Meeting Title: AI Text-to-SQL Spike Review Date: 2025-11-21 Meeting participants: Awaish Kumar, Casie Aviles, Samuel Roberts, Ashwini Sharma, Uttam Kumaran
WEBVTT
1 00:01:16.210 ⇒ 00:01:16.970 Samuel Roberts: Hey. Hi.
2 00:01:16.970 ⇒ 00:01:17.570 Casie Aviles: son.
3 00:01:20.300 ⇒ 00:01:21.209 Samuel Roberts: How’s it going?
4 00:01:23.850 ⇒ 00:01:27.519 Casie Aviles: Yeah, I think it’s doing good right now.
5 00:01:28.720 ⇒ 00:01:30.900 Casie Aviles: I was kind of making some last-minute
6 00:01:31.420 ⇒ 00:01:35.059 Casie Aviles: Changes here to the dock, but…
7 00:01:37.020 ⇒ 00:01:41.009 Casie Aviles: Yeah, I think this is mostly what I have so far.
8 00:01:41.320 ⇒ 00:01:44.219 Casie Aviles: Based on, like, the multiple tools.
9 00:01:45.490 ⇒ 00:01:50.270 Casie Aviles: Yeah, I can share as well the chat if you guys want to.
10 00:01:52.550 ⇒ 00:01:53.190 Casie Aviles: I’ll quickly…
11 00:01:53.190 ⇒ 00:01:54.220 Samuel Roberts: The Notion Dock?
12 00:01:55.020 ⇒ 00:01:57.510 Casie Aviles: Yes, I’ll also share human in this chat.
13 00:01:57.510 ⇒ 00:01:58.310 Samuel Roberts: Okay.
14 00:01:59.470 ⇒ 00:02:00.600 Samuel Roberts: Yeah, I got it.
15 00:02:03.190 ⇒ 00:02:06.409 Casie Aviles: Okay, yeah, I’ll do… let’s just wait.
16 00:02:06.510 ⇒ 00:02:10.480 Casie Aviles: A few more minutes, and see if anyone else will join.
17 00:02:10.900 ⇒ 00:02:11.630 Samuel Roberts: Totally.
18 00:03:46.450 ⇒ 00:03:50.759 Casie Aviles: Just send a quick note to the channel, if anyone else will join.
19 00:03:52.100 ⇒ 00:03:52.720 Samuel Roberts: Okay.
20 00:04:21.870 ⇒ 00:04:23.269 Casie Aviles: Oh, hi, Ashrin.
21 00:04:24.400 ⇒ 00:04:25.150 Ashwini Sharma: So…
22 00:04:28.660 ⇒ 00:04:35.139 Casie Aviles: Yeah, thank you for… For coming and also taking a look at the doc.
23 00:04:37.450 ⇒ 00:04:38.010 Casie Aviles: So I think…
24 00:04:38.010 ⇒ 00:04:38.560 Ashwini Sharma: Welcome.
25 00:04:39.080 ⇒ 00:04:41.329 Casie Aviles: Yeah, I think what you guys will…
26 00:04:41.580 ⇒ 00:04:43.869 Casie Aviles: Be very helpful with this, like…
27 00:04:44.000 ⇒ 00:04:47.949 Casie Aviles: More on the data side, since, admittedly, that’s, like.
28 00:04:48.780 ⇒ 00:04:51.399 Casie Aviles: also what I… I think what I need.
29 00:04:51.630 ⇒ 00:04:54.350 Casie Aviles: Your expertise on… Huh.
30 00:04:54.610 ⇒ 00:04:56.549 Casie Aviles: I was mostly looking at it
31 00:04:56.660 ⇒ 00:05:02.059 Casie Aviles: from an AI side, so it’d be great to hear, like, feedback on that as well.
32 00:05:09.440 ⇒ 00:05:12.910 Casie Aviles: Okay, C.
33 00:05:23.130 ⇒ 00:05:30.520 Casie Aviles: Alright, I’ll just go ahead and share my… Screen right now, and…
34 00:05:34.440 ⇒ 00:05:37.999 Casie Aviles: Okay, so can you guys see the Notion document right now?
35 00:05:39.270 ⇒ 00:05:39.800 Samuel Roberts: Yep.
36 00:05:41.750 ⇒ 00:05:42.500 Casie Aviles: Okay.
37 00:05:44.920 ⇒ 00:05:48.060 Casie Aviles: So, I guess I’ll start by just…
38 00:05:48.610 ⇒ 00:05:51.690 Casie Aviles: Prefacing, like, what’s, like, the goal here?
39 00:05:52.390 ⇒ 00:05:57.290 Casie Aviles: And then kind of give you, like, a quick overview of what I found.
40 00:05:58.140 ⇒ 00:06:00.089 Casie Aviles: What are the recommendations?
41 00:06:00.780 ⇒ 00:06:05.539 Casie Aviles: And… I guess… After that, I would love some…
42 00:06:06.020 ⇒ 00:06:14.240 Casie Aviles: feedback, some gaps that can be identified, and then if there are any action items that I can take.
43 00:06:14.880 ⇒ 00:06:21.129 Casie Aviles: So… Yeah, let’s… I’m going to start here with the goal and the requirements, so…
44 00:06:22.900 ⇒ 00:06:28.890 Casie Aviles: This is… this is… this spike is primarily for identifying, like, a viable…
45 00:06:29.480 ⇒ 00:06:41.549 Casie Aviles: way that we could, you know, bridge, like, natural language queries into SQL queries. And we’re trying to use, we’re looking at several AI tools.
46 00:06:41.810 ⇒ 00:06:45.480 Casie Aviles: Oh, I mean, platforms that let us… basically ask.
47 00:06:45.910 ⇒ 00:06:50.540 Casie Aviles: These kinds of questions, and have… A better…
48 00:06:50.650 ⇒ 00:06:55.610 Casie Aviles: Or be… and be able to query, like, the database with
49 00:06:55.760 ⇒ 00:07:00.240 Casie Aviles: Yeah, with lateral language queries, so… Weird.
50 00:07:00.850 ⇒ 00:07:04.549 Casie Aviles: yeah, like, we’re all doing this for Eden,
51 00:07:04.790 ⇒ 00:07:12.650 Casie Aviles: But also, ideally, like, across other clients and our internal, you know, our data team.
52 00:07:12.930 ⇒ 00:07:18.840 Casie Aviles: can also… From this bike, be able to, like, identify which tools would be the best.
53 00:07:19.280 ⇒ 00:07:20.710 Casie Aviles: That they could use.
54 00:07:20.900 ⇒ 00:07:27.770 Casie Aviles: So I know we’ve had a lot of attempts, and some attempts in the past, for, with…
55 00:07:27.910 ⇒ 00:07:31.100 Casie Aviles: When it comes to, like, looking for, like, tools that
56 00:07:31.380 ⇒ 00:07:40.649 Casie Aviles: we can use for text-to-SQL, so… I’m also going to briefly… Discussed that, and…
57 00:07:41.100 ⇒ 00:07:47.900 Casie Aviles: Yeah, I built… I basically built on top of what we had before, so… Yeah, so…
58 00:07:49.230 ⇒ 00:07:56.860 Casie Aviles: So since this is for Eden, like, one of the requirements that we wanted to have is BigQuery support, since they have, like, data there.
59 00:07:57.360 ⇒ 00:08:05.860 Casie Aviles: But ideally, we should be able to, like, support different sources. So, we would have Snowflake, BigQuery, Mother Duck.
60 00:08:06.050 ⇒ 00:08:11.600 Casie Aviles: for example, And even if, you know, like.
61 00:08:11.850 ⇒ 00:08:15.470 Casie Aviles: custom CSV uploads, that would be, also good.
62 00:08:17.600 ⇒ 00:08:24.259 Casie Aviles: Yeah, and then some example queries that I have over here that… We can try…
63 00:08:24.480 ⇒ 00:08:29.319 Casie Aviles: Like, we could test with would be, you know, being able to calculate revenue.
64 00:08:30.940 ⇒ 00:08:34.170 Casie Aviles: And so on. Like, these are just some of the examples to…
65 00:08:34.780 ⇒ 00:08:41.439 Casie Aviles: And in terms of, like, how we can interface with these tools, or how, ideally, we could have, like.
66 00:08:41.650 ⇒ 00:08:43.650 Casie Aviles: Or, like, these are just nice-to-haves.
67 00:08:43.929 ⇒ 00:08:49.409 Casie Aviles: That we’re able to… access it from Slap.
68 00:08:49.990 ⇒ 00:08:54.949 Casie Aviles: So some of these tools allow for Slack integrations, some of them no.
69 00:08:55.730 ⇒ 00:08:59.180 Casie Aviles: So far, like, I couldn’t find…
70 00:08:59.570 ⇒ 00:09:03.389 Casie Aviles: tool that… that also lets us use it in cursor.
71 00:09:03.920 ⇒ 00:09:05.030 Casie Aviles: At the moment.
72 00:09:07.400 ⇒ 00:09:14.929 Casie Aviles: So… yeah, and, like, another here, another one of the points here, ability to build on a response, so…
73 00:09:15.140 ⇒ 00:09:19.520 Casie Aviles: I guess being able to have, like, a conversational like, a thread.
74 00:09:19.730 ⇒ 00:09:23.190 Casie Aviles: And be able to, like, ask follow-up questions.
75 00:09:24.780 ⇒ 00:09:28.249 Casie Aviles: Those are some nice-to-haves, and also being able to, like.
76 00:09:29.030 ⇒ 00:09:37.729 Casie Aviles: Define a system prompt or tools, templates, include definitions. That would be, like, being able to bring in external context.
77 00:09:38.530 ⇒ 00:09:40.670 Casie Aviles: Which would be helpful.
78 00:09:41.120 ⇒ 00:09:51.039 Casie Aviles: So, yeah, I think… yeah, I’ll pause there for now. If… let me know, guys, if it’s clear, and if I can proceed, or if you have any questions so far.
79 00:09:55.880 ⇒ 00:09:57.880 Samuel Roberts: I guess my… first…
80 00:09:58.000 ⇒ 00:10:06.149 Samuel Roberts: biggest question is how… like, I saw Uten mention that, like, multiple types of sources or context, but also nice to have, like, how…
81 00:10:07.090 ⇒ 00:10:10.999 Samuel Roberts: big a deal is it to have that now versus just getting BigQuery?
82 00:10:12.820 ⇒ 00:10:19.550 Casie Aviles: Oh, yeah, I think for now, like, the… the biggest…
83 00:10:19.810 ⇒ 00:10:25.260 Casie Aviles: like, the main goal would be BigQuery, but… Yeah,
84 00:10:25.480 ⇒ 00:10:32.590 Casie Aviles: I think it’s also nice that we have multiple warehouses, since I believe we’re using also Mother Doc and Snowflake.
85 00:10:32.730 ⇒ 00:10:33.620 Casie Aviles: Internally.
86 00:10:33.620 ⇒ 00:10:34.400 Samuel Roberts: Yeah.
87 00:10:34.820 ⇒ 00:10:37.929 Casie Aviles: Yeah, so I think… It’s a huge plus, yeah.
88 00:10:38.160 ⇒ 00:10:43.180 Samuel Roberts: Cool, okay, I was just making sure, I wasn’t… like, I didn’t know how, pressing this is for Eden, kind of thing.
89 00:10:43.600 ⇒ 00:10:44.880 Casie Aviles: Okay, yeah.
90 00:10:45.240 ⇒ 00:10:45.840 Samuel Roberts: Okay.
91 00:10:47.760 ⇒ 00:10:48.480 Casie Aviles: Alright.
92 00:10:49.050 ⇒ 00:10:56.600 Casie Aviles: I’ll move on to, like, the tools we have so far. So I wasn’t able to test, like, two of these, but…
93 00:10:56.970 ⇒ 00:11:04.920 Casie Aviles: I… I was able to, like, test, Dagster on, you know, these, these 5 tools.
94 00:11:05.570 ⇒ 00:11:08.280 Casie Aviles: Omni and Bobby.
95 00:11:08.480 ⇒ 00:11:14.930 Casie Aviles: So, we also have… Custom solutions here as an option, but…
96 00:11:15.530 ⇒ 00:11:19.860 Casie Aviles: Yeah, this will have, like, its own…
97 00:11:20.010 ⇒ 00:11:26.470 Casie Aviles: challenges, I guess, and yeah, I’ll be talking about that in the summary table that I have, but…
98 00:11:26.780 ⇒ 00:11:31.950 Samuel Roberts: So far, these are, like, if you guys want to look at more detailed.
99 00:11:32.160 ⇒ 00:11:34.720 Casie Aviles: Items based on each tool that
100 00:11:35.000 ⇒ 00:11:37.529 Casie Aviles: I was able to find, so…
101 00:11:39.160 ⇒ 00:11:42.029 Casie Aviles: We could just breeze through quickly.
102 00:11:42.940 ⇒ 00:11:48.470 Casie Aviles: But, yeah, everything should be clear in the… clearer in the summary table that I have here.
103 00:11:50.720 ⇒ 00:11:58.180 Casie Aviles: Yeah, so… I think… And also, yeah, I’ll also go to the recommendations as well.
104 00:11:58.830 ⇒ 00:12:05.110 Casie Aviles: For now, like, I believe that the best tools that I’ve tried are Omni and Wabi.
105 00:12:05.900 ⇒ 00:12:10.630 Casie Aviles: So, compared to, like, the rest of the tools, the others are…
106 00:12:11.490 ⇒ 00:12:14.090 Casie Aviles: I would say quite early stage.
107 00:12:14.910 ⇒ 00:12:20.100 Casie Aviles: For example, Dagster Compass has Doesn’t have, very good…
108 00:12:21.530 ⇒ 00:12:24.970 Casie Aviles: like, I think it just released a week ago, so…
109 00:12:25.420 ⇒ 00:12:31.210 Casie Aviles: It’s Slack native. The good thing is we’re able to, like, connect different warehouses.
110 00:12:31.360 ⇒ 00:12:35.470 Casie Aviles: But it’s pretty much just… Slack-based, and…
111 00:12:35.910 ⇒ 00:12:40.880 Casie Aviles: There’s not a lot of other things we can do, and even here, like, the charts that it…
112 00:12:41.030 ⇒ 00:12:45.989 Casie Aviles: Create… it creates, or… still kind of…
113 00:12:47.100 ⇒ 00:12:48.729 Casie Aviles: Yeah, there can still be some…
114 00:12:48.920 ⇒ 00:12:55.099 Casie Aviles: it’s still pretty basic, and may contain inaccuracies, as it says here. So, yeah.
115 00:12:55.370 ⇒ 00:13:02.060 Casie Aviles: I would say, well, it’s interesting… it’s an interesting option, it’s still not, like…
116 00:13:03.130 ⇒ 00:13:07.830 Casie Aviles: At the, you know, it’s not, like, the most mature tool right now.
117 00:13:08.450 ⇒ 00:13:15.570 Casie Aviles: But what else? So, okay, so for, yeah, for contextual,
118 00:13:16.230 ⇒ 00:13:21.369 Casie Aviles: I believe that we… it was still in the beta phase, so that’s also something that
119 00:13:21.510 ⇒ 00:13:28.260 Casie Aviles: May not be the most… ideal right now, but they do conf… they did confirm that
120 00:13:29.020 ⇒ 00:13:34.300 Casie Aviles: they were able to set up BigQuery, and were able to connect with BigQuery, I mean.
121 00:13:34.450 ⇒ 00:13:39.660 Casie Aviles: And… Yeah, right now it’s, I wasn’t able to, like, test it yet.
122 00:13:41.050 ⇒ 00:13:50.879 Casie Aviles: But yeah, foc- focusing on… Wabi right now, I think… This one has pretty good… Results.
123 00:13:51.250 ⇒ 00:13:55.660 Casie Aviles: So, it was able to, like, support these warehouses already.
124 00:13:56.200 ⇒ 00:14:00.690 Casie Aviles: And I could quickly go through… Like, the interface right now.
125 00:14:01.340 ⇒ 00:14:06.430 Casie Aviles: So we have these data sources, we can upload CSPs.
126 00:14:07.850 ⇒ 00:14:10.370 Casie Aviles: We could have different data sources, so…
127 00:14:11.160 ⇒ 00:14:21.269 Casie Aviles: Yeah, we can have Mother Duck, Snowflake, Bakery, I think those are the ones we’re using right now, and then we can even do CSV uploads, so I think that’s pretty good.
128 00:14:21.730 ⇒ 00:14:24.320 Casie Aviles: And, for both Omni and Wabi.
129 00:14:24.670 ⇒ 00:14:30.340 Casie Aviles: We have, I think they released around 2023, so… We’re probably on…
130 00:14:31.260 ⇒ 00:14:34.080 Casie Aviles: The more mature side compared to the other tools.
131 00:14:36.240 ⇒ 00:14:43.600 Casie Aviles: Yeah, and then there’s… there’s, like, I think there’s more features that are here that are interesting.
132 00:14:46.060 ⇒ 00:14:51.000 Casie Aviles: Yeah, oh, so the conversation just, cleared, but…
133 00:14:51.430 ⇒ 00:14:55.410 Casie Aviles: What’s nice here, I think, is we’re able to, like, give instructions…
134 00:14:55.540 ⇒ 00:14:57.770 Casie Aviles: Here, with, like, just prompting it.
135 00:14:58.050 ⇒ 00:15:03.559 Casie Aviles: So… That’s important if we want to add, like, business context.
136 00:15:05.320 ⇒ 00:15:11.579 Casie Aviles: For other tools, there’s not really a straightforward way to do that, or there’s no way at all.
137 00:15:12.970 ⇒ 00:15:17.690 Casie Aviles: So I think that’s pretty good that we’re able to, like, add instructions here.
138 00:15:17.800 ⇒ 00:15:22.130 Casie Aviles: And consider, like, business contacts and…
139 00:15:22.430 ⇒ 00:15:25.410 Casie Aviles: There’s also, like, a semantic layer here.
140 00:15:25.900 ⇒ 00:15:30.399 Casie Aviles: that we could create, SQL templates and definitions over here.
141 00:15:31.540 ⇒ 00:15:35.549 Casie Aviles: Which I think… Can be useful as well.
142 00:15:37.930 ⇒ 00:15:47.740 Casie Aviles: Then we’re also able to have, like, models here, although I haven’t really tested the rest of these features extensively, just…
143 00:15:47.960 ⇒ 00:15:49.520 Casie Aviles: Mainly the agents.
144 00:15:50.050 ⇒ 00:15:54.500 Casie Aviles: And… Yeah, I like, for example, what I like about
145 00:15:56.420 ⇒ 00:15:58.370 Casie Aviles: Lobbyists were able to, like.
146 00:15:58.700 ⇒ 00:16:01.650 Casie Aviles: Perform multiple steps in the background, so it’s like…
147 00:16:02.370 ⇒ 00:16:04.700 Casie Aviles: Thinking about the steps, so it’s like…
148 00:16:05.620 ⇒ 00:16:10.489 Casie Aviles: You know, the reasoning models of… other, like, like…
149 00:16:10.610 ⇒ 00:16:15.569 Casie Aviles: for OpenAI, you know, the reasoning model, so it thinks in steps.
150 00:16:16.480 ⇒ 00:16:21.699 Casie Aviles: We’re able to ask queries like this, and then it’s also available on Slack.
151 00:16:22.180 ⇒ 00:16:24.110 Casie Aviles: If we need to connect it.
152 00:16:24.910 ⇒ 00:16:29.669 Casie Aviles: And then we have, like… Yeah, over here it also generates, like.
153 00:16:30.410 ⇒ 00:16:35.170 Casie Aviles: charts that… for more visualization, so I think that’s good.
154 00:16:37.720 ⇒ 00:16:43.699 Casie Aviles: And then for… Yeah, let’s see, for Omni, I think that’s also,
155 00:16:43.900 ⇒ 00:16:46.620 Casie Aviles: They’re also kind of similar in that
156 00:16:47.480 ⇒ 00:16:53.120 Casie Aviles: It’s easier… it’s very easy to just ask human queries here, and then we have, like.
157 00:16:53.580 ⇒ 00:16:56.040 Casie Aviles: We can generate, like, charts as well.
158 00:16:56.510 ⇒ 00:17:03.870 Casie Aviles: And we can even see… The queries that are generated, and they’re just,
159 00:17:04.310 ⇒ 00:17:10.219 Casie Aviles: a lot of, op, features and configurations that we can have. We can,
160 00:17:11.550 ⇒ 00:17:17.640 Casie Aviles: Yeah, that we have over here, so… I think. And then we also have other connectors.
161 00:17:18.190 ⇒ 00:17:20.999 Casie Aviles: Based on the documentation.
162 00:17:22.650 ⇒ 00:17:25.599 Casie Aviles: We can connect to… these.
163 00:17:25.790 ⇒ 00:17:32.080 Casie Aviles: So, bakery, and then a lot of the major ones as well that we have, so Snowflake, Mother Duck, primarily.
164 00:17:32.280 ⇒ 00:17:35.250 Casie Aviles: And the rest, so… I think that’s good.
165 00:17:36.510 ⇒ 00:17:38.020 Casie Aviles: Oops.
166 00:17:40.000 ⇒ 00:17:49.729 Casie Aviles: Okay, yeah, I think that’s, that’s for, like, the recommendations. The other tools that we have, for example, Bond, I was talking to Bond.
167 00:17:50.300 ⇒ 00:17:53.569 Casie Aviles: Yeah, they didn’t have, like.
168 00:17:54.900 ⇒ 00:18:03.910 Casie Aviles: we’re not able to connect any warehouses yet, so that’s still something that’s coming up. So I would say that we’re going to skip that.
169 00:18:04.380 ⇒ 00:18:06.499 Casie Aviles: I’m not sure who among…
170 00:18:06.710 ⇒ 00:18:08.449 Casie Aviles: The team was able to, like.
171 00:18:09.090 ⇒ 00:18:11.239 Casie Aviles: I have a call with Bond already, right?
172 00:18:13.030 ⇒ 00:18:15.060 Casie Aviles: Yeah, I think, it’s not…
173 00:18:15.960 ⇒ 00:18:21.820 Casie Aviles: Yeah, we’re still not able to connect to warehouses, so it’s… we can skip this for now.
174 00:18:22.740 ⇒ 00:18:29.179 Casie Aviles: And it’s only, like, Shopify, so… and Eden, particularly, doesn’t really have any Shopify, so…
175 00:18:30.640 ⇒ 00:18:34.300 Casie Aviles: And then I guess lastly, there’s also just the…
176 00:18:34.940 ⇒ 00:18:39.700 Casie Aviles: A custom agent that we can… this is also an option that we can do.
177 00:18:39.940 ⇒ 00:18:47.200 Casie Aviles: Wherein, basically, we have… We have a… it’s kind of similar to what…
178 00:18:49.310 ⇒ 00:18:55.409 Casie Aviles: I think what Ashwini was mentioning, where we could connect Gemini.
179 00:18:56.060 ⇒ 00:18:58.389 Casie Aviles: I think what we could do there is…
180 00:18:59.000 ⇒ 00:19:04.230 Casie Aviles: We can use a framework like LachChain, and… be able to…
181 00:19:05.920 ⇒ 00:19:10.950 Casie Aviles: be able to, like… so this is something we did for ABC as well, where
182 00:19:11.220 ⇒ 00:19:13.559 Casie Aviles: We have a SQL agent that
183 00:19:13.840 ⇒ 00:19:19.989 Casie Aviles: translates, human, text, human questions into SQL queries, so…
184 00:19:20.800 ⇒ 00:19:28.120 Casie Aviles: I think the only thing there is it’s gonna take a lot of, more upfront work
185 00:19:28.540 ⇒ 00:19:32.190 Casie Aviles: We’re gonna have to, you know, manage the infra.
186 00:19:33.390 ⇒ 00:19:37.059 Casie Aviles: The model is also going to be from ours.
187 00:19:37.670 ⇒ 00:19:44.600 Casie Aviles: But yeah, like, the only benefit… the benefit there is we have a lot more control, definitely, we can…
188 00:19:45.010 ⇒ 00:19:51.420 Casie Aviles: Just… Include whatever data that we want, or, like, additional context, prompts.
189 00:19:51.930 ⇒ 00:19:54.739 Casie Aviles: So, the way it works is it also uses,
190 00:19:55.560 ⇒ 00:20:03.190 Casie Aviles: prompt, so I can actually just quickly show how that actually looks on Mint Mill, just to give you guys an idea.
191 00:20:10.480 ⇒ 00:20:12.690 Casie Aviles: But, yeah, it’s, it’s, I think…
192 00:20:13.120 ⇒ 00:20:16.930 Casie Aviles: And… but the problem there is, yeah, there’s a lot more work,
193 00:20:17.090 ⇒ 00:20:18.909 Casie Aviles: We’re gonna have to build, like.
194 00:20:19.300 ⇒ 00:20:24.880 Casie Aviles: a UI for that. Right now, it’s just a webhook-based, so it’s getting…
195 00:20:25.110 ⇒ 00:20:30.599 Casie Aviles: A message, and then it’s spitting out the results of the query back.
196 00:20:30.720 ⇒ 00:20:34.589 Casie Aviles: And visualization is gonna be, I guess, trickier.
197 00:20:35.330 ⇒ 00:20:41.399 Casie Aviles: for our current… version of that, we don’t really have, like, the ability to generate charts.
198 00:20:44.390 ⇒ 00:20:50.860 Casie Aviles: So, yeah, it’s this tool that we have over here that we’re using for ABC.
199 00:20:52.450 ⇒ 00:20:57.319 Casie Aviles: Yeah, so we’re basically just instantiating our…
200 00:20:57.930 ⇒ 00:21:02.590 Casie Aviles: LLM here, and then we have, like, a system message, so it’s a pretty…
201 00:21:03.230 ⇒ 00:21:11.520 Casie Aviles: it’s a lot of, prompting here that I would do in order to, like, handle specific queries.
202 00:21:13.200 ⇒ 00:21:18.269 Casie Aviles: We’re also able to, like, add the schemas to the context using this method.
203 00:21:18.710 ⇒ 00:21:25.330 Casie Aviles: But yeah, this is… But yeah, it’s not definitely as robust as…
204 00:21:26.230 ⇒ 00:21:31.310 Casie Aviles: as the other, you know, the other tools that I’ve shown earlier.
205 00:21:31.960 ⇒ 00:21:35.210 Casie Aviles: So, maybe it’s more of a fallback.
206 00:21:35.870 ⇒ 00:21:37.069 Casie Aviles: Pop Shenya.
207 00:21:38.040 ⇒ 00:21:41.460 Casie Aviles: So… I think, yeah, that’s… that’s pretty much…
208 00:21:42.190 ⇒ 00:21:47.390 Casie Aviles: Those are pretty much my findings for… This spike,
209 00:21:48.080 ⇒ 00:21:51.820 Casie Aviles: Yeah, I would say that’s it, and, you know, if there are
210 00:21:52.680 ⇒ 00:21:55.439 Casie Aviles: Yeah, if you guys have any immediate thoughts, yeah.
211 00:21:55.650 ⇒ 00:21:57.100 Casie Aviles: Let me know, Anne.
212 00:21:58.600 ⇒ 00:22:02.460 Awaish Kumar: Yeah, I only have one recommendation here.
213 00:22:02.720 ⇒ 00:22:07.099 Awaish Kumar: I would love to see the… like, the…
214 00:22:07.510 ⇒ 00:22:10.690 Awaish Kumar: what’d you say, kind of MVP for…
215 00:22:10.930 ⇒ 00:22:16.880 Awaish Kumar: Like, take a small task or a project, And applied using…
216 00:22:17.050 ⇒ 00:22:20.999 Awaish Kumar: Two, two, three tools which you are suggesting.
217 00:22:21.800 ⇒ 00:22:28.630 Awaish Kumar: And then we can basically see the results, we can see the… the effort
218 00:22:28.840 ⇒ 00:22:33.960 Awaish Kumar: The time taken to reach to that point in individual tools?
219 00:22:35.050 ⇒ 00:22:37.180 Awaish Kumar: Engineering effort, everything.
220 00:22:37.740 ⇒ 00:22:41.319 Awaish Kumar: And the cost. And then we can make a decision.
221 00:22:43.320 ⇒ 00:22:44.110 Casie Aviles: Okay.
222 00:22:44.240 ⇒ 00:22:45.040 Casie Aviles: Yeah.
223 00:22:46.070 ⇒ 00:22:54.040 Awaish Kumar: I have a few… tools involved, like, one of the tools is called DataGPT,
224 00:22:54.800 ⇒ 00:22:59.269 Awaish Kumar: That is, like, requires… also requires upfront,
225 00:22:59.980 ⇒ 00:23:06.500 Awaish Kumar: Efforts, like, we might have to set it up, do a few things.
226 00:23:06.710 ⇒ 00:23:09.839 Awaish Kumar: To actually converse using that tool.
227 00:23:12.670 ⇒ 00:23:17.609 Awaish Kumar: You can include that in your… spike, and…
228 00:23:18.380 ⇒ 00:23:24.279 Awaish Kumar: Another one which I have been, looking at is Curio.
229 00:23:25.080 ⇒ 00:23:31.660 Awaish Kumar: That is also… Like, kind of tools where you actually
230 00:23:32.660 ⇒ 00:23:39.860 Awaish Kumar: have to meet with them to set it up, right? So, basically, these are two things. So, like…
231 00:23:40.180 ⇒ 00:23:51.300 Awaish Kumar: So I just want to see how much effort we have to put in to basically use those tools, right? I know there…
232 00:23:51.300 ⇒ 00:24:06.670 Awaish Kumar: there is some efforts required, but how much it is, and how much they are going to support, and how much it is going to cost compared to other tools, if I… we get answer to these… those three questions, along with some, like, a sample
233 00:24:07.740 ⇒ 00:24:10.580 Awaish Kumar: Project, or, like, small task.
234 00:24:11.080 ⇒ 00:24:12.360 Awaish Kumar: As a demo?
235 00:24:12.870 ⇒ 00:24:15.560 Awaish Kumar: It would be easier to make a decision.
236 00:24:17.790 ⇒ 00:24:23.839 Casie Aviles: Okay, yeah, right, pricing and data as well, okay.
237 00:24:24.290 ⇒ 00:24:31.510 Casie Aviles: So, for pricing, there are some here that I already have an idea of how much it will cost, but…
238 00:24:32.190 ⇒ 00:24:38.080 Casie Aviles: There are also some that… Don’t have any explicit pricing based on…
239 00:24:38.180 ⇒ 00:24:42.300 Casie Aviles: These are… these are the ones that I got from… their website.
240 00:24:44.070 ⇒ 00:24:51.820 Casie Aviles: Okay, so I think what, what, we can do there is… For the data source…
241 00:24:52.520 ⇒ 00:24:56.450 Casie Aviles: I can try to create one, so let’s see.
242 00:24:57.610 ⇒ 00:24:58.900 Casie Aviles: Next steps…
243 00:25:03.540 ⇒ 00:25:08.480 Casie Aviles: prepare data source… That’s good.
244 00:25:08.880 ⇒ 00:25:10.130 Casie Aviles: And then…
245 00:25:10.670 ⇒ 00:25:18.150 Casie Aviles: Do we want… do we also want to have, like, queries that we want to test across, the…
246 00:25:19.270 ⇒ 00:25:25.099 Awaish Kumar: Yeah, like, we can take, for example, there’s a table called Fact Transaction.
247 00:25:25.540 ⇒ 00:25:29.750 Awaish Kumar: And, in the, in the Eden project. That is your data source.
248 00:25:30.060 ⇒ 00:25:32.339 Awaish Kumar: So, take that as your data source.
249 00:25:32.450 ⇒ 00:25:43.470 Awaish Kumar: And, connect… try to connect it using multiple tools. And ask the questions related to revenue, week versus week, comparison of revenue.
250 00:25:44.010 ⇒ 00:25:48.309 Awaish Kumar: Ask about refunds or discounts, data.
251 00:25:48.860 ⇒ 00:25:51.720 Awaish Kumar: Oh, things like that, so…
252 00:25:58.240 ⇒ 00:25:58.940 Casie Aviles: Okay.
253 00:26:07.370 ⇒ 00:26:12.610 Casie Aviles: Yep, okay. Yeah, that’s… I’ll also work on that then, next.
254 00:26:15.030 ⇒ 00:26:20.009 Casie Aviles: So, just to go… I do have, like.
255 00:26:20.800 ⇒ 00:26:25.230 Casie Aviles: I believe I have, access to that BigQuery data, okay.
256 00:26:25.660 ⇒ 00:26:27.769 Casie Aviles: So, use that,
257 00:26:31.660 ⇒ 00:26:34.640 Casie Aviles: Yeah, I was using, like, a desk.
258 00:26:34.800 ⇒ 00:26:39.580 Casie Aviles: data, but… It’s just mostly AI-generated, so…
259 00:26:40.010 ⇒ 00:26:43.180 Awaish Kumar: Yeah, like, we can use the data, like.
260 00:26:44.550 ⇒ 00:26:47.320 Awaish Kumar: I think we can use the time data.
261 00:26:47.940 ⇒ 00:26:48.910 Awaish Kumar: Not sure.
262 00:26:52.080 ⇒ 00:26:53.100 Casie Aviles: Yeah, okay.
263 00:26:54.570 ⇒ 00:26:56.609 Awaish Kumar: Maybe use the order summary table.
264 00:26:56.750 ⇒ 00:27:01.649 Ashwini Sharma: Kathy, quick question on your… on your dummy data, demo data, right?
265 00:27:02.360 ⇒ 00:27:03.270 Casie Aviles: Yeah, yeah.
266 00:27:03.270 ⇒ 00:27:09.249 Awaish Kumar: Did you make some definitions? Like, how does the tool understand the relationship between.
267 00:27:09.250 ⇒ 00:27:11.270 Ashwini Sharma: Between the tables that you have described.
268 00:27:11.410 ⇒ 00:27:15.039 Ashwini Sharma: That you have given order, order lines, you know, and then one more.
269 00:27:15.390 ⇒ 00:27:17.880 Ashwini Sharma: table was there, I’m not able to recall right now.
270 00:27:18.560 ⇒ 00:27:23.130 Ashwini Sharma: But you had 3 datasets in your demo data, right?
271 00:27:23.890 ⇒ 00:27:30.850 Casie Aviles: Yes, I… yeah, I believe what we… although I didn’t really set it up, but…
272 00:27:31.810 ⇒ 00:27:35.629 Casie Aviles: I think it can… yeah, it’s able to, like, read the…
273 00:27:36.670 ⇒ 00:27:39.299 Casie Aviles: Like, schemas for the tables.
274 00:27:39.770 ⇒ 00:27:45.450 Casie Aviles: But ideally, we sh… it can also… we can also, like, inject
275 00:27:45.570 ⇒ 00:27:49.410 Casie Aviles: More context, and if we need to clarify how, how…
276 00:27:49.790 ⇒ 00:27:53.389 Casie Aviles: Thing, or how these tables, or what these tables are for.
277 00:27:53.960 ⇒ 00:27:58.290 Casie Aviles: Like, I believe there’s an option for that here.
278 00:27:58.700 ⇒ 00:28:00.620 Casie Aviles: Then that should be possible.
279 00:28:01.820 ⇒ 00:28:04.930 Casie Aviles: Lake Fear, for example, and then we also…
280 00:28:25.350 ⇒ 00:28:29.020 Ashwini Sharma: Yeah, I think I lost your voice, I’m not able to hear anything.
281 00:28:29.620 ⇒ 00:28:31.080 Ashwini Sharma: Maybe it’s just me?
282 00:28:33.460 ⇒ 00:28:35.650 Casie Aviles: Sorry, did I get cut off?
283 00:28:36.680 ⇒ 00:28:39.350 Ashwini Sharma: I wasn’t able to hear you, anything.
284 00:28:40.330 ⇒ 00:28:43.970 Casie Aviles: Oh… Sorry, can you hear me now?
285 00:28:44.220 ⇒ 00:28:44.920 Ashwini Sharma: Yes.
286 00:28:46.060 ⇒ 00:28:51.859 Casie Aviles: Yeah, yeah, I, it showed on my monitor that my internet was unstable, sorry.
287 00:28:52.740 ⇒ 00:28:57.479 Casie Aviles: Yeah, so I… yeah, what I was trying to say is that we’re able to, like.
288 00:28:58.060 ⇒ 00:29:03.790 Casie Aviles: Use, we were able to inject, custom, like, context into…
289 00:29:05.190 ⇒ 00:29:08.470 Casie Aviles: Into the, AI agents, and…
290 00:29:08.630 ⇒ 00:29:13.270 Casie Aviles: Also, it’s able to, like, read the schema, at least, for the sources.
291 00:29:13.720 ⇒ 00:29:14.520 Casie Aviles: But…
292 00:29:15.090 ⇒ 00:29:22.780 Casie Aviles: I really… I didn’t really test this out yet, like, the custom instructions, but that’s also an option that we can do.
293 00:29:22.910 ⇒ 00:29:27.389 Casie Aviles: In order to, like, provide a more comprehensive
294 00:29:28.020 ⇒ 00:29:32.310 Casie Aviles: I guess, understanding for the AI, and understand, like, what these tables are for.
295 00:29:35.690 ⇒ 00:29:39.389 Samuel Roberts: So the, like, next steps would be to test Omni…
296 00:29:39.700 ⇒ 00:29:40.990 Casie Aviles: wobble…
297 00:29:43.660 ⇒ 00:29:46.060 Samuel Roberts: Those are the two recommended ones, right?
298 00:29:47.130 ⇒ 00:29:48.370 Casie Aviles: Yes.
299 00:29:51.050 ⇒ 00:30:01.130 Samuel Roberts: Yeah, I played with Omni a little bit, for another proposal. They don’t have a Slack bot yet, but they had a demo of one that I was able to get the code for.
300 00:30:01.250 ⇒ 00:30:07.169 Samuel Roberts: And their API, you can hit it and generate the SQL from the query.
301 00:30:08.640 ⇒ 00:30:09.310 Casie Aviles: Hmm.
302 00:30:09.470 ⇒ 00:30:15.919 Samuel Roberts: I can try to share that with you at some point, but I know that when I asked for that, they said they were working on a production Slack bot as well.
303 00:30:16.840 ⇒ 00:30:18.990 Casie Aviles: Oh, okay, so, yeah, okay.
304 00:30:25.520 ⇒ 00:30:34.319 Casie Aviles: Alright, I think that’s all right, and I will just check in with you guys async, regarding the…
305 00:30:34.320 ⇒ 00:30:38.259 Uttam Kumaran: Hi guys, sorry. Are we, are we ending meeting?
306 00:30:40.990 ⇒ 00:30:46.320 Uttam Kumaran: But give me the TLDR. I left some notes, but I’m bummed, I’ll…
307 00:30:50.270 ⇒ 00:30:50.790 Samuel Roberts: Oh, hello?
308 00:30:51.570 ⇒ 00:30:53.459 Casie Aviles: Yeah, I think he got cut off.
309 00:30:53.890 ⇒ 00:30:55.339 Uttam Kumaran: Did I get cut off? Can you hear me?
310 00:30:55.340 ⇒ 00:30:56.070 Samuel Roberts: Little bit.
311 00:30:56.960 ⇒ 00:31:00.299 Uttam Kumaran: No, I said I’ll watch, I’ll rewatch the meeting, but…
312 00:31:01.010 ⇒ 00:31:01.390 Samuel Roberts: Okay.
313 00:31:04.910 ⇒ 00:31:11.090 Samuel Roberts: TLDR is Omni and Wobble look like the best bets to try testing out against some Eden data and see how
314 00:31:11.610 ⇒ 00:31:15.509 Samuel Roberts: The setup is, and… How well they performed.
315 00:31:16.040 ⇒ 00:31:16.730 Samuel Roberts: Mmm.
316 00:31:19.150 ⇒ 00:31:24.440 Samuel Roberts: I’m a little curious about the pricing, because I saw wobble pricing, but I don’t know what the pricing is based on this.
317 00:31:25.460 ⇒ 00:31:27.179 Samuel Roberts: How does Omni compare, do we know?
318 00:31:27.790 ⇒ 00:31:28.359 Samuel Roberts: I thought you might.
319 00:31:28.360 ⇒ 00:31:36.940 Uttam Kumaran: Oh, Omni, omni… Is gonna be a yearly cost.
320 00:31:37.140 ⇒ 00:31:42.770 Uttam Kumaran: So it’s gonna… so it’s a BI tool plus this capability, so it’s gonna be a minimum of, like, 18…
321 00:31:42.970 ⇒ 00:31:45.640 Uttam Kumaran: It’s a minimum of 20 grand a year.
322 00:31:48.110 ⇒ 00:31:48.690 Samuel Roberts: Okay.
323 00:31:54.760 ⇒ 00:31:55.090 Casie Aviles: Yeah.
324 00:31:55.090 ⇒ 00:31:55.680 Samuel Roberts: Okay.
325 00:31:57.190 ⇒ 00:32:02.409 Uttam Kumaran: How much, yeah, what’s… how… where do you find the Wabi, or Wobble? Whatever it’s called.
326 00:32:02.410 ⇒ 00:32:04.040 Casie Aviles: Yeah, it’s Wabi.
327 00:32:04.040 ⇒ 00:32:04.490 Samuel Roberts: That’s.
328 00:32:04.490 ⇒ 00:32:04.850 Casie Aviles: Sorry.
329 00:32:04.850 ⇒ 00:32:05.520 Samuel Roberts: Saying that wrong?
330 00:32:06.180 ⇒ 00:32:13.799 Casie Aviles: I found it… on just basically Google and Reddit, and I think… Some… it’s been…
331 00:32:14.290 ⇒ 00:32:16.810 Casie Aviles: They launched it around 2023, so…
332 00:32:17.870 ⇒ 00:32:24.180 Casie Aviles: I think it’s, yeah, more on the older side compared to other tools that we have here.
333 00:32:27.890 ⇒ 00:32:28.520 Uttam Kumaran: Okay.
334 00:32:30.270 ⇒ 00:32:30.670 Casie Aviles: Yep.
335 00:32:30.670 ⇒ 00:32:33.230 Uttam Kumaran: And then we didn’t end up trying, TextQL.
336 00:32:34.260 ⇒ 00:32:40.530 Casie Aviles: Oh, no, no. So these… let’s go back. These tools we weren’t able to test out.
337 00:32:42.240 ⇒ 00:32:43.750 Casie Aviles: Only 5.
338 00:32:47.010 ⇒ 00:32:47.930 Uttam Kumaran: Okay. Okay.
339 00:32:48.260 ⇒ 00:32:50.660 Uttam Kumaran: How was, how was the compass?
340 00:32:52.110 ⇒ 00:32:52.850 Casie Aviles: Daxter?
341 00:32:52.850 ⇒ 00:32:53.490 Uttam Kumaran: tablets.
342 00:32:54.050 ⇒ 00:32:54.610 Uttam Kumaran: Yeah.
343 00:32:54.610 ⇒ 00:32:55.180 Casie Aviles: pink.
344 00:32:55.350 ⇒ 00:33:03.589 Casie Aviles: Yeah, I think it’s still kind of early stage, and I mean, the slap… UI is not bad.
345 00:33:04.170 ⇒ 00:33:10.310 Casie Aviles: But… For example, here… Like,
346 00:33:10.520 ⇒ 00:33:16.839 Casie Aviles: the charts are still pretty basic, so it’s not, like, as good as what Omni can do.
347 00:33:17.590 ⇒ 00:33:27.529 Casie Aviles: produce… But yeah, it’s mostly Slap native, and, and… Although there’s no, like…
348 00:33:27.770 ⇒ 00:33:33.680 Casie Aviles: and other interface with, for it, so it’s really just Slack, and…
349 00:33:34.050 ⇒ 00:33:41.830 Casie Aviles: But what’s good there is we’re able to, like, connect other warehouses as well. I briefly showed that here.
350 00:33:43.660 ⇒ 00:33:44.840 Uttam Kumaran: Nice, okay.
351 00:33:45.430 ⇒ 00:33:51.909 Casie Aviles: I don’t think we have the ability to… Like, include more context, so…
352 00:33:52.880 ⇒ 00:33:59.820 Casie Aviles: Compared to, like, the other tools where we’re able to, like, do a system prompt, or, like… definitions.
353 00:34:00.700 ⇒ 00:34:02.839 Casie Aviles: I don’t think that’s,
354 00:34:03.070 ⇒ 00:34:05.859 Casie Aviles: That’s available yet, since it’s pretty early.
355 00:34:07.810 ⇒ 00:34:08.750 Uttam Kumaran: Okay, okay.
356 00:34:11.899 ⇒ 00:34:17.619 Casie Aviles: Yeah, that’s pretty much it, I guess, for… just, we were almost done, I guess, but…
357 00:34:17.620 ⇒ 00:34:18.230 Uttam Kumaran: Okay.
358 00:34:18.239 ⇒ 00:34:22.929 Casie Aviles: Our next… my… just my next steps is to just provide, like.
359 00:34:23.279 ⇒ 00:34:27.419 Casie Aviles: To test it with Eden data. I was just using, like, dummy data for now.
360 00:34:27.779 ⇒ 00:34:32.579 Casie Aviles: That I created in our own… Brainforged fraud, so…
361 00:34:33.559 ⇒ 00:34:36.569 Casie Aviles: Yeah, if I could test, and also, like.
362 00:34:36.729 ⇒ 00:34:43.849 Casie Aviles: yeah, different queries as well. Those were the suggestions, and… And just…
363 00:34:44.629 ⇒ 00:34:48.919 Casie Aviles: check in with the results. That’s pretty much my next steps.
364 00:34:49.590 ⇒ 00:34:50.179 Uttam Kumaran: Okay.
365 00:34:50.620 ⇒ 00:34:51.350 Uttam Kumaran: Okay.
366 00:34:53.620 ⇒ 00:34:56.589 Uttam Kumaran: Okay, cool. Ashwini, Alesh, what’d you guys think?
367 00:34:57.400 ⇒ 00:35:00.670 Awaish Kumar: Yeah, I shared the recommendations.
368 00:35:01.070 ⇒ 00:35:17.809 Awaish Kumar: With the lead on, like, what I would like to see is to connect with the EU data, and… and see, for example, connect with Fed transactions, and answer questions related to revenue, week-over-week comparisons, our revenue discounts, and…
369 00:35:18.440 ⇒ 00:35:32.560 Awaish Kumar: I suggested some tools also, like, I have been looking at QDIO, like, and DataGPT, but these… both tools are something where you have to basically ask.
370 00:35:32.830 ⇒ 00:35:35.780 Awaish Kumar: Ask them to connect and set up the demo.
371 00:35:36.730 ⇒ 00:35:37.340 Uttam Kumaran: Okay.
372 00:35:42.500 ⇒ 00:35:43.950 Uttam Kumaran: What do you think, Ashwini?
373 00:35:44.950 ⇒ 00:35:59.090 Ashwini Sharma: Yeah, I was suggesting, like, why not directly integrate with Gemini, and then see how it works out, right? But yeah, I think it’s a good thing to see how it works with real data, rather than with just dummy data, and see the results.
374 00:36:01.620 ⇒ 00:36:07.739 Uttam Kumaran: Yeah, I agree. Yeah, I don’t… I think we can have a menu of options, so we don’t need to arrive on, like.
375 00:36:07.970 ⇒ 00:36:12.119 Uttam Kumaran: one, but I just want to rule out the bad ones, basically.
376 00:36:12.350 ⇒ 00:36:18.479 Uttam Kumaran: You know, of course, if there’s some clients that are like, hey, we don’t want to use an external, we want everything governed within Google.
377 00:36:18.610 ⇒ 00:36:24.029 Uttam Kumaran: maybe we use… You know, Gemini for that, you know, so…
378 00:36:24.820 ⇒ 00:36:28.559 Awaish Kumar: Yeah, I just looked into Gemini, but that’s…
379 00:36:28.720 ⇒ 00:36:32.260 Awaish Kumar: Like, we don’t get an interface like ChatGPT, where you can…
380 00:36:32.260 ⇒ 00:36:32.750 Casie Aviles: Yeah.
381 00:36:32.750 ⇒ 00:36:35.870 Awaish Kumar: It was, like, prompting inside the…
382 00:36:36.070 ⇒ 00:36:41.250 Awaish Kumar: Google BigQuery UI, generate a query, and then basically gonna run it.
383 00:36:42.330 ⇒ 00:36:43.679 Uttam Kumaran: Hmm, okay, okay.
384 00:36:49.470 ⇒ 00:36:50.190 Uttam Kumaran: Okay.
385 00:36:50.570 ⇒ 00:36:52.549 Casie Aviles: Yeah, I think that’s all.
386 00:36:56.300 ⇒ 00:37:02.200 Uttam Kumaran: Okay, cool. How’s the rest of the day, guys? Anything else you need help with?
387 00:37:11.910 ⇒ 00:37:12.510 Uttam Kumaran: Okay.
388 00:37:12.670 ⇒ 00:37:16.809 Uttam Kumaran: I guess whatever you end up on Hedra, if you can send that to them.
389 00:37:18.250 ⇒ 00:37:22.230 Awaish Kumar: Or give them… we’ll give them an update, because she’ll probably ping me later.
390 00:37:23.460 ⇒ 00:37:25.550 Awaish Kumar: Okay, I haven’t…
391 00:37:26.960 ⇒ 00:37:27.520 Uttam Kumaran: Okay.
392 00:37:30.780 ⇒ 00:37:32.940 Uttam Kumaran: How’s the day, Ashwini? Sam?
393 00:37:34.720 ⇒ 00:37:35.290 Ashwini Sharma: Good.
394 00:37:35.400 ⇒ 00:37:36.140 Samuel Roberts: Excuse me?
395 00:37:36.280 ⇒ 00:37:42.540 Ashwini Sharma: I will be raising some PRs, like, the Metaplane instance that we have is… it’s only in…
396 00:37:42.960 ⇒ 00:37:44.789 Ashwini Sharma: What do you call it? Like,
397 00:37:45.600 ⇒ 00:37:49.229 Ashwini Sharma: Okay, can we, can we monitor dev tables from Metaplane?
398 00:37:50.220 ⇒ 00:37:54.220 Uttam Kumaran: Yeah, you just have to… you have to… Change of grants, probably.
399 00:37:55.690 ⇒ 00:38:06.440 Ashwini Sharma: Okay. Now, I wanted to test out a few things. So rather than going through the UI, right, it’s possible to define those monitors right in the dbt models?
400 00:38:06.870 ⇒ 00:38:12.429 Ashwini Sharma: And… Yeah, I feel that that would be a better way to…
401 00:38:12.850 ⇒ 00:38:16.740 Ashwini Sharma: You know, keep the monitor set at some central place, rather than…
402 00:38:17.080 ⇒ 00:38:21.929 Ashwini Sharma: You know, example, like, for example, you want to change some default configuration of the monitors, right?
403 00:38:22.090 ⇒ 00:38:25.840 Ashwini Sharma: Like, all the monitors should run every 1 hour rather than every 3 hours.
404 00:38:26.310 ⇒ 00:38:28.380 Ashwini Sharma: That’s basically the idea. Yeah, so you’re.
405 00:38:28.630 ⇒ 00:38:32.860 Uttam Kumaran: Where do you… so, can you say it again? So, where do you think you’re gonna maintain that one?
406 00:38:33.130 ⇒ 00:38:36.379 Ashwini Sharma: Along with the dbt, right, we can define a YAML file.
407 00:38:36.380 ⇒ 00:38:38.260 Uttam Kumaran: Oh, okay, okay, great.
408 00:38:38.530 ⇒ 00:38:41.680 Ashwini Sharma: Yep. And then put the monitor over there.
409 00:38:41.970 ⇒ 00:38:44.490 Uttam Kumaran: Oh, fantastic. I didn’t know you could do that.
410 00:38:45.610 ⇒ 00:38:49.039 Ashwini Sharma: Yeah, yeah, I mean, Metaplane says that we can do that.
411 00:38:49.040 ⇒ 00:38:52.710 Uttam Kumaran: That’s up to you to figure out. Yeah, yeah, yeah.
412 00:38:53.190 ⇒ 00:38:58.340 Uttam Kumaran: Okay, cool. I’m also speaking with Anomalo. I’m trying to speak with Anomalo.
413 00:38:59.310 ⇒ 00:39:04.690 Uttam Kumaran: If there’s any other ones you think we should chat with, I’m happy to go call them.
414 00:39:06.550 ⇒ 00:39:08.499 Uttam Kumaran: Unless you think Metaplane is fine.
415 00:39:38.520 ⇒ 00:39:42.759 Casie Aviles: Yeah, I guess I’ll drop off for now, guys.
416 00:39:42.760 ⇒ 00:39:43.250 Uttam Kumaran: Okay.
417 00:39:43.250 ⇒ 00:39:44.960 Casie Aviles: And yeah, I’ll work on…
418 00:39:45.640 ⇒ 00:39:50.719 Casie Aviles: the next steps, and just ping you on Slack whenever I get it. Okay.
419 00:39:51.580 ⇒ 00:39:54.499 Casie Aviles: Alright, yeah, thank you very much, guys, for…
420 00:39:54.500 ⇒ 00:39:54.900 Uttam Kumaran: GS.
421 00:39:54.900 ⇒ 00:39:55.820 Ashwini Sharma: Thank you.
422 00:39:56.500 ⇒ 00:39:58.820 Uttam Kumaran: Yeah, thanks, Casey. Great job. Talk to you soon.
423 00:39:59.300 ⇒ 00:40:00.010 Casie Aviles: Bye.