Meeting Title: Internal AI Team | Standup Date: 2025-04-14 Meeting participants: Miguel De Veyra, Casie Aviles, Amber Lin


WEBVTT

1 00:00:42.890 00:00:43.739 Miguel de Veyra: Hey, everyone.

2 00:00:44.890 00:00:49.776 Amber Lin: Oh, man, there you go.

3 00:00:57.270 00:00:57.850 Amber Lin: Okay.

4 00:00:57.850 00:00:58.360 Miguel de Veyra: Bye.

5 00:00:59.190 00:01:08.710 Amber Lin: Let’s let’s blast through this meeting. Oh, let’s see I’ll give you guys a quick update on ABC right now

6 00:01:08.870 00:01:25.599 Amber Lin: doing the rollouts. Not sure if Tim gave everyone access, checking with him, and then Janice will check with them as well. And then there’s errors here and there, but I think they can fix it on their end. So I’m just gonna send Janice everything. So I think we’re gonna be

7 00:01:25.870 00:01:29.860 Amber Lin: good for at least another

8 00:01:29.980 00:01:34.730 Amber Lin: one or 2 days. Do you guys remember if there’s anything we said we wanted to do.

9 00:01:38.860 00:01:42.609 Miguel de Veyra: I think there’s a couple, but I’m not sure if it’s this cycle.

10 00:01:47.410 00:01:52.060 Casie Aviles: Yeah, I think at the top of my head. There’s like.

11 00:01:52.960 00:01:55.949 Casie Aviles: I guess, a trainer bot! And then.

12 00:01:55.950 00:01:56.580 Amber Lin: Oh!

13 00:01:56.900 00:02:02.519 Casie Aviles: I’m not sure if you want to do the like this additional button for them to generate the oh, by the ways.

14 00:02:04.610 00:02:08.829 Amber Lin: I’m gonna for those 2 tests. How long do you think they will take.

15 00:02:15.060 00:02:23.430 Miguel de Veyra: I would say, the oh, by the way, generator with one. But with cause, that’s gonna include Google integration

16 00:02:23.800 00:02:28.689 Miguel de Veyra: and stuff like that, that’s probably gonna be 5 points case, you know.

17 00:02:31.517 00:02:41.260 Miguel de Veyra: And then as for the, as for the document generator, we have to break that down, I I think, because it’s too big of a ticket, and then

18 00:02:41.620 00:02:44.899 Miguel de Veyra: do we already know how they want it?

19 00:02:47.780 00:02:56.939 Amber Lin: I mean, we could just ha have a version for them to test with. Right? So we have all their past training documents.

20 00:02:57.670 00:02:58.920 Amber Lin: And

21 00:03:01.146 00:03:10.660 Amber Lin: essentially, we’re prompt giving a prompt to Gpt right? I guess we can start with maybe this week, just to create a custom prompt.

22 00:03:11.150 00:03:18.000 Amber Lin: What do you think like a Sys or a system prompt, and then

23 00:03:18.540 00:03:23.300 Amber Lin: next week, or if we don’t, if we have capacity, we can

24 00:03:23.470 00:03:30.619 Amber Lin: see how we can integrate that because we decided on just to use chatgpt and plug it into

25 00:03:31.434 00:03:33.849 Amber Lin: their Google workspace. Right?

26 00:03:35.480 00:03:36.210 Miguel de Veyra: Okay.

27 00:03:37.290 00:03:37.760 Amber Lin: Or is it.

28 00:03:37.760 00:03:41.390 Miguel de Veyra: Does, it still has to be deployed into their Google Workspace.

29 00:03:43.500 00:03:45.560 Amber Lin: Is that what we decided on.

30 00:03:46.697 00:03:49.280 Miguel de Veyra: I think we we were supposed to ask them.

31 00:03:50.540 00:03:55.529 Amber Lin: Oh, if it needs to be in their Google Workspace. Is that the question.

32 00:03:56.220 00:03:57.070 Miguel de Veyra: Yeah, yeah.

33 00:03:57.850 00:04:03.600 Amber Lin: Okay, I’ll make a note to ask Yvette or Janice.

34 00:04:03.810 00:04:06.740 Miguel de Veyra: But I’m but that for that one. I think it.

35 00:04:07.450 00:04:15.909 Miguel de Veyra: It’s probably if it’s just that, probably 5 points. Yeah, it’ll probably take a day. I’m not sure the Google Integration part because.

36 00:04:16.500 00:04:21.130 Miguel de Veyra: Casey, have we gotten over the hurdle? The hurdle of deploying 2 bots or not yet?

37 00:04:24.130 00:04:31.380 Casie Aviles: I mean, I think there’s a way. But you know the the blocker is just did with the 1st implementation of the bot itself.

38 00:04:33.140 00:04:33.800 Miguel de Veyra: Yeah.

39 00:04:33.800 00:04:39.740 Casie Aviles: Like the update, like the update, but itself, like the logic of, you know, editing the document.

40 00:04:40.980 00:04:44.709 Miguel de Veyra: Okay. But we can deploy multiple bots into Google Chat already. Right?

41 00:04:45.350 00:04:48.650 Casie Aviles: I believe we can, as long as we have, like 2 separate projects.

42 00:04:49.820 00:04:50.700 Miguel de Veyra: Okay. Okay.

43 00:04:52.160 00:05:03.709 Amber Lin: I see, are we? Cause? I thought last time when we talked about it, we’re no longer gonna make the edits to the document directly. We’re probably just gonna suggest something using Gpt. And then.

44 00:05:03.710 00:05:06.289 Miguel de Veyra: Yep, yep, that’s what we’re gonna do now.

45 00:05:06.490 00:05:13.220 Amber Lin: Oh, okay. So that would say, how long would that.

46 00:05:13.220 00:05:16.549 Miguel de Veyra: Create a ticket, for that’s probably going to be another 5 points.

47 00:05:16.550 00:05:20.070 Amber Lin: Okay, I think we do have the ticket. Let me go find it.

48 00:05:20.340 00:05:20.890 Miguel de Veyra: Okay.

49 00:05:23.160 00:05:31.249 Amber Lin: Trainer. But I’ll say estimate is 5 points, and then the

50 00:05:37.400 00:05:38.530 Amber Lin: where’s the other?

51 00:05:38.930 00:05:45.100 Amber Lin: Oh, the button! Do you guys want to do the button first, st or the

52 00:05:45.930 00:05:48.679 Amber Lin: I guess I could. I could ask Janice. Huh!

53 00:05:49.050 00:05:56.380 Miguel de Veyra: I think we work with the bottom first, st because that’s like tied to what all what they already have. And then the edit is like a new one.

54 00:05:57.880 00:06:03.799 Amber Lin: Sure. I mean, if we have time we can do the button, and then if we have time, we can do

55 00:06:04.363 00:06:12.349 Amber Lin: not the integration part, but maybe the prompt part, so I can demo to them. Hey, this is how cool it is.

56 00:06:12.550 00:06:15.560 Amber Lin: I hope that prompting wouldn’t take that much time.

57 00:06:16.130 00:06:16.680 Miguel de Veyra: Yeah.

58 00:06:16.960 00:06:19.379 Amber Lin: Just the just the prompting part.

59 00:06:19.820 00:06:21.879 Amber Lin: Yeah, I’ll make a subtask.

60 00:06:22.520 00:06:23.420 Amber Lin: Right?

61 00:06:34.870 00:06:38.149 Amber Lin: Sounds good. What about the AI side?

62 00:06:40.940 00:06:42.860 Casie Aviles: For you internal site. Yep.

63 00:06:43.861 00:06:49.210 Casie Aviles: yeah, we’re. I guess. We were planning the internal review with Patrick later.

64 00:06:50.720 00:06:53.160 Miguel de Veyra: Not invited to that one. What time is it.

65 00:06:54.680 00:06:59.009 Casie Aviles: It’s 1130, our time I I created like a an event.

66 00:07:00.110 00:07:05.990 Miguel de Veyra: Oh, it’s in internal AI review. Okay? Okay, I thought this was meeting.

67 00:07:07.000 00:07:11.990 Casie Aviles: Yeah, it’s going to be a separate one from the officers, I believe. So.

68 00:07:12.300 00:07:12.960 Miguel de Veyra: Officers.

69 00:07:12.960 00:07:16.309 Casie Aviles: And we don’t have one yet. We don’t have a.

70 00:07:16.605 00:07:16.900 Miguel de Veyra: Okay.

71 00:07:16.900 00:07:17.790 Casie Aviles: That one?

72 00:07:20.197 00:07:27.359 Casie Aviles: What else? So yeah, last week it’s more on just the zoom side. So I started with.

73 00:07:27.910 00:07:37.439 Casie Aviles: So I guess I have the code for that, and I just want I’m going to show that to Patrick later for the review. Yet in Github. So

74 00:07:38.521 00:07:43.160 Casie Aviles: there are some caveats, though. So that’s why I want to have it reviewed

75 00:07:43.340 00:07:47.530 Casie Aviles: like, yeah, it doesn’t fully satisfy the acceptance criteria.

76 00:07:50.920 00:07:53.210 Casie Aviles: but at the minimum. What it does, is it?

77 00:07:53.390 00:07:56.329 Casie Aviles: It’s able to pull like the metadata from Zoom.

78 00:07:56.490 00:08:00.470 Casie Aviles: So at least we we have that covered, and then we have our

79 00:08:01.615 00:08:04.440 Casie Aviles: workflows responsible for actually uploading

80 00:08:05.010 00:08:08.890 Casie Aviles: the zoom files to S 3. And yeah.

81 00:08:09.446 00:08:14.199 Casie Aviles: right now, I’m in progress with slack. So I’m also using dlt with that. And

82 00:08:14.310 00:08:20.270 Casie Aviles: yeah, I guess that’s my progress on the data stuff. I mean, yeah, I mean the

83 00:08:20.710 00:08:22.800 Casie Aviles: getting the context for the AI.

84 00:08:27.710 00:08:34.699 Miguel de Veyra: Yeah. And then sorry for the sa, for the sales. I think it’s just a matter of deploying it. I’m I’m have you asked

85 00:08:34.909 00:08:37.830 Miguel de Veyra: for feedback from.

86 00:08:38.059 00:08:39.500 Amber Lin: I asked.

87 00:08:40.409 00:08:49.330 Amber Lin: He hasn’t given me any more context or the message templates, so I guess we’re a little stuck on his on his end.

88 00:08:49.490 00:08:53.939 Miguel de Veyra: Yeah, I think Utam mentioned, we just deploy it into sales, for now, anyways.

89 00:08:55.010 00:08:56.090 Amber Lin: Oh! Did he.

90 00:08:57.160 00:08:57.980 Casie Aviles: Channel.

91 00:08:58.560 00:08:59.759 Amber Lin: Oh, I see.

92 00:08:59.970 00:09:16.270 Amber Lin: Sure, yeah. And I’ll nag him on the context because I do want to have a message suggestion for him. Do you think that would take a lot of time to suggest different types, and, like a few different messages based on the past messages.

93 00:09:16.270 00:09:22.400 Miguel de Veyra: I mean, it kinda is already giving that like. But again, I need context.

94 00:09:22.810 00:09:23.330 Amber Lin: Yeah.

95 00:09:23.330 00:09:26.289 Miguel de Veyra: Probably improve it a bit. It’ll probably take like.

96 00:09:26.700 00:09:30.649 Miguel de Veyra: I don’t know 2 points. So it’s short amount of time, probably. Max, 3 h.

97 00:09:30.930 00:09:34.029 Miguel de Veyra: But yeah, we have nothing to work with for now.

98 00:09:34.290 00:09:38.919 Amber Lin: Okay, yeah. Sounds good. I think that’s there. Would

99 00:09:39.530 00:09:48.539 Amber Lin: do. We have enough to create other client agents with our current system as opposed with Dlt.

100 00:09:50.810 00:09:51.200 Miguel de Veyra: Oh!

101 00:09:51.200 00:09:56.690 Casie Aviles: We we? Yeah, we are like, it depends on the client, because right now we are tracking, I think.

102 00:09:57.710 00:10:00.260 Casie Aviles: We have urban stems. Eden.

103 00:10:00.770 00:10:05.630 Casie Aviles: ABC, Javi, yeah, I think those

104 00:10:05.790 00:10:09.779 Casie Aviles: we have those. We have data from the client channels. I believe.

105 00:10:14.980 00:10:20.169 Miguel de Veyra: We could expand it. But, Casey, do you think it’s gonna take a it will take a day now.

106 00:10:22.670 00:10:30.500 Casie Aviles: Maybe faster. Since I’m just gonna copy the work existing workflow. And just, you know, change some things.

107 00:10:30.800 00:10:31.560 Casie Aviles: Yeah.

108 00:10:33.210 00:10:34.150 Miguel de Veyra: Okay. Yeah.

109 00:10:34.150 00:10:36.259 Amber Lin: Okay, yeah, I mean.

110 00:10:37.540 00:10:46.339 Amber Lin: I guess what our our goal is to. Okay, what are we gonna show on Friday? So I guess on Friday we’ll have pretty good progress on the context.

111 00:10:48.450 00:10:56.700 Amber Lin: the Javi bot! I’ll make sure that people learn how to use them, and then small tweaks here and there

112 00:10:58.050 00:11:00.739 Amber Lin: and then. What else do we want to show.

113 00:11:02.486 00:11:11.910 Miguel de Veyra: Think for in our current cycle sales office hours, probably that one we want to schedule that upgrade contacts.

114 00:11:12.840 00:11:17.889 Amber Lin: Yeah, I’m I mean more of our work. Demos.

115 00:11:18.504 00:11:25.139 Amber Lin: Do you think finishing up the context is a big enough milestone, because we are also

116 00:11:25.570 00:11:27.249 Amber Lin: last week, as well.

117 00:11:29.380 00:11:32.739 Miguel de Veyra: I mean, because we only have until Wednesday for this week. Right.

118 00:11:32.740 00:11:48.939 Amber Lin: Oh, okay, I keep forgetting. That’s you’re correct. That’s good. Okay, as long as that’s that’s gonna I’ll find other small things to pepper in there to make the presentation look bigger. But I think for Wednesday getting the context done is pretty much a big thing. Right?

119 00:11:49.790 00:11:54.140 Amber Lin: Okay? Yeah. And then we have like tiny ABC things. So I think this is good.

120 00:11:54.890 00:12:01.260 Miguel de Veyra: Okay, yeah. And then the other thing, I guess, is, Josh has already submitted his.

121 00:12:01.954 00:12:05.610 Amber Lin: Yeah, do you? Would you like me to send it to you?

122 00:12:06.301 00:12:10.218 Amber Lin: Sure, I mean, forward me the his email, or whatever

123 00:12:10.960 00:12:18.500 Amber Lin: I I suppose most of the technical part. You have to look at it right, but I’ll also watch it, too. I’ll see how he talks and how he communicates.

124 00:12:19.990 00:12:24.869 Miguel de Veyra: Actually, he did submit like a loom video cause, I told him, it’s either a loom video or documentation type of.

125 00:12:24.870 00:12:25.430 Amber Lin: Yeah.

126 00:12:26.230 00:12:29.570 Miguel de Veyra: Wait! Let me send this to you.

127 00:12:30.090 00:12:31.470 Amber Lin: Or email me.

128 00:12:33.080 00:12:45.780 Miguel de Veyra: I think it’s pretty good. Of course I also had AI help me out. It’s like 7 a 7.5 to to 8.5. It depends because we had a talk over the weekend. Basically, I told him, you know.

129 00:12:45.880 00:12:56.299 Miguel de Veyra: cause he couldn’t get the notion stuff, cause who does the notion database stuff. So I told him. Just base it off the air table, or what he knows, because it’s the same idea. Anyways.

130 00:12:56.920 00:12:58.589 Miguel de Veyra: as long as you get the course.

131 00:12:58.900 00:12:59.510 Amber Lin: Oh!

132 00:13:04.238 00:13:10.279 Miguel de Veyra: But yeah, I think that’s pretty much it. Is there anything else we want to work on this week?

133 00:13:14.710 00:13:23.540 Amber Lin: So, Miguel, would you be busy with any anything? Cause maybe Casey needs some help here and there.

134 00:13:23.960 00:13:28.069 Miguel de Veyra: I can probably to work on the update bot.

135 00:13:30.880 00:13:31.580 Amber Lin: Sure.

136 00:13:32.050 00:13:34.689 Miguel de Veyra: Yeah. And honestly, I could probably work on both

137 00:13:35.060 00:13:40.670 Miguel de Veyra: the the update bought and the oh, by the way, suggestion, I think that’s not really that hard.

138 00:13:41.410 00:13:42.690 Amber Lin: Sure.

139 00:13:42.710 00:13:44.240 Miguel de Veyra: Yeah, I’ll I’ll talk with Casey.

140 00:13:44.240 00:13:52.530 Amber Lin: And then I will. I will flesh out that if you guys can help flesh out the technical requirements on the tickets, that’ll be great, because I don’t know what goes in there.

141 00:13:52.940 00:13:53.730 Miguel de Veyra: Okay. Okay.

142 00:13:53.730 00:13:55.750 Amber Lin: Yeah, sounds good.

143 00:13:58.410 00:14:04.659 Amber Lin: Okay. Thank you. Guys for the meeting. I’ll talk to you guys. I’ll see you guys later at the other meeting. Then.

144 00:14:05.010 00:14:08.740 Miguel de Veyra: I’ll probably because I’ll be at the data meeting and then I’ll hop in.

145 00:14:09.860 00:14:15.570 Miguel de Veyra: Yeah, this is for 1131, right? Yeah, I have a 1 on one with them at around 11, so we’ll see how that goes.

146 00:14:18.770 00:14:20.479 Miguel de Veyra: Okay, thanks. Everyone. Bye, bye.

147 00:14:20.480 00:14:26.960 Amber Lin: I think data platform meeting at 11. You might want to check with him if he can go.

148 00:14:27.750 00:14:28.429 Amber Lin: Oh, okay.

149 00:14:28.430 00:14:29.020 Miguel de Veyra: I mean.

150 00:14:29.020 00:14:30.920 Amber Lin: 1130. Okay, that’s good.

151 00:14:32.060 00:14:32.929 Miguel de Veyra: Okay. Okay.

152 00:14:33.140 00:14:34.330 Amber Lin: Bye, guys.

153 00:14:34.780 00:14:36.129 Miguel de Veyra: Bye, everyone, thank you.

154 00:14:36.410 00:14:37.020 Casie Aviles: Thank you.