Meeting Title: ABC Project Weekly Sync Date: 2026-03-04 Meeting participants: Casie Aviles, Amber Lin, Mustafa Raja, Pranav Narahari, Samuel Roberts
WEBVTT
1 00:00:07.830 ⇒ 00:00:09.170 Amber Lin: Hello!
2 00:00:09.590 ⇒ 00:00:10.700 Amber Lin: Hi, team!
3 00:00:16.710 ⇒ 00:00:20.230 Amber Lin: Is Sam still out of office, or is he coming in a bit?
4 00:00:20.660 ⇒ 00:00:22.880 Mustafa Raja: I think he’s just a few minutes late.
5 00:00:23.280 ⇒ 00:00:24.840 Amber Lin: Okay, that’s all good.
6 00:00:26.670 ⇒ 00:00:32.659 Amber Lin: So, just so you guys know, we are working on transitioning CSO to Pranav.
7 00:00:32.710 ⇒ 00:00:41.669 Amber Lin: So, I think today we’re gonna lead this meeting together, and then soon, probably next week or so.
8 00:00:41.710 ⇒ 00:00:51.680 Amber Lin: Bernard will start taking over the CSO rules, which means he will also help translate the requirements from the clients and help with planning.
9 00:00:51.680 ⇒ 00:01:03.220 Amber Lin: I think Prinav’s skills on the AI side will make it a lot easier and a lot clearer to communicate, especially on architecting, on the technical side, so I look forward to that.
10 00:01:06.090 ⇒ 00:01:13.529 Pranav Narahari: Yeah, so I think today, Amber, you’re kind of gonna just start off with, like, going over Instagant.
11 00:01:13.690 ⇒ 00:01:15.290 Pranav Narahari: Is that right?
12 00:01:15.810 ⇒ 00:01:26.850 Amber Lin: Yeah, so I will just… I’ll run through the different workflows, just to reestablish what… what we’re working on, so that next time, you can just directly take over.
13 00:01:27.150 ⇒ 00:01:35.920 Amber Lin: So this is just another rundown of what we’re doing, and what workflows we’re currently working on, any blockers we have.
14 00:01:37.390 ⇒ 00:01:38.560 Pranav Narahari: Perfect, yep.
15 00:01:39.010 ⇒ 00:01:43.520 Pranav Narahari: Yeah, and I actually did sync with Casey yesterday on a couple of the workflows, so…
16 00:01:43.600 ⇒ 00:01:45.020 Amber Lin: Awesome. That was great. Awesome.
17 00:01:45.120 ⇒ 00:02:04.359 Amber Lin: Sounds good. So, let me… let me share. I think the biggest thing in mind I want to ask first is on the… on the migration side. I saw that the AI models are still blocked by Tim. Anything else that’s stopping us from rolling out this week?
18 00:02:07.310 ⇒ 00:02:15.119 Casie Aviles: Yeah, for this week, we’re… I’m still in the process of just doing the… this one, the E2E validation.
19 00:02:17.740 ⇒ 00:02:19.440 Amber Lin: Sorry, which one?
20 00:02:20.230 ⇒ 00:02:23.279 Casie Aviles: With… let’s see, it’s number 42.
21 00:02:23.660 ⇒ 00:02:24.400 Amber Lin: Okay.
22 00:02:25.680 ⇒ 00:02:28.140 Casie Aviles: This one is actually in progress.
23 00:02:29.150 ⇒ 00:02:30.340 Amber Lin: And…
24 00:02:31.100 ⇒ 00:02:37.150 Casie Aviles: I think… so we’re testing it with the dev version right now, so I managed to fix some… a couple of errors.
25 00:02:37.300 ⇒ 00:02:39.450 Amber Lin: But what’s…
26 00:02:39.950 ⇒ 00:02:43.820 Casie Aviles: I guess what’s kind of blocking right now is…
27 00:02:44.600 ⇒ 00:02:50.350 Casie Aviles: The… we found earlier that the response times may not be the most ideal, so maybe it’s like a…
28 00:02:50.350 ⇒ 00:02:51.060 Amber Lin: Mmm.
29 00:02:51.060 ⇒ 00:02:52.599 Casie Aviles: So we’re investigating that.
30 00:02:54.650 ⇒ 00:02:59.009 Casie Aviles: Yeah, and hopefully we were going to try to resolve that also, kind of.
31 00:02:59.140 ⇒ 00:03:02.270 Casie Aviles: Sorry, in this working session, but…
32 00:03:02.500 ⇒ 00:03:06.550 Casie Aviles: Yeah, just to give you guys, like, an idea of what… how it is right now.
33 00:03:07.410 ⇒ 00:03:16.230 Amber Lin: Gotcha, okay. So do you guys think we’re still able to have them tested by the end of this week, or is that not a realistic timeline anymore?
34 00:03:18.620 ⇒ 00:03:23.179 Casie Aviles: I think we should still be able to, like.
35 00:03:23.340 ⇒ 00:03:27.209 Casie Aviles: do the testing, at least, like, complete it within this week.
36 00:03:27.860 ⇒ 00:03:37.480 Casie Aviles: But… I’m not sure if we can, like, have the… the best state of Andy.
37 00:03:37.630 ⇒ 00:03:38.460 Casie Aviles: released within.
38 00:03:38.460 ⇒ 00:03:42.799 Amber Lin: Gotcha. Okay, I see. So…
39 00:03:43.080 ⇒ 00:03:57.949 Amber Lin: I would suggest holding it off, or just having one of them tested, because if we roll out to any… anyone else, if they’re confused, they’re gonna say, oh, this is not good, they’re gonna think this is the end state, so…
40 00:03:58.150 ⇒ 00:03:59.000 Casie Aviles: Yeah, yeah.
41 00:03:59.000 ⇒ 00:04:05.159 Amber Lin: If you think this is not great, don’t cause bad reviews, I’d rather just… Hold it off.
42 00:04:06.160 ⇒ 00:04:15.860 Casie Aviles: Okay, okay, yeah, that makes sense. Okay. So we’re just gonna focus on the testing still, and make, you know, ironing out the flaws, okay?
43 00:04:15.860 ⇒ 00:04:21.270 Amber Lin: Cool. And then AI models were blocked by Tim, right?
44 00:04:22.160 ⇒ 00:04:25.499 Mustafa Raja: Yeah, Tim just, replied to me…
45 00:04:25.630 ⇒ 00:04:30.180 Amber Lin: Okay. In the thread, he says that he’ll do it today.
46 00:04:30.180 ⇒ 00:04:40.770 Mustafa Raja: So, yeah, it just, it depends on Tim if he gets it out. As soon as he gets it out, we should be able to, you know, update the models.
47 00:04:40.990 ⇒ 00:04:42.289 Mustafa Raja: To Gemini. Okay.
48 00:04:43.160 ⇒ 00:04:53.960 Amber Lin: Sounds good. What about, alright, tell me about these…
49 00:04:54.860 ⇒ 00:04:59.549 Amber Lin: task. Oh, sorry, one… one more thing. Is this done? The feedback?
50 00:05:00.040 ⇒ 00:05:01.780 Amber Lin: Thumbs up.
51 00:05:01.780 ⇒ 00:05:04.540 Mustafa Raja: I’ll be taking it this week.
52 00:05:05.280 ⇒ 00:05:07.029 Amber Lin: Okay, so…
53 00:05:07.030 ⇒ 00:05:08.440 Mustafa Raja: Let’s move it to this weekend.
54 00:05:09.460 ⇒ 00:05:19.239 Amber Lin: Okay, got it. I think that’s important before… we need that to do any testing with clients, because that’s the only way we’ll know their feedback.
55 00:05:20.070 ⇒ 00:05:20.940 Mustafa Raja: Yep.
56 00:05:23.330 ⇒ 00:05:24.300 Amber Lin: Alright.
57 00:05:24.570 ⇒ 00:05:38.089 Amber Lin: Alright, that’s… we’ll discuss the things you want to just remember what we want to discuss. I just want to run through all the workflows first. How’s it on the zip code side? I know we’re doing final QA.
58 00:05:39.580 ⇒ 00:05:42.230 Casie Aviles: Yeah, so for the zip codes.
59 00:05:42.260 ⇒ 00:05:43.280 Amber Lin: I’m…
60 00:05:43.320 ⇒ 00:05:49.969 Casie Aviles: Yeah, we’ve, we’ve… yesterday… last week, sorry, last week, we’ve already, like, reconciled, like, inspector sheet.
61 00:05:50.520 ⇒ 00:05:57.059 Casie Aviles: I’m missing, so I found out that there are about 3,000 assignments that were added.
62 00:05:57.270 ⇒ 00:05:58.070 Casie Aviles: So…
63 00:05:58.070 ⇒ 00:05:58.720 Amber Lin: Hmm.
64 00:05:59.050 ⇒ 00:06:02.309 Casie Aviles: So those should be there now, but…
65 00:06:02.570 ⇒ 00:06:11.560 Casie Aviles: I guess I just want to be careful with what we tell them, because I don’t think we can guarantee, like, 100%, you know, accuracy.
66 00:06:11.560 ⇒ 00:06:12.230 Amber Lin: Hmm.
67 00:06:12.610 ⇒ 00:06:16.230 Casie Aviles: Because, just given the nature of the data that they have, so I don’t.
68 00:06:16.230 ⇒ 00:06:16.570 Amber Lin: Yes.
69 00:06:16.570 ⇒ 00:06:22.930 Casie Aviles: to, like, say that it’s perfect, it’s all good. I… what I… what I do think we can…
70 00:06:23.070 ⇒ 00:06:33.229 Casie Aviles: say instead is that maybe, like, you know, for example, the triage tickets related to the zip codes would go lower, so I think that’s what we can…
71 00:06:33.780 ⇒ 00:06:38.339 Casie Aviles: Gotcha. Have, like, a tangible outcome of this.
72 00:06:39.930 ⇒ 00:06:52.140 Amber Lin: Okay, I also remember we said we’re doing, so I know we did an assignment QA, and then we said we wanted to do the triage questions. Did we do that?
73 00:06:53.400 ⇒ 00:06:59.590 Casie Aviles: For the 3D Edge? No, no, we haven’t cleared too much yet, but we already have the workflow there.
74 00:07:00.070 ⇒ 00:07:01.250 Casie Aviles: Gosh, okay.
75 00:07:02.200 ⇒ 00:07:03.779 Casie Aviles: the GH tickets.
76 00:07:03.940 ⇒ 00:07:04.660 Amber Lin: Okay.
77 00:07:05.130 ⇒ 00:07:07.730 Casie Aviles: I will deduct it, yeah, I’ll deduct it there.
78 00:07:08.040 ⇒ 00:07:11.849 Amber Lin: triage, punch… I think I need to clean up Linear.
79 00:07:12.090 ⇒ 00:07:22.220 Amber Lin: Optimization… so I’ll add a test… Workflow for retesting…
80 00:07:31.780 ⇒ 00:07:33.380 Amber Lin: So that was…
81 00:07:34.410 ⇒ 00:07:37.210 Casie Aviles: Yeah, the workflow is working.
82 00:07:38.570 ⇒ 00:07:40.720 Casie Aviles: Yeah, I haven’t really run it.
83 00:07:41.990 ⇒ 00:07:47.079 Casie Aviles: You know, like, extensively to clear out the current triage tickets.
84 00:07:47.420 ⇒ 00:07:53.650 Amber Lin: Okay. Is that something we plan to do this week, or, like, what’s our plan there?
85 00:07:55.790 ⇒ 00:07:56.390 Casie Aviles: Do you have…
86 00:07:56.390 ⇒ 00:07:58.900 Amber Lin: Other things you think is more important right now?
87 00:07:59.180 ⇒ 00:08:05.870 Casie Aviles: I’m… Well, I’m just focusing on the migration work at the moment.
88 00:08:06.030 ⇒ 00:08:10.640 Casie Aviles: Since I think it’s been installed for now, so I wanted to, like…
89 00:08:10.840 ⇒ 00:08:18.090 Casie Aviles: I wanted to make some progress there. Makes sense. That’s why I slowed down a little bit on the zip code work, since there’s just…
90 00:08:18.090 ⇒ 00:08:18.820 Amber Lin: I see.
91 00:08:18.820 ⇒ 00:08:23.070 Casie Aviles: I don’t want to spend, like, more time here, because we already spent, like, a lot.
92 00:08:23.520 ⇒ 00:08:32.290 Amber Lin: Yeah, it makes sense. I think that that’s the way to go. So I’ll add another task, but we can schedule it for the week later.
93 00:08:32.299 ⇒ 00:08:33.349 Casie Aviles: Yeah, I would…
94 00:08:38.309 ⇒ 00:08:41.339 Amber Lin: I’ll put it here, then.
95 00:08:42.390 ⇒ 00:08:43.340 Casie Aviles: Okay.
96 00:08:43.730 ⇒ 00:08:44.600 Amber Lin: Okay.
97 00:08:50.120 ⇒ 00:08:54.420 Amber Lin: Should I call this QA… Bond.
98 00:08:55.970 ⇒ 00:09:00.150 Casie Aviles: Hmm, let’s see, so this is the QA for… oh, okay.
99 00:09:01.140 ⇒ 00:09:04.089 Amber Lin: Is this the QA, or did we name it wrong?
100 00:09:05.090 ⇒ 00:09:07.579 Casie Aviles: I think this is fine, since,
101 00:09:07.900 ⇒ 00:09:10.080 Casie Aviles: The goal here is to make sure that
102 00:09:10.680 ⇒ 00:09:13.380 Casie Aviles: the generated SQL queries are.
103 00:09:14.210 ⇒ 00:09:14.590 Amber Lin: Hmm.
104 00:09:14.770 ⇒ 00:09:21.929 Casie Aviles: Correct? Because… Yeah, even if, like, the data is already in the tables, if the query is wrong, then…
105 00:09:22.950 ⇒ 00:09:26.989 Casie Aviles: So that’s also kind of what’s going to be tested by the.
106 00:09:26.990 ⇒ 00:09:27.830 Amber Lin: info.
107 00:09:28.580 ⇒ 00:09:30.420 Casie Aviles: retesting workflow.
108 00:09:30.880 ⇒ 00:09:32.210 Amber Lin: Gotcha, okay.
109 00:09:37.280 ⇒ 00:09:38.230 Amber Lin: Alright.
110 00:09:40.860 ⇒ 00:09:47.310 Amber Lin: Alright, so that SIPCO side. Anything on the central dock side, Ms. Sofa?
111 00:09:48.130 ⇒ 00:09:48.960 Mustafa Raja: Yes.
112 00:09:49.360 ⇒ 00:09:52.009 Mustafa Raja: We have the mechanical one, so.
113 00:09:52.980 ⇒ 00:09:58.599 Mustafa Raja: Oh, no… I just want someone to QA, if someone is available.
114 00:09:59.300 ⇒ 00:10:08.809 Mustafa Raja: else I can, you know, make sure that we have everything in place for that. Once we’re good with that, what we would want to do is we would want to embed that.
115 00:10:09.070 ⇒ 00:10:13.060 Mustafa Raja: And see if it makes, you know, good,
116 00:10:13.330 ⇒ 00:10:16.400 Mustafa Raja: Good changes with the responses, you know?
117 00:10:16.610 ⇒ 00:10:18.309 Mustafa Raja: So that’s September.
118 00:10:18.330 ⇒ 00:10:20.870 Amber Lin: When you say QA, what does that include?
119 00:10:20.980 ⇒ 00:10:32.290 Mustafa Raja: That includes that, we, we want to make sure that, we have everything in place in this new one, so we are not missing out on any of the content.
120 00:10:32.750 ⇒ 00:10:36.560 Mustafa Raja: I say this because, most of this is,
121 00:10:36.720 ⇒ 00:10:41.079 Mustafa Raja: AI helping me out, you know, laying the content in. So…
122 00:10:42.040 ⇒ 00:10:45.210 Mustafa Raja: Yeah, some human eyes might be nice, but I can take that.
123 00:10:45.710 ⇒ 00:10:49.739 Amber Lin: I see. How would we…
124 00:10:50.110 ⇒ 00:10:54.080 Amber Lin: do that. Will we just take a piece of information.
125 00:10:54.080 ⇒ 00:10:54.800 Mustafa Raja: I think.
126 00:10:54.800 ⇒ 00:10:55.480 Amber Lin: shit.
127 00:10:55.940 ⇒ 00:11:05.119 Mustafa Raja: I don’t think that, I think, what we can do is, just, pick some random,
128 00:11:05.350 ⇒ 00:11:10.229 Mustafa Raja: definitions, and spot check it, basically, you know, for a few ones.
129 00:11:10.900 ⇒ 00:11:12.560 Mustafa Raja: Those are good, it should be good.
130 00:11:13.260 ⇒ 00:11:13.950 Samuel Roberts: Hail.
131 00:11:14.660 ⇒ 00:11:15.350 Mustafa Raja: Hey.
132 00:11:15.350 ⇒ 00:11:16.049 Amber Lin: And he said, like…
133 00:11:16.050 ⇒ 00:11:16.680 Samuel Roberts: that.
134 00:11:17.150 ⇒ 00:11:19.129 Amber Lin: Transcript went syncing.
135 00:11:20.070 ⇒ 00:11:21.920 Samuel Roberts: Just trying to get that fixed in.
136 00:11:22.580 ⇒ 00:11:36.210 Pranav Narahari: Galapa, one quick question on that test run and QA. Is that going to have to do with question categorization? Because I know we were talking about being able to differentiate on a department level, or is this completely unrelated?
137 00:11:39.520 ⇒ 00:11:58.260 Amber Lin: I think I can take that. So, when we did the question categorization, why we wanted it by department was because, mechanical, pest, law, and home improvement are different departments, and they ask slightly different questions. But right now, I think we’re only running the dock for mechanical, so…
138 00:11:58.620 ⇒ 00:12:03.820 Amber Lin: they’re… they’re related, but they’re separate things in this case, so,
139 00:12:04.210 ⇒ 00:12:13.290 Amber Lin: just no need to, like, I don’t think we need to look at the questions asked yet, we’re just comparing, did we get all the information from the previous Central Doc?
140 00:12:14.960 ⇒ 00:12:17.090 Pranav Narahari: If we got it from the previous what, sorry?
141 00:12:17.090 ⇒ 00:12:28.609 Amber Lin: The previous central doc. So the workflow is… we’re trying to optimize it and rewrite it, but we’re concerned that we might have lost something.
142 00:12:28.610 ⇒ 00:12:29.270 Pranav Narahari: I see.
143 00:12:29.270 ⇒ 00:12:30.450 Amber Lin: in translation.
144 00:12:30.950 ⇒ 00:12:32.109 Pranav Narahari: That makes sense, okay.
145 00:12:35.970 ⇒ 00:12:44.960 Pranav Narahari: Yeah, Mustafa, we can, like, sync there to talk about, like, your ideas on how to, you know, mitigate that. I’m curious to see, like, how you plan on doing that.
146 00:12:45.680 ⇒ 00:12:46.360 Mustafa Raja: Okay.
147 00:12:48.980 ⇒ 00:12:50.140 Mustafa Raja: Yeah, that works.
148 00:12:50.820 ⇒ 00:12:54.170 Pranav Narahari: Perfect, yeah. It’ll just give me more context to, like, the central dock, too, that I’ve.
149 00:12:54.170 ⇒ 00:12:54.780 Mustafa Raja: No.
150 00:12:54.780 ⇒ 00:13:00.120 Pranav Narahari: And I feel like I understand it pretty well, but this might be a good thing for us to just, like, pair on a little bit.
151 00:13:00.120 ⇒ 00:13:08.420 Mustafa Raja: Yeah, and I also want to, you know, have this cleared as soon as possible, so we can move forward with embeddings.
152 00:13:09.070 ⇒ 00:13:09.960 Pranav Narahari: Perfect, yeah.
153 00:13:13.450 ⇒ 00:13:13.800 Mustafa Raja: Yeah.
154 00:13:15.560 ⇒ 00:13:18.799 Amber Lin: Central Dog, zip code, that…
155 00:13:21.910 ⇒ 00:13:37.519 Amber Lin: I think that’s the main workflows we’re currently doing. I did want to talk about, improving how we deal with the triage questions, because we really spend very minimal time there, less than we should be, so…
156 00:13:37.740 ⇒ 00:13:42.639 Amber Lin: Last time, we said we want to sync the logs into the
157 00:13:42.770 ⇒ 00:13:45.730 Amber Lin: tickets. Do we have time to do that?
158 00:13:49.360 ⇒ 00:13:52.049 Casie Aviles: Syncing the logs to the tickets, I think.
159 00:13:52.050 ⇒ 00:13:53.440 Amber Lin: Yeah.
160 00:13:53.890 ⇒ 00:14:01.329 Casie Aviles: I think there’s… Does that mean, like, for example, the execution logs on N8N? I believe Mr.
161 00:14:01.330 ⇒ 00:14:01.980 Amber Lin: Yeah.
162 00:14:01.980 ⇒ 00:14:05.480 Casie Aviles: Something… Where he, included, like.
163 00:14:05.670 ⇒ 00:14:12.090 Casie Aviles: the exec… the execution IDs to the… to each, like, Log, or row.
164 00:14:12.660 ⇒ 00:14:16.789 Casie Aviles: I believe there was, that’s already been implemented.
165 00:14:19.040 ⇒ 00:14:20.830 Casie Aviles: But… Hmm.
166 00:14:21.610 ⇒ 00:14:27.149 Casie Aviles: What’s, like, what else to be… do we need there? I’m trying to think.
167 00:14:32.670 ⇒ 00:14:43.100 Casie Aviles: I think, yeah, we should be linking it, I believe, right, to the tickets, the execution logs. I don’t think we’ve done that yet, at least. Like, I mean, on the…
168 00:14:43.230 ⇒ 00:14:46.480 Casie Aviles: On the view, like, if you click on a ticket, there’s no…
169 00:14:46.980 ⇒ 00:14:50.630 Casie Aviles: linked to the workflow yet. I think that’s what mis… that’s what’s missing.
170 00:14:50.630 ⇒ 00:14:51.230 Mustafa Raja: Yep.
171 00:14:52.400 ⇒ 00:14:53.330 Amber Lin: I see.
172 00:14:54.820 ⇒ 00:15:03.669 Amber Lin: Well, my overall goal is just to reduce the time, or reduce the manual work we have on the triage tickets, so we can
173 00:15:04.140 ⇒ 00:15:20.570 Amber Lin: solve more in a shorter period of time, whether… whether it’s adding the execution logs, whether it’s having AI categorize the questions, like, whatever will help us, in doing it faster, is another, like.
174 00:15:21.110 ⇒ 00:15:33.379 Amber Lin: tech debt workstream, I would… not that… like, internal work stream, I would like to do. But anyways, that’s… that’s, like, that’s my main… the main workflows.
175 00:15:33.960 ⇒ 00:15:42.169 Amber Lin: what is the next thing we want to discuss in this working session? Like, I know we said we had the migration thing we wanted to talk about.
176 00:15:45.190 ⇒ 00:15:53.399 Casie Aviles: Okay, yeah, I can basically just give, show my screen, share my screen, and
177 00:15:53.900 ⇒ 00:15:58.469 Casie Aviles: Give you guys… An update on, like, what’s… what’s going on there.
178 00:16:03.100 ⇒ 00:16:06.880 Casie Aviles: Okay, yeah, so basically right now, we have this
179 00:16:07.450 ⇒ 00:16:09.909 Casie Aviles: the version of Andy, and .
180 00:16:10.250 ⇒ 00:16:13.049 Pranav Narahari: We already have it on Google Cloud.
181 00:16:13.190 ⇒ 00:16:17.389 Casie Aviles: So that’s… that should be working now. I’ve also resolved, like, the port issue.
182 00:16:17.790 ⇒ 00:16:19.099 Casie Aviles: That we were getting…
183 00:16:20.440 ⇒ 00:16:30.910 Casie Aviles: So, additionally, there’s, there are also some other changes, updates that I implemented. So, for example, I think one of the triage tickets had this issue where
184 00:16:32.500 ⇒ 00:16:40.869 Casie Aviles: Whenever a CSR would just write, without specifying the service type, which is, you know, commercial or residential.
185 00:16:41.280 ⇒ 00:16:44.519 Casie Aviles: I think Andy should be able to do that, but…
186 00:16:44.860 ⇒ 00:16:49.939 Casie Aviles: I’ll have to test this a bit more if it happens, like, regularly now, but…
187 00:16:50.790 ⇒ 00:16:54.089 Casie Aviles: That’s one thing that I added,
188 00:16:54.610 ⇒ 00:17:01.639 Casie Aviles: And then the other thing is… hold on, so, these are, like…
189 00:17:01.900 ⇒ 00:17:04.949 Casie Aviles: the, the, the logs that Mustafa worked on.
190 00:17:05.780 ⇒ 00:17:09.870 Casie Aviles: So some of them are still kind of long.
191 00:17:10.089 ⇒ 00:17:15.980 Casie Aviles: And I think what makes it long would be the DB agent.
192 00:17:16.390 ⇒ 00:17:24.069 Casie Aviles: There’s, like, 13 seconds here, 12 seconds, but then there’s also the… the routing step.
193 00:17:24.849 ⇒ 00:17:30.840 Casie Aviles: Which takes around 6 seconds, so, you know, combining those, it’ll take…
194 00:17:31.600 ⇒ 00:17:39.500 Casie Aviles: It’ll take quite long, so around 20 seconds, so I… ideally, we should be getting Light.
195 00:17:39.800 ⇒ 00:17:47.140 Casie Aviles: Around 5… In total, for each, like, response, so… .
196 00:17:47.140 ⇒ 00:17:49.179 Samuel Roberts: These are both using 4R, though, still?
197 00:17:49.600 ⇒ 00:17:50.320 Mustafa Raja: Yeah.
198 00:17:51.120 ⇒ 00:17:55.780 Samuel Roberts: Okay. So that may improve itself when we switch models anyway.
199 00:17:55.780 ⇒ 00:17:57.050 Mustafa Raja: But… Yeah.
200 00:17:57.560 ⇒ 00:18:01.379 Samuel Roberts: Still something to keep an eye on. And we have to test those models anyway, so…
201 00:18:03.090 ⇒ 00:18:03.860 Mustafa Raja: Thank you.
202 00:18:06.900 ⇒ 00:18:15.290 Casie Aviles: I was also wondering, like, could it… would it… could it… could the factor be, like, the Google Cloud itself, or…
203 00:18:15.410 ⇒ 00:18:19.380 Casie Aviles: Maybe with the code, maybe…
204 00:18:19.860 ⇒ 00:18:22.640 Casie Aviles: There are some things that make it, like.
205 00:18:23.330 ⇒ 00:18:28.950 Casie Aviles: Longer, or is it purely, like… The AI generation step.
206 00:18:29.560 ⇒ 00:18:30.719 Casie Aviles: I think that’s something I…
207 00:18:30.720 ⇒ 00:18:35.389 Mustafa Raja: Both of these steps are just, you know, returning your surface back.
208 00:18:37.420 ⇒ 00:18:38.720 Casie Aviles: Hmm. Is…
209 00:18:38.720 ⇒ 00:18:43.650 Samuel Roberts: Is this step breakdown, do you program that for specifically the steps?
210 00:18:44.070 ⇒ 00:18:48.329 Mustafa Raja: Those two steps, or… No, no, no, so.
211 00:18:48.330 ⇒ 00:18:49.459 Samuel Roberts: It’ll show whatever the steps are.
212 00:18:49.460 ⇒ 00:18:55.830 Mustafa Raja: Yeah, this is going to be whatever steps is, whatever steps are taken, you know… Okay, okay.
213 00:18:56.060 ⇒ 00:19:00.560 Samuel Roberts: then yeah, it’s probably just that. I mean, we can do a little more experimentation that way, but I think once we…
214 00:19:00.910 ⇒ 00:19:04.829 Samuel Roberts: It’s probably not worth too much optimization of these numbers until we…
215 00:19:05.490 ⇒ 00:19:08.000 Samuel Roberts: Switch over and see how bad they are with the new models.
216 00:19:08.280 ⇒ 00:19:09.470 Mustafa Raja: Hmm, yeah, Tim.
217 00:19:09.470 ⇒ 00:19:13.760 Samuel Roberts: And we also may want to play with the models, like, the routing might not need the same model as the DB, for example.
218 00:19:13.760 ⇒ 00:19:18.529 Mustafa Raja: Yeah, I think a much lighter model should be able to handle that.
219 00:19:18.810 ⇒ 00:19:20.250 Samuel Roberts: I think so, yeah.
220 00:19:21.250 ⇒ 00:19:26.720 Mustafa Raja: And Tim got back to me today, so he… Oh, good it might… Do it.
221 00:19:27.180 ⇒ 00:19:29.190 Mustafa Raja: I don’t know, but we’ll see.
222 00:19:29.860 ⇒ 00:19:30.610 Samuel Roberts: Okay, that’d be…
223 00:19:30.610 ⇒ 00:19:38.630 Mustafa Raja: He was just confirming that, we do not need Vertex AI, so I just said, no, we do not need… just the Gemini API should be enough.
224 00:19:42.380 ⇒ 00:19:43.590 Samuel Roberts: Say that again? Sorry.
225 00:19:43.590 ⇒ 00:19:48.700 Mustafa Raja: So, Tim was asking if we need a vertex AI, you know?
226 00:19:48.700 ⇒ 00:19:49.429 Samuel Roberts: Oh, yeah, yeah, yeah.
227 00:19:49.430 ⇒ 00:19:55.279 Mustafa Raja: I just said, the Gemini API should be, should be enough for us.
228 00:19:56.450 ⇒ 00:19:58.570 Samuel Roberts: Okay, yeah, that’s totally perfect.
229 00:20:01.640 ⇒ 00:20:11.500 Casie Aviles: Okay, so for… yeah, for next steps here, I guess we’ll have to… yeah, we’ll have to test this with the Gemini models that we’re going to…
230 00:20:12.410 ⇒ 00:20:22.149 Casie Aviles: Yeah, we’re gonna have, and I think, yeah, that’s what’s the next step that we could do to see if we could reduce, like, the execution times.
231 00:20:23.440 ⇒ 00:20:25.370 Casie Aviles: Okay.
232 00:20:26.260 ⇒ 00:20:31.980 Casie Aviles: But yeah, right now, it’s okay, it’s working right now. It’s just the execution time, since…
233 00:20:32.300 ⇒ 00:20:35.529 Casie Aviles: Like, I believe, like, one of the goals that we wanted to is…
234 00:20:35.780 ⇒ 00:20:46.410 Casie Aviles: Achieve, like, with this migration is to, you know, reduce the… Execution time, so… Yeah.
235 00:20:46.410 ⇒ 00:20:50.719 Samuel Roberts: He’s… I mean, is it mostly the DB ones that are the long ones?
236 00:20:51.940 ⇒ 00:20:53.810 Casie Aviles: Yeah, I have it, like.
237 00:20:54.030 ⇒ 00:20:59.530 Casie Aviles: tested it with central doc-related questions. These are mostly VB ones.
238 00:20:59.960 ⇒ 00:21:03.570 Samuel Roberts: Okay, yeah, I mean, it looks like that was the… the bulk of the time, so hopefully that’s…
239 00:21:03.720 ⇒ 00:21:06.419 Samuel Roberts: The ones that we can optimize, then, with the model.
240 00:21:08.490 ⇒ 00:21:09.350 Casie Aviles: Okay.
241 00:21:11.550 ⇒ 00:21:17.370 Casie Aviles: Let’s see, I guess another… Yeah, like…
242 00:21:18.780 ⇒ 00:21:22.869 Casie Aviles: Another thing that we… I talked on briefly about, Mustafa, was
243 00:21:23.250 ⇒ 00:21:27.530 Casie Aviles: Whether we should set up, like, another…
244 00:21:28.900 ⇒ 00:21:33.420 Casie Aviles: endpoint here in N8N that basically gets the…
245 00:21:34.100 ⇒ 00:21:43.960 Casie Aviles: output from the current Andy, and then also runs it through the Bastra Andy, so…
246 00:21:44.360 ⇒ 00:21:48.079 Casie Aviles: We would have, like, A direct comparison of…
247 00:21:48.950 ⇒ 00:21:51.720 Casie Aviles: the results as it is now, and then… Yes.
248 00:21:52.750 ⇒ 00:21:53.830 Samuel Roberts: Definitely.
249 00:21:54.710 ⇒ 00:21:58.480 Samuel Roberts: I think it’s a good idea, yeah, and so it’s after the response. Yeah, that’s perfect.
250 00:21:59.780 ⇒ 00:22:00.880 Casie Aviles: Okay.
251 00:22:00.880 ⇒ 00:22:03.399 Samuel Roberts: We even want to do after the execution time?
252 00:22:06.350 ⇒ 00:22:07.839 Casie Aviles: I see, yeah. Okay.
253 00:22:07.840 ⇒ 00:22:12.279 Samuel Roberts: Whatever information we have, we just want to run that again, because it doesn’t matter, you know.
254 00:22:13.890 ⇒ 00:22:17.850 Samuel Roberts: You know, no one’s seeing that result except us, so I think it’s fine to do it wherever it matters.
255 00:22:18.690 ⇒ 00:22:19.290 Casie Aviles: Hmm, I saw that.
256 00:22:19.290 ⇒ 00:22:20.830 Samuel Roberts: Find all the information I need.
257 00:22:20.970 ⇒ 00:22:22.600 Casie Aviles: Respond to work, yeah.
258 00:22:22.600 ⇒ 00:22:27.159 Samuel Roberts: Yeah, definitely after that, but then potentially… is there any other stuff? This is just all writing stuff?
259 00:22:27.360 ⇒ 00:22:40.259 Samuel Roberts: the summary score… Yeah, we’ve deactivated some of these even, so it’s… Okay, yeah, I would say, after that execution time one, let’s run it, that way we can log probably both of those together, side by side, even.
260 00:22:41.590 ⇒ 00:22:50.899 Samuel Roberts: Okay, yeah, great. In somewhere else, and then, you know, I wouldn’t worry about tying them together in a snowflake, but just to get a log of, yeah, that would be very helpful.
261 00:22:52.620 ⇒ 00:22:58.210 Casie Aviles: Okay, so that’s one more thing we will work on as well. Cool.
262 00:22:59.150 ⇒ 00:23:05.169 Casie Aviles: Alright, are there, like, any other checks or tests that we need to be doing?
263 00:23:05.670 ⇒ 00:23:06.909 Casie Aviles: In order to, like…
264 00:23:07.380 ⇒ 00:23:12.710 Casie Aviles: make sure that, this is all good now. Because right now I’m just testing, like, the…
265 00:23:14.390 ⇒ 00:23:19.260 Casie Aviles: The response quality, and then the execution time, so…
266 00:23:21.160 ⇒ 00:23:25.320 Samuel Roberts: I think adding that piece and letting it run for a day or two would be…
267 00:23:25.750 ⇒ 00:23:35.430 Samuel Roberts: a ton of testing for us, effectively, and then we can evaluate from there if the execution times are still too long. Obviously, the model, so there’s a little bit of timing there, but,
268 00:23:36.760 ⇒ 00:23:39.500 Samuel Roberts: I think that’ll give us a bunch of real-world testing, because…
269 00:23:40.920 ⇒ 00:23:44.799 Samuel Roberts: That will be real usage, just not end result going to the user.
270 00:23:46.170 ⇒ 00:23:46.880 Casie Aviles: Okay.
271 00:23:47.980 ⇒ 00:23:49.139 Casie Aviles: Sounds good.
272 00:23:50.030 ⇒ 00:23:56.790 Casie Aviles: Make sure to add. So I’ll just operate, like, another… a sub-workflow here that will… do that.
273 00:23:58.110 ⇒ 00:24:02.940 Casie Aviles: Okay, yeah, but other than that, I think, I think we’re okay now. This is already, like.
274 00:24:03.520 ⇒ 00:24:12.189 Casie Aviles: Working, you know, the continuous deployment is working, so whenever we push changes to the…
275 00:24:13.790 ⇒ 00:24:17.109 Casie Aviles: The branch, it’s going to deploy automatically.
276 00:24:17.470 ⇒ 00:24:20.649 Casie Aviles: I guess one question I had there was…
277 00:24:21.520 ⇒ 00:24:26.180 Casie Aviles: What would be the best, like, or what would be the ideal…
278 00:24:26.390 ⇒ 00:24:29.220 Casie Aviles: Set up with the branches, because right now it’s a bit…
279 00:24:30.380 ⇒ 00:24:30.880 Mustafa Raja: Hmm…
280 00:24:33.010 ⇒ 00:24:39.250 Casie Aviles: We only have, like, the migration progress, that’s kind of the one that we’re actively… it’s kind of our development branch.
281 00:24:40.020 ⇒ 00:24:44.990 Casie Aviles: Where we do, like, you know, we update, but it’s not the one deployed.
282 00:24:45.600 ⇒ 00:24:50.860 Casie Aviles: Or the one that’s live in fraud, so maybe we need to, like, create
283 00:24:51.400 ⇒ 00:24:56.689 Casie Aviles: Similar to what we did with Lilo, we have two, like, a staging and then a live
284 00:24:57.130 ⇒ 00:25:01.700 Casie Aviles: Branch. Would that kind of make sense here as well?
285 00:25:03.400 ⇒ 00:25:07.700 Samuel Roberts: Yeah, so what is the current deployed one if it’s not migration progress?
286 00:25:08.590 ⇒ 00:25:15.929 Casie Aviles: Oh, I mean… yeah, sorry, it’s a bit confusing, but this one is just the dev one, the deployed dev one.
287 00:25:16.410 ⇒ 00:25:18.560 Samuel Roberts: Oh, oh, I see, okay, yeah, yeah.
288 00:25:19.320 ⇒ 00:25:23.190 Casie Aviles: And then… Let’s see…
289 00:25:23.640 ⇒ 00:25:29.200 Casie Aviles: There’s also, like, main as well, so… That’s the one, I believe.
290 00:25:31.010 ⇒ 00:25:32.450 Samuel Roberts: Yeah, I mean, as long as…
291 00:25:32.620 ⇒ 00:25:35.569 Samuel Roberts: it’s more… I don’t know exactly how it’s set up in…
292 00:25:35.920 ⇒ 00:25:48.079 Samuel Roberts: in GCP, but as long as we can do the same kind of watch folder situation as Railway, we could probably just use main. If we can’t, and it’s more like Heroku, then we should definitely set up the production.
293 00:25:48.260 ⇒ 00:25:50.520 Samuel Roberts: Like, a monster production branch.
294 00:25:52.040 ⇒ 00:25:52.990 Casie Aviles: Okay.
295 00:25:54.490 ⇒ 00:25:55.330 Casie Aviles: Okay.
296 00:25:55.330 ⇒ 00:26:05.400 Samuel Roberts: I don’t know, because, like, yeah, Heroku was rebuilding every time a thing got pushed to a main, which was… not ideal. Railway at least checks the watch folder first. I don’t know what GCP does.
297 00:26:05.740 ⇒ 00:26:07.240 Samuel Roberts: If there’s a way to specify.
298 00:26:10.690 ⇒ 00:26:12.350 Casie Aviles: Okay, yeah, yeah.
299 00:26:15.510 ⇒ 00:26:18.659 Casie Aviles: Mmm… yeah, I think, I think that’s, that’s…
300 00:26:19.070 ⇒ 00:26:20.579 Mustafa Raja: All I have right now.
301 00:26:20.580 ⇒ 00:26:21.899 Casie Aviles: Yeah, go ahead, Mustafa.
302 00:26:21.900 ⇒ 00:26:25.899 Mustafa Raja: Also, I’m wondering if, if changing the region
303 00:26:26.480 ⇒ 00:26:29.900 Mustafa Raja: Would also improve the response times, because this is…
304 00:26:29.900 ⇒ 00:26:30.360 Samuel Roberts: Oh, yeah.
305 00:26:30.360 ⇒ 00:26:31.380 Casie Aviles: Yes, we…
306 00:26:32.050 ⇒ 00:26:36.089 Samuel Roberts: Yeah, the… the other ones, the database and the model.
307 00:26:36.090 ⇒ 00:26:36.430 Mustafa Raja: Should be…
308 00:26:36.430 ⇒ 00:26:37.450 Samuel Roberts: deployed in…
309 00:26:37.450 ⇒ 00:26:47.529 Mustafa Raja: Yeah, because even when I… I’m working with a Postgres database locally on my computer, it takes… it takes a lot of time to respond.
310 00:26:47.680 ⇒ 00:26:52.670 Mustafa Raja: So, this could just be it, or at least one of the reasons.
311 00:26:53.450 ⇒ 00:26:57.930 Mustafa Raja: So… This might improve the response time, I believe?
312 00:26:58.590 ⇒ 00:26:59.240 Casie Aviles: Yeah.
313 00:27:00.310 ⇒ 00:27:02.630 Mustafa Raja: Make sure that we have… we have…
314 00:27:03.140 ⇒ 00:27:07.080 Mustafa Raja: Same regions for, database and this.
315 00:27:07.630 ⇒ 00:27:12.759 Mustafa Raja: I think, I don’t know which region, I forgot which region is it.
316 00:27:12.760 ⇒ 00:27:16.270 Samuel Roberts: I think it’s, like, US South something. Yeah, yeah, yeah. South 1?
317 00:27:17.330 ⇒ 00:27:18.830 Mustafa Raja: Let me confirm that.
318 00:27:21.210 ⇒ 00:27:24.130 Samuel Roberts: But that’s, like, in Texas, so I figured that would be the best for them.
319 00:27:24.770 ⇒ 00:27:25.780 Mustafa Raja: Yeah.
320 00:27:35.540 ⇒ 00:27:36.680 Samuel Roberts: Yeah, I don’t know how.
321 00:27:38.120 ⇒ 00:27:40.579 Samuel Roberts: There’s a good way to easily move stuff.
322 00:27:50.270 ⇒ 00:27:56.140 Samuel Roberts: It just says to move, you have to create a new one and delete the old one, I don’t know if there’s a good way to actually move it.
323 00:27:57.680 ⇒ 00:28:02.070 Casie Aviles: Yeah, I guess I’ll just create, like, a… a new…
324 00:28:02.670 ⇒ 00:28:03.010 Samuel Roberts: Yeah.
325 00:28:04.180 ⇒ 00:28:05.990 Mustafa Raja: Hmm, it’s your south one.
326 00:28:06.130 ⇒ 00:28:06.970 Mustafa Raja: For reference.
327 00:28:07.690 ⇒ 00:28:08.360 Samuel Roberts: Cool.
328 00:28:09.500 ⇒ 00:28:10.210 Casie Aviles: R.
329 00:28:12.440 ⇒ 00:28:13.599 Casie Aviles: Just set that.
330 00:28:15.480 ⇒ 00:28:17.139 Samuel Roberts: Yeah, Dallas, cool. Okay.
331 00:28:17.390 ⇒ 00:28:18.460 Casie Aviles: U.S.
332 00:28:18.700 ⇒ 00:28:21.870 Casie Aviles: So… oh, okay, Dallas.
333 00:28:24.140 ⇒ 00:28:24.660 Samuel Roberts: Cool.
334 00:28:25.350 ⇒ 00:28:29.079 Samuel Roberts: Yeah, it’ll be interesting to see if that changes the times in any measurable way.
335 00:28:29.290 ⇒ 00:28:30.190 Mustafa Raja: Yeah.
336 00:28:40.220 ⇒ 00:28:46.010 Casie Aviles: Okay, yeah, I’ll just… I’ll set this, but… you guys have anything else,
337 00:28:49.840 ⇒ 00:28:53.209 Casie Aviles: I think that’s all I had for ABC right now.
338 00:29:00.060 ⇒ 00:29:09.799 Pranav Narahari: Another part of the migration effort is, giving that thumbs up, thumbs down feedback, right? I think currently that’s being supported by NADN.
339 00:29:11.020 ⇒ 00:29:11.670 Casie Aviles: Yes.
340 00:29:12.430 ⇒ 00:29:13.160 Pranav Narahari: Okay.
341 00:29:15.730 ⇒ 00:29:20.879 Pranav Narahari: Is that gonna… is that part of the scope of this migration, or is that different?
342 00:29:21.200 ⇒ 00:29:22.799 Mustafa Raja: Yeah, it’s part of this.
343 00:29:22.800 ⇒ 00:29:25.659 Samuel Roberts: Yeah, there’s a… there’s something in the Gantt, I think, for it.
344 00:29:26.400 ⇒ 00:29:27.760 Pranav Narahari: Oh, okay, cool.
345 00:29:28.550 ⇒ 00:29:34.720 Pranav Narahari: Yeah, Amber, I’m not sure for the Gantt if you’re able to invite me, I think it might just be Riko or Utena.
346 00:29:34.720 ⇒ 00:29:40.419 Amber Lin: Oh, yeah, tag Rico, he should be able to invite you. Let me check, let me check.
347 00:29:40.420 ⇒ 00:29:42.040 Mustafa Raja: I think I’m looking white.
348 00:29:42.280 ⇒ 00:29:48.580 Amber Lin: Yeah, let me see, projects…
349 00:29:49.740 ⇒ 00:29:52.479 Mustafa Raja: I think whoever owns the project can invite or something.
350 00:29:54.240 ⇒ 00:29:58.339 Amber Lin: Yeah, let me grab your email, I should be able to add you.
351 00:30:26.830 ⇒ 00:30:32.989 Amber Lin: Okay, I think it… I… it should have sent an invite to you, so maybe check your…
352 00:30:33.780 ⇒ 00:30:36.829 Amber Lin: Check your brain forge email.
353 00:30:37.620 ⇒ 00:30:38.880 Pranav Narahari: Yeah, let me check.
354 00:30:52.760 ⇒ 00:30:54.420 Pranav Narahari: Cool, yep, I got it.
355 00:30:54.590 ⇒ 00:30:56.010 Amber Lin: Awesome.
356 00:31:03.820 ⇒ 00:31:04.500 Pranav Narahari: Nice.
357 00:31:11.750 ⇒ 00:31:20.620 Amber Lin: Cool. Anything else? Would you like to use this working session to just do some of the work, or do you guys want to take it offline?
358 00:31:23.410 ⇒ 00:31:27.590 Casie Aviles: I don’t mind staying on, I’ll just… I’ll still work on ABC anyway.
359 00:31:27.860 ⇒ 00:31:28.400 Amber Lin: Cool.
360 00:31:28.400 ⇒ 00:31:29.419 Casie Aviles: Even if I hop off.
361 00:31:29.420 ⇒ 00:31:33.839 Amber Lin: Thank you. Awesome, awesome. Pranav, do you want to talk about the slides?
362 00:31:34.900 ⇒ 00:31:36.650 Pranav Narahari: Yeah, we can do that now.
363 00:31:36.920 ⇒ 00:31:37.620 Amber Lin: Okay.
364 00:31:37.980 ⇒ 00:31:39.610 Amber Lin: Cool.
365 00:31:39.840 ⇒ 00:31:44.509 Amber Lin: And folks, if you don’t want to hear our constant talking, you can feel free to hop off.
366 00:31:45.140 ⇒ 00:31:47.980 Samuel Roberts: Okay. Let me know if you need me for anything, I’ll…
367 00:31:48.320 ⇒ 00:31:48.690 Amber Lin: Awesome.
368 00:31:49.010 ⇒ 00:31:50.630 Mustafa Raja: Also, Pranav, let me know…
369 00:31:50.950 ⇒ 00:31:55.449 Mustafa Raja: Let me know when you’ll be available to talk about the, Central Doc stuff.
370 00:31:55.980 ⇒ 00:31:59.369 Pranav Narahari: Yeah, definitely. Right after this, I’ll, find some time.
371 00:31:59.620 ⇒ 00:32:01.589 Mustafa Raja: Okay, okay, yeah, just ping me. Thank you.
372 00:32:01.800 ⇒ 00:32:03.150 Pranav Narahari: Yeah, thanks, Mustafa.
373 00:32:03.150 ⇒ 00:32:04.120 Mustafa Raja: Okay, bye.
374 00:32:04.120 ⇒ 00:32:04.850 Pranav Narahari: Thanks, guys.
375 00:32:05.980 ⇒ 00:32:06.770 Casie Aviles: Q.
376 00:32:09.840 ⇒ 00:32:12.969 Pranav Narahari: Cool. Yeah, so maybe we can just start off with just, like.
377 00:32:13.100 ⇒ 00:32:15.430 Pranav Narahari: How you usually go about creating slides?
378 00:32:15.430 ⇒ 00:32:22.680 Amber Lin: Yeah, let me… I’ll share screen. I also… I’m trying to fix this scan so it’s easier for you to read for the first time, so…
379 00:32:23.360 ⇒ 00:32:24.170 Amber Lin: think…
380 00:32:25.160 ⇒ 00:32:37.019 Amber Lin: these two are just central docs. I think the other ones are more explanatory. So, first, let me share this with you. We currently just keep a running dock, so it shouldn’t be…
381 00:32:39.260 ⇒ 00:32:49.200 Amber Lin: Shouldn’t… be too many to catch up on. All the past slides are kind of here, so…
382 00:32:49.520 ⇒ 00:33:06.229 Amber Lin: Last week is a bit different. Last week, we did more of a, what can we analyze on the transcript side, but usually, it’s pretty straightforward, just… I start off with usage, which is a screenshot with a link, so I don’t update this all that much.
383 00:33:06.350 ⇒ 00:33:08.889 Amber Lin: I do usage.
384 00:33:09.040 ⇒ 00:33:18.700 Amber Lin: then I do the workflows we’re doing. So, zip codes, transcripts, central Dock.
385 00:33:18.900 ⇒ 00:33:24.819 Amber Lin: And then migration, if there’s any. So, I think pretty simple…
386 00:33:25.290 ⇒ 00:33:36.760 Amber Lin: I would just update the date, 2… March 5th… Copy the usage.
387 00:33:37.230 ⇒ 00:33:44.359 Amber Lin: And then copy the main workflows that we are doing.
388 00:33:44.790 ⇒ 00:33:51.960 Amber Lin: So, copy that… It usually takes me about, like, 15, 20 minutes.
389 00:33:52.260 ⇒ 00:33:54.789 Amber Lin: To update this, so…
390 00:33:55.060 ⇒ 00:34:03.640 Amber Lin: I start with this. I start with this framework, and then just write in any updates. What did we accomplish this week? What are we going to do next week?
391 00:34:03.640 ⇒ 00:34:04.270 Pranav Narahari: Yep.
392 00:34:04.550 ⇒ 00:34:13.900 Amber Lin: And then, if there’s any specific things we would show, ideally, they prefer screenshots, so I would paste in,
393 00:34:14.210 ⇒ 00:34:17.300 Amber Lin: An image like…
394 00:34:17.580 ⇒ 00:34:26.480 Amber Lin: like that, of, oh, this is what… this is what we’re able to do, this is what we’re doing. So, slides are not pretty, but just…
395 00:34:26.600 ⇒ 00:34:39.560 Amber Lin: framework to guide them through what we’re talking about. And you don’t always have to talk about all the sections. If there’s nothing important happening there, like, feel free to leave it out.
396 00:34:41.429 ⇒ 00:34:41.879 Pranav Narahari: Gotcha.
397 00:34:42.219 ⇒ 00:34:43.340 Amber Lin: Yeah, so…
398 00:34:43.340 ⇒ 00:34:47.670 Pranav Narahari: The usage in the last 7 days, so you said this is usually, like,
399 00:34:47.790 ⇒ 00:34:51.040 Pranav Narahari: A link? Okay. Oh, so this image… Alexis.
400 00:34:51.040 ⇒ 00:35:03.100 Amber Lin: goes… it goes to real. Sometimes I update this, so I would take a new screenshot. Make sure that up here, it is the last 7 days. I usually take
401 00:35:03.340 ⇒ 00:35:09.399 Amber Lin: a new screenshot on Thursday, but you don’t have to. Feel free to just click this link and go here.
402 00:35:09.760 ⇒ 00:35:17.269 Amber Lin: and show them this. Sometimes they will like you to click down into the specific persons under there, but…
403 00:35:17.820 ⇒ 00:35:18.960 Amber Lin: All good.
404 00:35:19.340 ⇒ 00:35:22.219 Amber Lin: So… let’s see…
405 00:35:22.220 ⇒ 00:35:26.109 Pranav Narahari: Okay, looks like I need access for this, too. So I just requested.
406 00:35:26.730 ⇒ 00:35:30.549 Amber Lin: Okay, let me… let me check if I can grant you access.
407 00:35:31.300 ⇒ 00:35:35.900 Amber Lin: Mmm… Hmm.
408 00:35:36.350 ⇒ 00:35:42.120 Amber Lin: Maybe you’ll need someone to… Add you.
409 00:35:42.720 ⇒ 00:35:48.840 Amber Lin: Settings… Yeah… Okay.
410 00:35:49.150 ⇒ 00:35:52.929 Amber Lin: rico or someone will call back in.
411 00:35:53.520 ⇒ 00:35:56.129 Pranav Narahari: Perfect, yeah, let me just ping them real quick now.
412 00:35:56.130 ⇒ 00:35:56.930 Amber Lin: Awesome.
413 00:35:57.980 ⇒ 00:35:59.150 Amber Lin: And…
414 00:35:59.860 ⇒ 00:36:17.389 Amber Lin: as we gather the client’s requests throughout the week, for example, on Monday, they would say, hey, can you look at this? This is a thing that we want to work on. I usually mention them in this meeting, just to make sure they know we’re working on them, that we’ve heard them. So, in this case, on this.
415 00:36:17.390 ⇒ 00:36:20.950 Amber Lin: To do, ideally, if we can do that this week.
416 00:36:20.950 ⇒ 00:36:27.410 Amber Lin: is, review, Unknown.
417 00:36:27.990 ⇒ 00:36:32.220 Amber Lin: And re… Assign in.
418 00:36:32.930 ⇒ 00:36:33.740 Amber Lin: Fred.
419 00:36:40.140 ⇒ 00:36:47.229 Amber Lin: So, how we… a little bit more context here. So, we have a few unknowns right here.
420 00:36:47.720 ⇒ 00:36:58.090 Amber Lin: So, we have quite a few people here. And how we assigned them, their…
421 00:36:59.380 ⇒ 00:37:06.210 Amber Lin: their departments is with this. So, let me make sure you have access here, too.
422 00:37:06.910 ⇒ 00:37:11.440 Amber Lin: So I’ll send this to you.
423 00:37:13.050 ⇒ 00:37:25.750 Amber Lin: And how we did it is we take this sheet, we have their name, and most importantly, their email, and we ask the clients to add the email, add the department, and role.
424 00:37:26.880 ⇒ 00:37:32.440 Amber Lin: So… Let me copy and paste this under… Here…
425 00:37:33.650 ⇒ 00:37:53.639 Amber Lin: So, they said, hey, we’ve already added the people under unknowns. I don’t know if they added every single person there, so that’s something I want Casey to check, say, hey, have they added this, or if they’ve added it, why hasn’t it been updated in… in real? That’s… that’s the thing there.
426 00:37:53.760 ⇒ 00:37:58.160 Amber Lin: I think not too many progress on trans… oh, go ahead.
427 00:37:58.550 ⇒ 00:38:05.960 Pranav Narahari: Yeah, just so I understand, like, there’s… this unknown is just, like, there’s users that are not assigned to a department, and…
428 00:38:06.100 ⇒ 00:38:09.780 Pranav Narahari: they’re wondering, okay, why is that the case? And then… Yeah.
429 00:38:10.120 ⇒ 00:38:11.349 Pranav Narahari: Okay, gotcha.
430 00:38:11.350 ⇒ 00:38:18.449 Amber Lin: Yeah, because I told them this updates automatically based on the spreadsheet. Maybe I was wrong, maybe it’s not a live sync.
431 00:38:18.560 ⇒ 00:38:19.090 Amber Lin: So…
432 00:38:19.540 ⇒ 00:38:20.749 Amber Lin: We’ll need to check.
433 00:38:21.320 ⇒ 00:38:21.800 Pranav Narahari: Okay.
434 00:38:22.680 ⇒ 00:38:41.829 Amber Lin: On the workflows, I think transcript side, we’re… no, not really movement there, because we’re still waiting for them to talk about renewal, because I think Utam says this is less of an Andy project, more of a discovery workflow, so they’re talking about renewal there, so…
435 00:38:41.830 ⇒ 00:38:48.890 Amber Lin: maybe this week, I would… I would delete that, so… no need to talk about it, I know… unless you wanna…
436 00:38:49.030 ⇒ 00:38:51.070 Amber Lin: Push them towards renewal.
437 00:38:52.870 ⇒ 00:39:02.409 Amber Lin: And so, I think migration would be the first part, so we would say what we’re doing this week,
438 00:39:03.260 ⇒ 00:39:08.489 Amber Lin: Like, don’t need to… Go too technical for this client.
439 00:39:08.730 ⇒ 00:39:22.800 Amber Lin: So, mostly business impact of, oh, we’re doing this, so soon you will be able to do this. So, very, very non-technical audience. So, if we go too deep, they will zone out, they will have no clue what you’re talking about.
440 00:39:23.020 ⇒ 00:39:26.780 Pranav Narahari: Gotcha. So I can just be, like, updated certain GCP parameters to…
441 00:39:26.820 ⇒ 00:39:31.420 Amber Lin: That’s too technical. They don’t know what GCP parameters are.
442 00:39:31.420 ⇒ 00:39:39.740 Pranav Narahari: Okay, okay, fair. So then, yeah, latency… yeah, latency.
443 00:39:39.740 ⇒ 00:39:42.749 Amber Lin: What we can see… what we can say, then, is…
444 00:39:43.360 ⇒ 00:39:46.799 Amber Lin: Working on making Andy respond faster.
445 00:39:47.190 ⇒ 00:39:48.280 Pranav Narahari: Yeah. So…
446 00:39:48.280 ⇒ 00:40:06.000 Amber Lin: Okay. We found some ways to make Andy respond faster. We’re testing, different ways to make it faster, and then you can say something about the timeline of, oh, we think, you should be able to test this
447 00:40:06.210 ⇒ 00:40:11.770 Amber Lin: This week, we can say, Tim is helping…
448 00:40:12.180 ⇒ 00:40:24.679 Amber Lin: migrate AI models, you can say, hey, good news, we have a new AI model from whatever that might make it faster. So, very high level, no technical details, because that will distract them.
449 00:40:24.830 ⇒ 00:40:27.200 Amber Lin: There’s, like, simple, simple, simple stuff.
450 00:40:29.130 ⇒ 00:40:29.890 Pranav Narahari: Gotcha.
451 00:40:30.150 ⇒ 00:40:30.700 Amber Lin: Yeah.
452 00:40:32.130 ⇒ 00:40:34.290 Amber Lin: Improving…
453 00:40:37.490 ⇒ 00:40:39.860 Amber Lin: Yeah, and on the…
454 00:40:41.470 ⇒ 00:40:58.059 Amber Lin: Final QA… we can talk about here on the zip code side, Casey said, hey, it’s not going to be 100%, but we do think we’ll see a reduction in triage tickets on the zip code side, and…
455 00:40:59.460 ⇒ 00:41:11.640 Amber Lin: We can say we created workflow to auto… 2… Automatically retest old triage questions.
456 00:41:11.770 ⇒ 00:41:16.970 Amber Lin: So we’re gonna say we’re gonna run that workflow, and…
457 00:41:18.680 ⇒ 00:41:27.500 Amber Lin: It’s also important to, talk about your asks of the client. So… In this case.
458 00:41:27.660 ⇒ 00:41:31.280 Amber Lin: I think our ask for them is that
459 00:41:31.780 ⇒ 00:41:37.059 Amber Lin: hey, can you keep the database up to date? If there’s any changes.
460 00:41:37.270 ⇒ 00:41:47.730 Amber Lin: please, if you update it in the spreadsheet, please also update it in the database, because that helps us give accurate responses. So, essentially, data freshness request.
461 00:41:47.790 ⇒ 00:41:58.519 Amber Lin: For them. But just make that really, really clear, and we will also follow up in the following meetings, and what specifically do we need them to do.
462 00:41:59.490 ⇒ 00:42:00.660 Pranav Narahari: Cool, yeah.
463 00:42:03.620 ⇒ 00:42:08.310 Pranav Narahari: And when we say database, it’s just that Excel sheet, or that Google Sheet, right?
464 00:42:09.000 ⇒ 00:42:15.999 Amber Lin: The database here, we mean the admin UI. So, I think Casey shared the link with you…
465 00:42:16.000 ⇒ 00:42:18.819 Pranav Narahari: Yeah, they make updates straight in there.
466 00:42:20.580 ⇒ 00:42:33.080 Amber Lin: So, Janiece does the updates in there. I think they… because they so use their spreadsheets, they do also update there, but Janice takes that update and inputs it into the database.
467 00:42:33.430 ⇒ 00:42:34.990 Pranav Narahari: Okay, gotcha.
468 00:42:39.090 ⇒ 00:42:39.790 Amber Lin: Cool.
469 00:42:41.670 ⇒ 00:42:44.369 Pranav Narahari: That’s the one pass, I think.
470 00:42:44.710 ⇒ 00:42:51.849 Amber Lin: I see. I think… I’ll just put it there. I think it’s okay, the link will expire.
471 00:42:52.080 ⇒ 00:42:55.560 Amber Lin: Or I can grab… I can grab the link.
472 00:42:56.960 ⇒ 00:42:57.585 Amber Lin: Hmm…
473 00:43:01.040 ⇒ 00:43:02.689 Pranav Narahari: Okay, whatever.
474 00:43:02.690 ⇒ 00:43:04.450 Amber Lin: You know where the link is.
475 00:43:05.250 ⇒ 00:43:05.920 Pranav Narahari: Yeah, yeah.
476 00:43:08.310 ⇒ 00:43:12.249 Amber Lin: Last thing is the central dock overhaul.
477 00:43:14.650 ⇒ 00:43:18.270 Amber Lin: What did we do this week? We didn’t do much this week.
478 00:43:18.620 ⇒ 00:43:20.269 Amber Lin: We valid…
479 00:43:20.270 ⇒ 00:43:22.489 Pranav Narahari: The stuff I was talking about, yeah.
480 00:43:22.490 ⇒ 00:43:28.729 Amber Lin: Yeah, validated… matched? Mapped content?
481 00:43:32.660 ⇒ 00:43:36.709 Amber Lin: how do I… how do we call it? We essentially look for missing content.
482 00:43:37.490 ⇒ 00:43:38.440 Pranav Narahari: Yeah.
483 00:43:38.980 ⇒ 00:43:40.840 Pranav Narahari: On a department level.
484 00:43:40.840 ⇒ 00:43:50.920 Amber Lin: Yeah, compared… New versus old dogs for… Any missing content?
485 00:43:51.730 ⇒ 00:44:03.499 Amber Lin: Next steps… Mmm… use AI to test, like, something… even something like embeddings is too much.
486 00:44:03.720 ⇒ 00:44:11.999 Amber Lin: So, gloss over those and just say, we’re gonna test it. We’re gonna compare things. So, no.
487 00:44:12.000 ⇒ 00:44:16.299 Pranav Narahari: So when you see things like that, do they… Kind of just…
488 00:44:16.490 ⇒ 00:44:19.980 Pranav Narahari: turn off? Did they just ask more questions? Like…
489 00:44:19.980 ⇒ 00:44:32.530 Amber Lin: Like, if it’s too much of a technical term, it just doesn’t make sense, and then, they… it might cause them to ignore other important details you have in the same sentence.
490 00:44:32.780 ⇒ 00:44:36.269 Amber Lin: So, in order to keep their attention,
491 00:44:36.600 ⇒ 00:44:55.560 Amber Lin: like, be very, very, very simple and specific. They’ll have business questions if I want to see this as a goal I want for my department, but on the technical execution, just keep that, internal, unless they ask, and then we have to explain it in very, like, simple terms.
492 00:44:55.910 ⇒ 00:44:56.680 Pranav Narahari: Gotcha.
493 00:45:01.720 ⇒ 00:45:06.929 Amber Lin: Well, the benefit is the slide updates are really simple.
494 00:45:06.930 ⇒ 00:45:07.270 Pranav Narahari: Right.
495 00:45:07.270 ⇒ 00:45:14.370 Amber Lin: You don’t have to spend too much time here. And… but the downside is that you’ll have to push internally
496 00:45:14.610 ⇒ 00:45:26.350 Amber Lin: For these technical things to happen, because the clients are not going to specify how you would do it. They’re generally impressed by our technical abilities, but
497 00:45:26.760 ⇒ 00:45:31.480 Amber Lin: Unless it becomes a business… impact for them. They don’t really…
498 00:45:32.180 ⇒ 00:45:35.040 Amber Lin: It doesn’t render in their minds.
499 00:45:35.300 ⇒ 00:45:35.890 Amber Lin: So…
500 00:45:35.890 ⇒ 00:45:40.670 Pranav Narahari: Yeah, so I think that makes sense. In terms of, like…
501 00:45:41.450 ⇒ 00:45:45.859 Pranav Narahari: their primary concerns that I’ve noticed in the last couple meetings is…
502 00:45:46.110 ⇒ 00:45:51.279 Pranav Narahari: Usage is one, and then also just accuracy, of course.
503 00:45:51.280 ⇒ 00:45:52.190 Amber Lin: Yeah.
504 00:45:52.620 ⇒ 00:46:12.509 Amber Lin: Okay. Because their main concern, I think, understandably for Andy, is, is it really improving? Is it actually giving accurate responses and making our agent’s life better? Because they have been pushing a lot for usage, which I really thank them for, because that contributes to our…
505 00:46:12.620 ⇒ 00:46:25.979 Amber Lin: Like, that pushes her to a higher payment tier. So we’re already at Tier 2, but, like, the looming fear I have is that even if they use it more, it’s not really as helpful, or it’s never as accurate.
506 00:46:26.260 ⇒ 00:46:31.540 Amber Lin: So that’s their concern, too, of how do we know things are…
507 00:46:31.690 ⇒ 00:46:41.660 Amber Lin: accurate? How do we make sure that things are up to date, that they follow the update workflow to keep it up to date?
508 00:46:42.140 ⇒ 00:46:45.039 Amber Lin: That means we need to look at the triage questions.
509 00:46:45.220 ⇒ 00:46:50.530 Amber Lin: Like, I know we’re looking… we’re working on stuff of the migration to make…
510 00:46:51.380 ⇒ 00:47:04.689 Amber Lin: architecturally better, but I feel like, we’ve been missing the smaller steps of the day-to-day minor improvements. We’ve not been able to give time to that.
511 00:47:04.690 ⇒ 00:47:05.280 Pranav Narahari: Gotcha.
512 00:47:05.660 ⇒ 00:47:21.100 Amber Lin: And without those, like, Andy could be on better architecture, but it will not have the data to give correct responses. So that’s the same with the zip code databases as well. So, a lot of manual work on this client.
513 00:47:21.380 ⇒ 00:47:29.320 Amber Lin: And ideally, after we migrate, we want to, like, automate these manual improvements to have
514 00:47:29.910 ⇒ 00:47:32.660 Amber Lin: A good feedback flow.
515 00:47:32.810 ⇒ 00:47:37.430 Amber Lin: So I would love to work on that with you.
516 00:47:38.160 ⇒ 00:47:38.940 Pranav Narahari: Gotcha.
517 00:47:39.400 ⇒ 00:47:39.910 Amber Lin: Yeah.
518 00:47:39.910 ⇒ 00:47:41.250 Pranav Narahari: with,
519 00:47:41.990 ⇒ 00:47:48.789 Pranav Narahari: what was I going to? Oh, yeah, so for these, the usage tiers, right, where can I read about that?
520 00:47:49.110 ⇒ 00:48:07.979 Amber Lin: Good question. The only thing I remember is there’s… that’s the first contract, this expansion contract we signed, but first tier is, like, up to 2K, second tier tier is 2K to 5K, so we’ll be in this tier for a while.
521 00:48:08.930 ⇒ 00:48:11.040 Pranav Narahari: 2K to 5K… what?
522 00:48:11.040 ⇒ 00:48:13.270 Amber Lin: Per month. Usage is per month.
523 00:48:15.320 ⇒ 00:48:19.240 Pranav Narahari: So… 2K to 5K requests.
524 00:48:19.820 ⇒ 00:48:36.499 Amber Lin: You see the slide on the usage one? So, exchanges. So, the dashboard that I show, if we have… the current one I show is weekly, but for monthly, if we were able to surpass 5K,
525 00:48:36.660 ⇒ 00:48:40.090 Amber Lin: Then we’re at another higher tier, but…
526 00:48:40.800 ⇒ 00:48:46.489 Amber Lin: Let me see if we can find the SOP, or the SOW we signed.
527 00:48:47.570 ⇒ 00:48:51.370 Pranav Narahari: Okay. Yeah, I’m also trying to see, like, where is.
528 00:48:51.370 ⇒ 00:48:53.570 Amber Lin: Where are you seeing 2K?
529 00:48:53.790 ⇒ 00:48:57.769 Pranav Narahari: Because the total exchange is, say, 627, right?
530 00:48:57.770 ⇒ 00:49:00.610 Amber Lin: Yeah, that’s for weekly, so.
531 00:49:00.610 ⇒ 00:49:01.090 Pranav Narahari: Gotcha.
532 00:49:01.090 ⇒ 00:49:08.009 Amber Lin: walk you through how to navigate the rail. Are you able to… did you get access?
533 00:49:08.210 ⇒ 00:49:09.779 Pranav Narahari: Yeah.
534 00:49:10.330 ⇒ 00:49:10.810 Amber Lin: Awesome.
535 00:49:10.810 ⇒ 00:49:14.210 Pranav Narahari: I think Rico already gave me access, and also… Yeah. Yeah, so I’m in.
536 00:49:14.210 ⇒ 00:49:20.569 Amber Lin: Yeah, so this is… this is where we start. It’s the Conversation Log Metrics Dashboard.
537 00:49:21.210 ⇒ 00:49:32.890 Amber Lin: And I usually go into pivot so we can… to see it in a tabular format. Up here is where you can adjust the… the date range. So, if we say…
538 00:49:33.280 ⇒ 00:49:38.319 Amber Lin: I think I wanna look at… What did last month.
539 00:49:38.920 ⇒ 00:49:46.620 Amber Lin: Anyways, so I’ll do last month, so that’s… 1 to 28…
540 00:49:47.120 ⇒ 00:49:49.350 Amber Lin: Apply, and then you can see this.
541 00:49:49.520 ⇒ 00:49:54.730 Amber Lin: In here, there’s a bookmark view. I don’t know if this gives it…
542 00:49:55.520 ⇒ 00:50:06.869 Amber Lin: if you can see this either, but you can create bookmarks, and I’ll click that so we can… I did past 6 months, monthly usage, you can drag these up here.
543 00:50:07.130 ⇒ 00:50:13.819 Amber Lin: Yeah. So, you can see, starting from December, we’ve been over 2K.
544 00:50:14.400 ⇒ 00:50:18.360 Amber Lin: So that’s going to get over next here, yeah.
545 00:50:18.360 ⇒ 00:50:25.669 Pranav Narahari: Gotcha. And so this is all, like, production exchanges from… from them, right? Like.
546 00:50:25.670 ⇒ 00:50:28.650 Amber Lin: Yeah, so I excluded the…
547 00:50:29.210 ⇒ 00:50:42.780 Amber Lin: test users, but I think occasionally, like, there’s one usage from Sam or one usage from Mustafa that kind of gets leaked through, so sometimes you have to manually filter them out.
548 00:50:43.170 ⇒ 00:50:43.800 Pranav Narahari: Yep.
549 00:50:43.800 ⇒ 00:50:46.149 Amber Lin: But, like, usually it’s fine.
550 00:50:46.820 ⇒ 00:50:47.890 Pranav Narahari: Okay, cool.
551 00:50:48.020 ⇒ 00:50:49.070 Pranav Narahari: Yeah.
552 00:50:49.550 ⇒ 00:50:52.570 Pranav Narahari: Yeah, that would be cool just for me to see, like…
553 00:50:53.200 ⇒ 00:50:58.659 Pranav Narahari: have numbers to know, like, what’s an SOW, like, once we get to that 5K mark?
554 00:50:58.660 ⇒ 00:51:02.189 Amber Lin: Okay, let me… let me check.
555 00:51:06.280 ⇒ 00:51:09.170 Pranav Narahari: I wonder if it’s in, I know a lot of this stuff is in…
556 00:51:09.170 ⇒ 00:51:12.190 Amber Lin: Maybe it’s in sales?
557 00:51:12.350 ⇒ 00:51:14.940 Amber Lin: Oh, that’s true. That’s true.
558 00:51:15.860 ⇒ 00:51:16.659 Pranav Narahari: I can take a look there.
559 00:51:17.080 ⇒ 00:51:17.940 Amber Lin: Okay.
560 00:51:18.290 ⇒ 00:51:23.739 Amber Lin: ABC Discovery… nope, it’s not there, maybe in Sales.
561 00:51:25.930 ⇒ 00:51:27.400 Amber Lin: Okay, I don’t know.
562 00:51:27.600 ⇒ 00:51:28.210 Amber Lin: I don’t know.
563 00:51:28.210 ⇒ 00:51:29.920 Pranav Narahari: All good, yeah, I’ll.
564 00:51:29.920 ⇒ 00:51:30.570 Amber Lin: possible.
565 00:51:31.020 ⇒ 00:51:32.020 Pranav Narahari: I’ll figure that out.
566 00:51:32.280 ⇒ 00:51:43.649 Amber Lin: All right, yeah, and if it’s helpful on the slides, as a last note, add images of what we’re doing. Right now, I’m… I don’t have much. I usually ask for them.
567 00:51:43.730 ⇒ 00:52:01.049 Amber Lin: on Thursday, I asked Mustafa, Casey, can you give me screenshots for this particular thing? And then that helps… that helps, drive discussion, because they’ll see an image, and then they will help discuss, which, like, discussion is how… how I’ve been running these…
568 00:52:01.190 ⇒ 00:52:14.239 Amber Lin: these updates to get them more involved. So… but because it’s on Thursday noon, we don’t always have something ready by Wednesday, so I usually ask on Thursday, I hate giving you a screenshot.
569 00:52:15.130 ⇒ 00:52:16.860 Pranav Narahari: Yeah, gotcha, that makes sense.
570 00:52:17.030 ⇒ 00:52:24.630 Amber Lin: Okay, cool. Yeah, that’s all from me. I’ll still run this this week, but, next week, you can start running it.
571 00:52:25.470 ⇒ 00:52:38.180 Pranav Narahari: Yeah, sounds good. So I’m in my CSO call in a couple hours, like, Utsama kind of wants to do a dry run, so, I mean, I’m going to be presenting this there.
572 00:52:39.210 ⇒ 00:52:47.279 Pranav Narahari: getting, like, just kind of, like, feedback on… from them. Probably just from Bhutan, because he knows the client the best. So… Yeah.
573 00:52:47.610 ⇒ 00:52:50.639 Pranav Narahari: Yeah, even tomorrow, we can talk about how we can…
574 00:52:51.030 ⇒ 00:52:55.699 Pranav Narahari: like, tag-team it or something, like, I can just kind of do the initial pass of, like, each of…
575 00:52:56.080 ⇒ 00:53:04.540 Pranav Narahari: Or, honestly, it maybe makes sense for you to just do it this week, because I think what I also want to do is just set up individual calls with,
576 00:53:04.740 ⇒ 00:53:06.790 Pranav Narahari: Yvette, and .
577 00:53:08.540 ⇒ 00:53:09.430 Amber Lin: Denise.
578 00:53:09.580 ⇒ 00:53:18.050 Pranav Narahari: Engine, yeah, sorry. So I’m gonna set up calls with them, just to kind of introduce myself, talk a little bit more, just so, I can get a better picture of just how…
579 00:53:18.220 ⇒ 00:53:20.430 Pranav Narahari: They want things run as well.
580 00:53:20.430 ⇒ 00:53:30.920 Amber Lin: Yeah. I invited you to next Monday’s call, but I think an introduction this Thursday would be great. They already know you, but, like, and then set up calls and all that.
581 00:53:30.920 ⇒ 00:53:34.780 Pranav Narahari: I’ll just probably do, like, some one-on-one calls with them. Yeah, probably, like, tomorrow or Friday.
582 00:53:35.270 ⇒ 00:53:35.920 Amber Lin: Okay.
583 00:53:36.840 ⇒ 00:53:38.090 Amber Lin: Sounds good.
584 00:53:38.760 ⇒ 00:53:39.590 Pranav Narahari: Yeah.
585 00:53:40.300 ⇒ 00:53:45.160 Pranav Narahari: I think transition-wise, I feel pretty good about how things are moving now. Anything…
586 00:53:45.820 ⇒ 00:53:48.140 Pranav Narahari: Yeah, let me know, though, because, like…
587 00:53:48.140 ⇒ 00:53:49.500 Amber Lin: concerns, yeah.
588 00:53:51.010 ⇒ 00:53:58.640 Amber Lin: Cool. There’s just a lot of different work streams, and it’s not the best documented, because I didn’t have EP time, I only had CSO time.
589 00:53:58.720 ⇒ 00:54:07.410 Pranav Narahari: Cool. Yeah, no, makes sense. I mean, like, Casey’s been super helpful, just kind of, like, walking me through, like, the different workflows from a technical perspective, too.
590 00:54:08.420 ⇒ 00:54:08.970 Amber Lin: Okay.
591 00:54:08.970 ⇒ 00:54:15.709 Pranav Narahari: So, yeah, that should definitely… that definitely helped me out a lot yesterday. So, yeah, I’m feeling pretty good, too. Awesome.
592 00:54:16.100 ⇒ 00:54:23.509 Amber Lin: Alright, I don’t think I’ll be there at the CSO meeting, but do you feel like you have enough to do the dry run there?
593 00:54:23.950 ⇒ 00:54:27.839 Pranav Narahari: Yeah, definitely. I don’t think it’s gonna be, like, super formal. It’s not a big deal.
594 00:54:28.060 ⇒ 00:54:32.870 Amber Lin: Okay, sounds good. I canceled our meeting later, so that’ll be it for today.
595 00:54:33.430 ⇒ 00:54:34.600 Pranav Narahari: Cool. Alright, thanks, Amber.
596 00:54:34.600 ⇒ 00:54:36.919 Amber Lin: Awesome. Thanks, great, bye.
597 00:54:37.350 ⇒ 00:54:37.900 Pranav Narahari: See ya.