇒ 00:58:55.620 Samuel Roberts: If they’re moving fully to this as a…

621 00:58:56.330 ⇒ 00:58:59.059 Samuel Roberts: Forecasting model, rather than using the spreadsheets.

622 00:59:00.150 ⇒ 00:59:05.930 Samuel Roberts: Do we have access to their GitHub yet? I saw that was mentioned, but…

623 00:59:07.440 ⇒ 00:59:11.449 Samuel Roberts: I don’t think that happened yet, as far as I could tell, Lufan.

624 00:59:12.210 ⇒ 00:59:15.749 Uttam Kumaran: No, I haven’t gotten yet, but I can ask them for that today.

625 00:59:15.750 ⇒ 00:59:25.670 Samuel Roberts: Okay, yeah, because I’m curious, I just… I don’t fully have a great view of, like, what the status of everything is, but they seem like they were working with someone else that is not moving fast enough, is that…

626 00:59:26.120 ⇒ 00:59:27.150 Samuel Roberts: Also true.

627 00:59:28.180 ⇒ 00:59:30.720 Uttam Kumaran: Yeah, that’s… that’s basically the case.

628 00:59:31.070 ⇒ 00:59:32.170 Samuel Roberts: Okay, cool.

629 00:59:32.480 ⇒ 00:59:34.739 Samuel Roberts: Hopefully the code is decent.

630 00:59:37.270 ⇒ 00:59:52.730 Samuel Roberts: Great, okay. So besides that forecasting model, were there other pieces that… or is that, like, the big thing they want first? I thought that there was talk about other things for a platform for them, but I wasn’t sure where the focus is gonna be.

631 00:59:53.590 ⇒ 00:59:54.430 Samuel Roberts: Or is that still locked?

632 00:59:55.100 ⇒ 00:59:56.290 Samuel Roberts: Again and find out.

633 00:59:56.290 ⇒ 00:59:58.679 Uttam Kumaran: it’s still kind of dig in, I guess.

634 00:59:58.680 ⇒ 01:00:01.749 Samuel Roberts: Okay. So one thing I can ask for is for the…

635 01:00:02.170 ⇒ 01:00:07.689 Uttam Kumaran: Basically for the code. I think if I go to my…

636 01:00:07.940 ⇒ 01:00:12.230 Uttam Kumaran: open up Slack. Basically, they’re… want this monthly forecast builder.

637 01:00:12.690 ⇒ 01:00:13.500 Samuel Roberts: Right.

638 01:00:14.520 ⇒ 01:00:17.009 Uttam Kumaran: They shared this prototype.

639 01:00:17.370 ⇒ 01:00:17.730 Samuel Roberts: Yep.

640 01:00:17.730 ⇒ 01:00:24.799 Uttam Kumaran: So, like, they kind of just want, like, basically a version of this.

641 01:00:25.010 ⇒ 01:00:26.080 Uttam Kumaran: Yeah.

642 01:00:26.970 ⇒ 01:00:31.710 Uttam Kumaran: You know, so there’s, like, some scope around… Migrating this, or moving this.

643 01:00:33.510 ⇒ 01:00:40.209 Uttam Kumaran: And then, yeah, this is another tracker. So again, I think it’s just, like, kind of simple inputs, and then some…

644 01:00:40.470 ⇒ 01:00:41.160 Samuel Roberts: Yeah.

645 01:00:41.450 ⇒ 01:00:47.769 Uttam Kumaran: some charting. I think the bigger piece here is that they,

646 01:00:48.910 ⇒ 01:00:53.809 Uttam Kumaran: I guess what I’m trying to understand is, like, are we building them these

647 01:00:54.560 ⇒ 01:01:00.170 Uttam Kumaran: Applications, or is it more related, like, how much of this is, like,

648 01:01:01.020 ⇒ 01:01:03.860 Uttam Kumaran: Like, how much of the… both of those leverage AI?

649 01:01:05.900 ⇒ 01:01:06.560 Casie Aviles: Hmm.

650 01:01:06.560 ⇒ 01:01:11.680 Uttam Kumaran: And this, like, third item, which is, like, they want to develop something using the ForePlay API.

651 01:01:14.660 ⇒ 01:01:15.899 Samuel Roberts: Yes, I saw that.

652 01:01:16.380 ⇒ 01:01:17.869 Samuel Roberts: I’m too familiar with that.

653 01:01:18.790 ⇒ 01:01:20.370 Uttam Kumaran: Yeah, so they basically…

654 01:01:20.370 ⇒ 01:01:21.110 Samuel Roberts: Yeah.

655 01:01:21.110 ⇒ 01:01:22.700 Uttam Kumaran: They developed this, like.

656 01:01:22.880 ⇒ 01:01:27.719 Uttam Kumaran: this is, I think, something that is, like, really, really AI-heavy, which is just, like, they want help

657 01:01:28.240 ⇒ 01:01:34.959 Uttam Kumaran: designing, like, visual briefs for some of their ad generations.

658 01:01:35.180 ⇒ 01:01:36.190 Uttam Kumaran: And so…

659 01:01:36.190 ⇒ 01:01:36.860 Samuel Roberts: Hmm.

660 01:01:48.250 ⇒ 01:01:56.189 Uttam Kumaran: this is something that’s more in our wheelhouse, so there’s kind of, like, 3 distinct work streams here. I guess, I guess more on the, like, to kind of talk about, like, what’s…

661 01:01:56.330 ⇒ 01:02:04.099 Uttam Kumaran: What’s next?

662 01:02:04.320 ⇒ 01:02:18.540 Uttam Kumaran: I’m basically gonna put together, like, a document outlining each of these. I guess, like, what technical questions can I highlight on… on each of these that you guys would want to know before we kind of put a timeline together? So right now, I can go ahead and ask…

663 01:02:18.950 ⇒ 01:02:23.069 Uttam Kumaran: for, the GitHub.

664 01:02:28.130 ⇒ 01:02:29.490 Uttam Kumaran: But what else?

665 01:02:38.520 ⇒ 01:02:41.700 Samuel Roberts: So these are gonna be built into the existing…

666 01:02:42.470 ⇒ 01:02:44.480 Uttam Kumaran: Well, I’d like… that’s… I’m not sure yet.

667 01:02:44.770 ⇒ 01:02:47.370 Samuel Roberts: Okay, that’s the other, I think that, like, because, yeah, that will… I mean…

668 01:02:47.650 ⇒ 01:02:55.669 Samuel Roberts: that definitely makes a difference, a little bit. I saw there was some stuff about, like, e-access and moving off AWS and stuff, like, is that all…

669 01:02:55.670 ⇒ 01:03:03.159 Uttam Kumaran: Well, yeah, they… they’re… it’s all under, like, their own managed, like, their other vendors managing AWS, we would just move it to their…

670 01:03:03.900 ⇒ 01:03:04.470 Samuel Roberts: Got it, okay.

671 01:03:04.470 ⇒ 01:03:05.760 Uttam Kumaran: Lilo owned.

672 01:03:06.200 ⇒ 01:03:07.460 Uttam Kumaran: Situation.

673 01:03:07.840 ⇒ 01:03:14.759 Uttam Kumaran: But, like, do you, like… Does anything seem, like, super…

674 01:03:15.940 ⇒ 01:03:26.020 Uttam Kumaran: difficult. If not, like, what I will basically work to do is, like, kind of talk… work on, like, a little bit of a Gantt chart for each of these that we can sort of talk through. I mean.

675 01:03:26.590 ⇒ 01:03:35.699 Uttam Kumaran: again, like, we can only provide the estimate of, like, what we know, but, I kind of want to think about this, like, they do have these, like, features, and so we can isolate that, but…

676 01:03:35.820 ⇒ 01:03:44.420 Uttam Kumaran: roughly, like, do you… are you guys familiar with, like, these… these types of, like, rough… rough charts? Like, we can use recharts or whatever.

677 01:03:44.640 ⇒ 01:03:48.040 Samuel Roberts: I think more of the complication for me is, like, you guys feel comfortable.

678 01:03:48.240 ⇒ 01:03:52.749 Uttam Kumaran: With the backend logic for… Doing this type of data work.

679 01:03:53.210 ⇒ 01:03:55.280 Uttam Kumaran: Like, within the UI, like, what do you think?

680 01:03:59.780 ⇒ 01:04:01.100 Samuel Roberts: Yeah, I think a question…

681 01:04:01.100 ⇒ 01:04:02.420 Uttam Kumaran: Calculating something…

682 01:04:02.420 ⇒ 01:04:03.080 Gabriel Lam: Yeah.

683 01:04:03.080 ⇒ 01:04:06.310 Samuel Roberts: I’m not too worried about that. I think the question becomes, is it… is it…

684 01:04:06.660 ⇒ 01:04:10.210 Samuel Roberts: part of this? Like, how… what is the,

685 01:04:10.810 ⇒ 01:04:27.109 Samuel Roberts: programming language setup look like? There’s some Python workbooks, but if we’re building UIs, then it’s JavaScript, and like, what is the logic front-end, back-end? Like, where is some of this happening? It needs to get thought out. Okay. I think that kind of depends on what… what they already have, and they have some of the workbook here that’s got some stuff.

686 01:04:27.240 ⇒ 01:04:28.570 Samuel Roberts: But, if it’s…

687 01:04:28.860 ⇒ 01:04:40.260 Samuel Roberts: so I just… if we need to migrate some of that or not, if we can just use that or not, because this seems like it was tied to the spreadsheet, so, like, logic… logic-wise, it’s there. The question is, is that…

688 01:04:40.600 ⇒ 01:04:44.580 Samuel Roberts: Something that can be easily pulled in to do.

689 01:04:44.580 ⇒ 01:04:45.130 Uttam Kumaran: Fair.

690 01:04:45.610 ⇒ 01:04:46.020 Samuel Roberts: the stuff.

691 01:04:46.020 ⇒ 01:04:48.300 Uttam Kumaran: Yeah, I’m actually not worried about…

692 01:04:48.600 ⇒ 01:04:58.709 Uttam Kumaran: I’m actually… yeah, okay, so if you’re thinking of it like, okay, we’re gonna take some inputs, we may hit some Python code somewhere to do, like, data frame calculations.

693 01:04:58.710 ⇒ 01:04:59.469 Samuel Roberts: that results out.

694 01:04:59.470 ⇒ 01:05:03.650 Uttam Kumaran: Okay, I guess I wasn’t like, okay, that makes sense, yeah.

695 01:05:03.660 ⇒ 01:05:06.560 Samuel Roberts: That’s one way to do it. The other way is, like, we…

696 01:05:06.810 ⇒ 01:05:16.490 Samuel Roberts: like, take that business logic into something different, like, whether or not it’s a TypeScript or Python backend that’s hosting this or whatever, and…

697 01:05:16.660 ⇒ 01:05:22.339 Samuel Roberts: build it into that, or keep these as endpoints? Like, I… there’s just a few different ways to think about it.

698 01:05:23.120 ⇒ 01:05:26.239 Samuel Roberts: I don’t think any of them is… is really too…

699 01:05:26.440 ⇒ 01:05:37.669 Samuel Roberts: crazy to do. I haven’t dug into the Python code, so I don’t know exactly what the, you know, math going on there is, but I can’t imagine it’s… I’ve done some sub-forecasting like this before, and I feel like I…

700 01:05:37.670 ⇒ 01:05:40.580 Uttam Kumaran: So, that’s that. We could leave to the data folks.

701 01:05:40.910 ⇒ 01:05:43.359 Samuel Roberts: That’s why I’m not too worried about that. Yeah, exactly.

702 01:05:43.360 ⇒ 01:05:44.180 Uttam Kumaran: Okay, okay.

703 01:05:44.860 ⇒ 01:05:49.180 Samuel Roberts: I don’t… I don’t worry too, too much about that, but yeah, tying it together,

704 01:05:54.120 ⇒ 01:06:03.659 Samuel Roberts: Yeah. I mean, the other stuff, yeah, just becomes access to things for moving stuff, and where their data currently lives, if it’s AWS, on the, you know, how we…

705 01:06:04.150 ⇒ 01:06:06.140 Samuel Roberts: Make that transition work.

706 01:06:06.580 ⇒ 01:06:09.810 Samuel Roberts: What we’re gonna… what roadblocks we’re gonna hit from…

707 01:06:10.210 ⇒ 01:06:14.129 Samuel Roberts: their other vendor, whatever,

708 01:06:16.270 ⇒ 01:06:24.369 Samuel Roberts: But again, that’s not technically, like, too big. It doesn’t seem like too… like, the real crazy thing here could be that

709 01:06:24.480 ⇒ 01:06:30.029 Samuel Roberts: That last piece. That’s the big question mark, the, was it ForePlay API stuff?

710 01:06:30.330 ⇒ 01:06:37.780 Samuel Roberts: Which is probably the more interesting part, and the quest… the part I just know less about at this point, so that definitely needs to get a little more fleshed out in my mind, but…

711 01:06:38.660 ⇒ 01:06:39.240 Uttam Kumaran: Okay.

712 01:06:40.170 ⇒ 01:06:50.769 Uttam Kumaran: Okay, cool. Alright, I think I have enough to sort of put something together for these folks, so… Cool. Yeah, I mean, if this kind of goes through, I think it’ll be a mix of this crew working on us, so… that’s great.

713 01:06:50.910 ⇒ 01:06:55.470 Uttam Kumaran: I’ll do that later today. Maybe we can…

714 01:06:55.940 ⇒ 01:07:05.150 Uttam Kumaran: transition to just, I guess, like… and maybe, Sam, if you want to jump to the other meeting, you can, I’m sure, or if they…

715 01:07:05.150 ⇒ 01:07:06.480 Samuel Roberts: I was gonna ask you, yeah, okay.

716 01:07:08.000 ⇒ 01:07:09.759 Uttam Kumaran: Yeah, if you want to jump to that, that… because that’s just.

717 01:07:09.760 ⇒ 01:07:11.070 Samuel Roberts: So now I may come back, we’ll see.

718 01:07:11.070 ⇒ 01:07:11.990 Uttam Kumaran: Okay, okay.

719 01:07:12.630 ⇒ 01:07:14.299 Samuel Roberts: Okay, sounds good. I’ll be back.

720 01:07:15.530 ⇒ 01:07:17.660 Uttam Kumaran: And then… for…

721 01:07:18.330 ⇒ 01:07:28.849 Uttam Kumaran: the case study system, can I go ahead and pull it up on my side, and we want to walk through, or do we want… are we good to kind of ship it live?

722 01:07:32.570 ⇒ 01:07:33.349 Mustafa Raja: I guess.

723 01:07:33.350 ⇒ 01:07:35.039 Gabriel Lam: Do you want to bring up the latest one?

724 01:07:36.540 ⇒ 01:07:41.109 Mustafa Raja: Yeah, it should be, it should be up.

725 01:07:41.650 ⇒ 01:07:48.879 Mustafa Raja: Apart from that, we are working on a few, few updates, so yeah.

726 01:07:49.250 ⇒ 01:07:55.590 Mustafa Raja: It’s the same, review app for now. I’ll share in the Zoom chat.

727 01:07:57.560 ⇒ 01:07:58.990 Mustafa Raja: Oh yeah, Casey did.

728 01:08:11.270 ⇒ 01:08:11.820 Uttam Kumaran: Okay.

729 01:08:27.290 ⇒ 01:08:30.449 Gabriel Lam: So the goal here, really, is to clean up this flow.

730 01:08:30.640 ⇒ 01:08:36.550 Gabriel Lam: to make it… A little less confusing, between what’s a draft, what’s…

731 01:08:36.910 ⇒ 01:08:50.139 Gabriel Lam: awaiting interviews. We’re gonna add a couple buttons on the bottom, so instead of picking a status, you either save a draft to come back for later, or you just instantly send it out for interviews. So that’s gonna be something that’s gonna be… sorry. That’s gonna be something that will be…

732 01:08:50.300 ⇒ 01:08:51.120 Gabriel Lam: Fixed.

733 01:08:58.170 ⇒ 01:09:02.640 Uttam Kumaran: Okay. So if we were to walk through… so I think the other thing is we still don’t have…

734 01:09:03.630 ⇒ 01:09:04.930 Uttam Kumaran: Past clients?

735 01:09:04.930 ⇒ 01:09:05.970 Gabriel Lam: So…

736 01:09:08.859 ⇒ 01:09:09.599 Mustafa Raja: Yeah, I knew.

737 01:09:09.600 ⇒ 01:09:10.010 Uttam Kumaran: all.

738 01:09:10.290 ⇒ 01:09:11.319 Mustafa Raja: I’ll enable those.

739 01:09:14.229 ⇒ 01:09:16.569 Uttam Kumaran: Yeah, that’s gonna be a big thing.

740 01:09:17.229 ⇒ 01:09:18.099 Samuel Roberts: Yeah, that should be…

741 01:09:18.100 ⇒ 01:09:18.729 Gabriel Lam: I know.

742 01:09:18.920 ⇒ 01:09:19.930 Samuel Roberts: Pretty quick, don’t

743 01:09:22.979 ⇒ 01:09:27.020 Samuel Roberts: Okay, yeah, we’re actually gonna tweak that process a little bit there, too.

744 01:09:27.029 ⇒ 01:09:29.129 Uttam Kumaran: Yeah, Gabe just mentioned that, yeah.

745 01:09:29.130 ⇒ 01:09:34.160 Samuel Roberts: Cool, okay, good. Yeah, I got the Slack notification sending and stuff, so pretty close.

746 01:09:35.439 ⇒ 01:09:38.119 Uttam Kumaran: Okay, cool, and then any big, like, changes here?

747 01:09:38.950 ⇒ 01:09:40.810 Samuel Roberts: You can edit the description.

748 01:09:41.649 ⇒ 01:09:44.270 Samuel Roberts: As the interviewee, there’s a… Oh, on the website.

749 01:09:44.270 ⇒ 01:09:44.790 Uttam Kumaran: Okay.

750 01:09:45.240 ⇒ 01:09:46.160 Samuel Roberts: Yeah.

751 01:09:46.160 ⇒ 01:09:47.049 Uttam Kumaran: Okay, cool.

752 01:09:48.000 ⇒ 01:09:55.820 Samuel Roberts: So yeah, anyone can do that now, that’s… Actually, I don’t know if anyone can do that, if you come to this page and you’re not the interviewee, but I didn’t really think it was worth locking down yet, so…

753 01:09:56.140 ⇒ 01:09:59.689 Uttam Kumaran: Okay. I would probably say, like, please provide a description.

754 01:10:00.610 ⇒ 01:10:01.460 Uttam Kumaran: Sure.

755 01:10:01.460 ⇒ 01:10:02.210 Samuel Roberts: I’m gonna fill in.

756 01:10:02.210 ⇒ 01:10:03.010 Uttam Kumaran: Yeah, understood.

757 01:10:03.220 ⇒ 01:10:08.259 Uttam Kumaran: And then, okay, and then anything’s changed with, like, the typical voice process?

758 01:10:09.330 ⇒ 01:10:13.500 Gabriel Lam: Which is correct that Mustafa handled that, and the velocity is… It’s…

759 01:10:13.500 ⇒ 01:10:13.820 Uttam Kumaran: Cool.

760 01:10:13.820 ⇒ 01:10:14.330 Gabriel Lam: Yeah.

761 01:10:18.140 ⇒ 01:10:18.830 Uttam Kumaran: Nice.

762 01:10:22.570 ⇒ 01:10:27.279 Gabriel Lam: So it should be able to capture both if you want it more prescriptive, or just free flow.

763 01:10:27.640 ⇒ 01:10:29.499 Gabriel Lam: It’s a little more flexible now.

764 01:10:32.800 ⇒ 01:10:40.870 Uttam Kumaran: Yeah, I had to… I had to overcome… Just, like, rough…

765 01:10:41.050 ⇒ 01:10:43.079 Uttam Kumaran: Just, like, bad data from vendors.

766 01:10:54.050 ⇒ 01:10:55.729 Uttam Kumaran: Wait, what the heck is this?

767 01:10:55.730 ⇒ 01:10:57.389 Samuel Roberts: What is that, yeah?

768 01:10:57.390 ⇒ 01:11:01.510 Mustafa Raja: Yeah, that… that is the transcribing model.

769 01:11:01.910 ⇒ 01:11:02.880 Mustafa Raja: acting up.

770 01:11:04.110 ⇒ 01:11:05.970 Gabriel Lam: Okay, I haven’t seen this before.

771 01:11:05.970 ⇒ 01:11:06.640 Uttam Kumaran: Oh, okay.

772 01:11:07.190 ⇒ 01:11:10.780 Uttam Kumaran: Yeah, I noticed this in, in the previous…

773 01:11:11.220 ⇒ 01:11:13.820 Uttam Kumaran: stuff as well, in that, like, I think it’s…

774 01:11:14.180 ⇒ 01:11:18.350 Uttam Kumaran: something’s wrong with this, like, is this not coming from OpenAI?

775 01:11:19.590 ⇒ 01:11:20.480 Mustafa Raja: Yeah, this is coming.

776 01:11:20.480 ⇒ 01:11:20.840 Uttam Kumaran: driving.

777 01:11:20.840 ⇒ 01:11:22.060 Mustafa Raja: And on the fly.

778 01:11:22.460 ⇒ 01:11:22.980 Uttam Kumaran: Oh.

779 01:11:22.980 ⇒ 01:11:28.289 Mustafa Raja: No, no, no, so the API also transcribes whatever we say.

780 01:11:28.440 ⇒ 01:11:34.030 Mustafa Raja: And whatever you said, it’s… it transcribed it into… into this thing.

781 01:11:35.830 ⇒ 01:11:37.190 Uttam Kumaran: Oh, okay.

782 01:11:37.190 ⇒ 01:11:44.380 Mustafa Raja: But if you look at the model’s output, we see that it does understand what we actually said, and not this.

783 01:11:44.380 ⇒ 01:11:45.400 Uttam Kumaran: I see, I see.

784 01:11:45.890 ⇒ 01:11:46.980 Uttam Kumaran: Okay, okay.

785 01:11:47.610 ⇒ 01:11:50.899 Samuel Roberts: So… Okay, alright, I think, yeah, yeah, alright.

786 01:11:51.430 ⇒ 01:11:58.420 Uttam Kumaran: Yeah, I guess it’s, like, I’m not too concerned. I guess even overall, may not even be important to show

787 01:11:59.390 ⇒ 01:12:04.829 Uttam Kumaran: the user, like, what they’re saying. I don’t know. Not convinced that even matters. I think more.

788 01:12:04.830 ⇒ 01:12:05.289 Samuel Roberts: Yeah, we…

789 01:12:05.290 ⇒ 01:12:06.230 Uttam Kumaran: seeing this.

790 01:12:06.490 ⇒ 01:12:06.960 Samuel Roberts: Yeah.

791 01:12:10.170 ⇒ 01:12:15.060 Uttam Kumaran: Also, I mean, I don’t know, even, like, the transcript, maybe it’s not important to have that at all.

792 01:12:17.510 ⇒ 01:12:24.400 Mustafa Raja: Yeah, it’s just for, you know, if you want to read, something that I do.

793 01:12:26.310 ⇒ 01:12:32.789 Mustafa Raja: So I let the interviewer talk, and then I might look at what it said again.

794 01:12:33.430 ⇒ 01:12:34.200 Uttam Kumaran: Oh…

795 01:12:34.200 ⇒ 01:12:35.939 Samuel Roberts: You bring up an interesting… an interesting…

796 01:12:35.940 ⇒ 01:12:38.419 Gabriel Lam: I think it gives you the optionality.

797 01:12:38.840 ⇒ 01:12:39.710 Samuel Roberts: Yeah.

798 01:12:40.390 ⇒ 01:12:45.799 Samuel Roberts: But the transcription is what matters at the end of the day, because that’s what’s getting fed into the next step, right?

799 01:12:46.090 ⇒ 01:12:46.980 Mustafa Raja: Yes.

800 01:12:47.150 ⇒ 01:12:50.939 Mustafa Raja: Yes, whatever is being written here is going to the…

801 01:12:50.940 ⇒ 01:12:51.680 Samuel Roberts: Exactly.

802 01:12:51.680 ⇒ 01:12:53.290 Mustafa Raja: Next assistant.

803 01:12:56.240 ⇒ 01:13:06.649 Samuel Roberts: But there’s still a little bit of a disconnect between the real time and what the transcription is, so, like, theoretically, there could be something like this happening here, where the transcription gets the wrong thing that will get passed in, but the

804 01:13:07.090 ⇒ 01:13:10.159 Samuel Roberts: Interviewer isn’t aware of that.

805 01:13:11.140 ⇒ 01:13:12.890 Uttam Kumaran: Okay, so I just hit submit.

806 01:13:13.260 ⇒ 01:13:15.799 Samuel Roberts: And then, at this point.

807 01:13:15.800 ⇒ 01:13:20.809 Uttam Kumaran: I can go back and see the output.

808 01:13:22.740 ⇒ 01:13:24.270 Mustafa Raja: Yeah, right.

809 01:13:24.420 ⇒ 01:13:25.059 Samuel Roberts: It should be right.

810 01:13:25.190 ⇒ 01:13:29.409 Mustafa Raja: It takes a few seconds, but yeah.

811 01:13:29.410 ⇒ 01:13:29.960 Gabriel Lam: Fun.

812 01:13:30.640 ⇒ 01:13:31.500 Gabriel Lam: So I think if you click.

813 01:13:31.500 ⇒ 01:13:37.480 Uttam Kumaran: It was a very long interview, yeah, so… Yeah, yeah, yeah. So I think another thing is, like, we should have a

814 01:13:38.180 ⇒ 01:13:40.500 Uttam Kumaran: We should make it clear for people, like.

815 01:13:41.600 ⇒ 01:13:44.949 Uttam Kumaran: the other action, which is, like, opening the case study, I don’t know.

816 01:13:45.940 ⇒ 01:13:48.739 Samuel Roberts: Yeah, I was… I was just clicking that, should be… I noticed it’.

817 01:13:48.740 ⇒ 01:13:49.809 Uttam Kumaran: Oh, I guess this is…

818 01:13:49.810 ⇒ 01:13:59.600 Samuel Roberts: Yeah, it highlights each little cell, but it does… the only thing it does is open the thing. I can try to see if I can tweak that… what displays there, but…

819 01:13:59.930 ⇒ 01:14:02.420 Samuel Roberts: The action should be the same.

820 01:14:04.190 ⇒ 01:14:08.680 Uttam Kumaran: Okay, and so… I mean, it looks like…

821 01:14:08.680 ⇒ 01:14:09.250 Casie Aviles: Mmm.

822 01:14:09.250 ⇒ 01:14:15.060 Uttam Kumaran: Either the prompt is good, as how did it… how did it know… Rill and stuff.

823 01:14:18.130 ⇒ 01:14:23.350 Casie Aviles: Oh, I think that might be because of, like, examples I provided.

824 01:14:23.920 ⇒ 01:14:24.650 Samuel Roberts: -Oh.

825 01:14:24.920 ⇒ 01:14:25.960 Uttam Kumaran: Okay.

826 01:14:26.570 ⇒ 01:14:28.989 Uttam Kumaran: So I think one point is, like…

827 01:14:29.650 ⇒ 01:14:32.369 Uttam Kumaran: We just want to be careful that it doesn’t hallucinate.

828 01:14:32.540 ⇒ 01:14:34.580 Uttam Kumaran: Or it doesn’t fill in the gaps.

829 01:14:35.110 ⇒ 01:14:41.259 Uttam Kumaran: So, like, for example, it should say this… this… it’s not… we don’t have enough information.

830 01:14:41.830 ⇒ 01:14:48.629 Uttam Kumaran: Like, the one worry I have is if we go publish shit that we didn’t do. Like, this has happened before, where…

831 01:14:48.760 ⇒ 01:14:53.310 Uttam Kumaran: Where marketing will be like, oh yeah, we do real-time data. We don’t do any real-time stuff.

832 01:14:53.710 ⇒ 01:15:04.560 Uttam Kumaran: So, AI is all… anytime you work… talk about data with AI, it’s always gonna push towards, like, these types of things, but we… you could tell all I provided it was something, so I think…

833 01:15:04.680 ⇒ 01:15:11.220 Uttam Kumaran: I want to just be a lot more conscious that, like, the output only It’s only really based on…

834 01:15:12.040 ⇒ 01:15:16.870 Uttam Kumaran: what we mentioned as much as possible. I don’t mind it filling out the gaps.

835 01:15:17.510 ⇒ 01:15:21.360 Uttam Kumaran: like, in terms of the wording, but I do mind that it’s, like.

836 01:15:22.280 ⇒ 01:15:28.940 Uttam Kumaran: it just makes stuff up, you know? Because, think about it, what if I did everything except give it the results, and then it did this?

837 01:15:29.600 ⇒ 01:15:29.920 Samuel Roberts: Right.

838 01:15:30.790 ⇒ 01:15:33.370 Gabriel Lam: Yeah, like, it should say, like, we don’t have enough information.

839 01:15:33.370 ⇒ 01:15:36.010 Uttam Kumaran: It should say, yeah, it should say we don’t have enough info.

840 01:15:36.270 ⇒ 01:15:47.000 Samuel Roberts: Yeah, that might be a change to the prompt or another step that we can do that will take them both in and make sure that the previous thing is stripped of whatever is not actually

841 01:15:47.740 ⇒ 01:15:48.880 Samuel Roberts: mentioned.

842 01:15:51.510 ⇒ 01:15:52.920 Samuel Roberts: Let me add that now.

843 01:16:04.210 ⇒ 01:16:09.930 Gabriel Lam: You know, something else we were talking about earlier was, the credit usage. I know this has been something that

844 01:16:10.330 ⇒ 01:16:15.749 Gabriel Lam: was brought up yesterday, and we had a short discussion about it this morning. Sam, if…

845 01:16:15.910 ⇒ 01:16:19.219 Gabriel Lam: You’re able to elaborate on it more, but…

846 01:16:19.220 ⇒ 01:16:20.720 Samuel Roberts: Yeah.

847 01:16:20.720 ⇒ 01:16:22.360 Gabriel Lam: on observability.

848 01:16:22.610 ⇒ 01:16:24.559 Samuel Roberts: Yeah, so because this is not…

849 01:16:24.710 ⇒ 01:16:34.199 Samuel Roberts: a master agent, it doesn’t automatically get logged to LankView, so we were looking to specifically log the part that is the interview

850 01:16:34.430 ⇒ 01:16:38.099 Samuel Roberts: to LangFuse to get the pricing, just so you have a better sense of, like.

851 01:16:38.240 ⇒ 01:16:49.939 Samuel Roberts: what does a case study cost to run this? Because these are… this is a fairly expensive model for the real-time. There is an 04 mini real-time that I do want to test out and see how it compares.

852 01:16:51.470 ⇒ 01:16:58.810 Samuel Roberts: But I think I wanted to get that baseline price to really know if this is crazy or not. And then how much…

853 01:16:59.270 ⇒ 01:17:08.739 Samuel Roberts: worse it’ll be for the mini one, it might not be worse at all, I don’t really know. We might just be, you know, this might be overkill, but that’s… that was what we discussed this morning.

854 01:17:10.510 ⇒ 01:17:11.130 Uttam Kumaran: Okay.

855 01:17:15.340 ⇒ 01:17:22.750 Samuel Roberts: Yeah, I think there’s definitely, like, a cost with the copy generation and stuff, but that’s minimal compared to the back-and-forth real-time voice stuff.

856 01:17:30.890 ⇒ 01:17:35.600 Uttam Kumaran: Okay, cool, so I guess, like, do we feel like we’ll be able to get,

857 01:17:37.740 ⇒ 01:17:41.219 Uttam Kumaran: Get this out with those remaining, kind of, fixes today.

858 01:17:41.330 ⇒ 01:17:45.360 Uttam Kumaran: like… What do you guys think?

859 01:17:45.840 ⇒ 01:17:49.569 Uttam Kumaran: Or, I mean, I would… I would… I was ideally hoping that we could demo…

860 01:17:49.990 ⇒ 01:17:55.209 Uttam Kumaran: this, and Gabe, you can demo this for in the next… in, like, our company All Hands that’s coming up.

861 01:17:56.190 ⇒ 01:18:10.709 Uttam Kumaran: I was just gonna carve out a slide, and we could just share screen and demo it. But do we feel, like, comfortable, like, kind of roughing out the remaining edges and, like, getting this out so people can start using it? Even in its current state, I feel like

862 01:18:11.370 ⇒ 01:18:12.330 Uttam Kumaran: It’s great.

863 01:18:12.660 ⇒ 01:18:19.210 Uttam Kumaran: So, the kind of small comments I have are, I think, as you guys know, is just, like, I think some edge cases, so…

864 01:18:19.810 ⇒ 01:18:21.550 Uttam Kumaran: Yeah. Yeah.

865 01:18:22.390 ⇒ 01:18:22.759 Samuel Roberts: Yeah, I think.

866 01:18:22.760 ⇒ 01:18:23.990 Gabriel Lam: Yeah, I think when…

867 01:18:24.100 ⇒ 01:18:25.330 Samuel Roberts: Go ahead. No, go for it.

868 01:18:25.710 ⇒ 01:18:42.510 Samuel Roberts: I was just gonna say, I think, like, there will be more things to refine as we use it anyway, and keep going, so I think, like, it’s in a pretty good state. A few of these things will be good fixes now, and then we’ll just have to probably set up, like, a check-in with, you know, marketing, and make sure that the outputs are still, you know.

869 01:18:42.710 ⇒ 01:18:46.049 Samuel Roberts: Good for what we’re actually getting at the final result.

870 01:18:46.880 ⇒ 01:18:47.460 Uttam Kumaran: Okay.

871 01:18:48.530 ⇒ 01:18:52.000 Samuel Roberts: Because I just don’t want to keep this iterating, because I think, you know.

872 01:18:52.200 ⇒ 01:18:53.010 Uttam Kumaran: Yeah, yeah, yeah.

873 01:18:53.010 ⇒ 01:18:58.950 Samuel Roberts: Even if it’s based on where we’re at, but yeah, I think it’s definitely worth, you know, a few of these things, few of these fixes. I think also.

874 01:18:59.050 ⇒ 01:19:07.979 Samuel Roberts: The hallucinations will hopefully be a little less of an issue if it’s a full interview and not just something that’s got no… no context, just from a little test, so I’m hoping that that’s, like.

875 01:19:08.600 ⇒ 01:19:15.570 Samuel Roberts: we can tweak the prompts, but I don’t think it’ll be as big a deal, but… Okay. But yeah, we’ll keep an eye on that moving forward.

876 01:19:15.570 ⇒ 01:19:17.540 Uttam Kumaran: But are those, are those prompts in LinkViews?

877 01:19:18.390 ⇒ 01:19:23.190 Casie Aviles: No, there’s still, like… In the repo itself…

878 01:19:23.970 ⇒ 01:19:26.509 Casie Aviles: So, that’s also something I was…

879 01:19:27.240 ⇒ 01:19:29.219 Casie Aviles: Thinking about whether we should

880 01:19:29.480 ⇒ 01:19:34.289 Casie Aviles: move it to LangFuse, but I guess the only good thing about having it within the repo

881 01:19:34.470 ⇒ 01:19:37.910 Casie Aviles: is that a cursor can directly edit those, and I can just.

882 01:19:37.910 ⇒ 01:19:43.070 Samuel Roberts: Yeah, I… my thought is for, like, development process, it should be…

883 01:19:43.730 ⇒ 01:19:48.630 Samuel Roberts: In the repo for now, and then, like, once it’s in a good state, we can move it to length views.

884 01:19:48.790 ⇒ 01:19:55.289 Samuel Roberts: To track versions after that, but it definitely would make it harder to,

885 01:19:55.750 ⇒ 01:19:59.830 Samuel Roberts: harder to use Cursor, and harder to make quick changes, and… keeping the…

886 01:20:00.110 ⇒ 01:20:07.920 Samuel Roberts: sync between LangView’s versions and any other code changes that happen, I think, during development is not worth it right now.

887 01:20:08.400 ⇒ 01:20:12.820 Samuel Roberts: But yeah, probably, I mean, I would say, like, if we can get this out and used, then I would say maybe, like.

888 01:20:13.300 ⇒ 01:20:17.480 Samuel Roberts: in the… like, next week, we move it to Langviews once it’s kind of…

889 01:20:17.600 ⇒ 01:20:20.760 Samuel Roberts: Proven that it does a few case studies well.

890 01:20:22.190 ⇒ 01:20:23.470 Samuel Roberts: Yeah, I think it’s worth…

891 01:20:23.470 ⇒ 01:20:24.310 Uttam Kumaran: So…

892 01:20:24.560 ⇒ 01:20:40.149 Uttam Kumaran: It’s worth seeing how other teams are doing it. I mean, LinkViews, I think they have a… they have a Discord or something. It’s worth asking how other teams are managing. I mean, I’m comfortable in… for development, but I… guys, I really think we want to democratize people

893 01:20:40.350 ⇒ 01:20:42.479 Uttam Kumaran: Editing and improving prompts.

894 01:20:42.740 ⇒ 01:20:43.790 Samuel Roberts: 100%.

895 01:20:43.790 ⇒ 01:21:03.269 Uttam Kumaran: it’s… it’s like, it’s just gonna be the one piece that you guys don’t… I’m telling you, you don’t want to, like, have to make these, like, small prompt tweaks. Like, I would rather… I would… I could totally, like, if you’re like, oh yeah, go into length views and do whatever you want, I’ll go do it. You guys… 100%, but doing it in code, I think… this is just one piece that I think we want to move there, so as soon as you guys are comfortable

896 01:21:03.310 ⇒ 01:21:16.109 Uttam Kumaran: that would be my suggestion, or longer term, think through something that’s like, okay, in development, we can have it in code, but as part of the PR process, it should get migrated to LaneFuse. That’s exactly… yeah, that’s…

897 01:21:16.350 ⇒ 01:21:21.380 Uttam Kumaran: Everybody in the company, we have, luckily, are very good users of ChatGBT, so everybody’s…

898 01:21:21.530 ⇒ 01:21:24.250 Uttam Kumaran: At least gonna be able to supply feedback there.

899 01:21:25.960 ⇒ 01:21:26.670 Uttam Kumaran: You know.

900 01:21:27.450 ⇒ 01:21:37.370 Samuel Roberts: Definitely. Yeah, no, that’s… that’s… you pretty much summed up exactly what I was thinking when we were talking about it before, was during development, and then, yeah, once it’s PR’d into main, maybe that’s when we…

901 01:21:38.130 ⇒ 01:21:42.080 Samuel Roberts: get it into Langviews so that if people need to make changes, they can, but…

902 01:21:42.320 ⇒ 01:21:46.770 Samuel Roberts: For quick iterations, I think, which is most of the dev flee.

903 01:21:47.510 ⇒ 01:21:50.480 Samuel Roberts: I wouldn’t worry about it, but yeah, we can do that as part of the PR today.

904 01:21:51.480 ⇒ 01:21:52.060 Uttam Kumaran: Okay.

905 01:21:52.740 ⇒ 01:21:56.960 Gabriel Lam: I think in terms of an MV… or a demoable MVP,

906 01:21:57.100 ⇒ 01:22:01.199 Gabriel Lam: I… I do feel like the features that we get out today

907 01:22:01.310 ⇒ 01:22:05.350 Gabriel Lam: Are relatively minor compared to, like, getting people

908 01:22:05.690 ⇒ 01:22:17.990 Gabriel Lam: getting the idea that we can conduct these interviews with an AI and have copy generated, I think that’s the bigger, like, mind shift for people to be like, oh, I don’t have to set a time with Hannah anymore. And I think if we’re able.

909 01:22:17.990 ⇒ 01:22:18.520 Uttam Kumaran: Yes.

910 01:22:18.520 ⇒ 01:22:21.430 Gabriel Lam: To demonstrate that, that’s the most important thing.

911 01:22:22.430 ⇒ 01:22:23.340 Samuel Roberts: Agreed.

912 01:22:25.720 ⇒ 01:22:26.320 Uttam Kumaran: Okay.

913 01:22:29.320 ⇒ 01:22:34.199 Samuel Roberts: What time is that, all hands, that’s it.

914 01:22:34.200 ⇒ 01:22:34.850 Uttam Kumaran: It’s like an hour.

915 01:22:34.850 ⇒ 01:22:36.190 Samuel Roberts: Okay, yeah.

916 01:22:37.950 ⇒ 01:22:46.499 Samuel Roberts: I’m hoping to get the Slack message PR’d in, or into the PR, so that you can maybe even show that people get notifications and stuff.

917 01:22:46.680 ⇒ 01:22:52.949 Samuel Roberts: And I could probably have that in the next hour, so… we can add that to part of the demo, so people are used to that.

918 01:22:54.240 ⇒ 01:23:01.709 Samuel Roberts: And then… yeah, the other tweaks, I think, are probably less critical to show there, as long as we’re showing a good demo.

919 01:23:02.410 ⇒ 01:23:05.399 Samuel Roberts: But… Yeah, we gotta get them by the end of the day.

920 01:23:09.780 ⇒ 01:23:10.530 Uttam Kumaran: Great.

921 01:23:11.300 ⇒ 01:23:12.130 Uttam Kumaran: Alright.

922 01:23:12.890 ⇒ 01:23:14.779 Uttam Kumaran: Thanks, guys. I’ll talk to you then.

923 01:23:15.320 ⇒ 01:23:15.800 Samuel Roberts: Alrighty.

924 01:23:16.550 ⇒ 01:23:17.190 Uttam Kumaran: Okay.

Brainforge Knowledge

Explorer

2025-11-14_ai-default-abc_standup_3d946a89

Graph View