Meeting Title: Brainforge Data Integration Sync Date: 2026-04-01 Meeting participants: Robert Tseng, Shivani Amar
WEBVTT
1 00:02:46.230 ⇒ 00:02:47.110 Robert Tseng: Hey, Shivani.
2 00:02:52.360 ⇒ 00:02:53.790 Shivani Amar: Use cloud code.
3 00:02:55.230 ⇒ 00:02:56.140 Robert Tseng: I did not.
4 00:02:56.960 ⇒ 00:02:58.160 Shivani Amar: What do you use?
5 00:02:59.040 ⇒ 00:03:01.379 Robert Tseng: We run everything out of cursor.
6 00:03:01.730 ⇒ 00:03:02.689 Shivani Amar: Out of what?
7 00:03:02.960 ⇒ 00:03:03.730 Robert Tseng: Cursor.
8 00:03:04.180 ⇒ 00:03:07.570 Shivani Amar: For a surf, okay. Does that have an AI component, too?
9 00:03:08.260 ⇒ 00:03:10.019 Robert Tseng: Yeah,
10 00:03:10.480 ⇒ 00:03:15.249 Robert Tseng: are less technical. I mean, we have… we have… we have a quad code. I mean, we have 3 different…
11 00:03:15.830 ⇒ 00:03:20.229 Robert Tseng: We have, like, a brain-forged-skinned version of, like, a…
12 00:03:20.430 ⇒ 00:03:24.650 Robert Tseng: Simple, like, of a quad code, and then we…
13 00:03:24.800 ⇒ 00:03:29.030 Robert Tseng: have some people using Claude Code, and then we have most of the team using Cursor.
14 00:03:29.700 ⇒ 00:03:30.290 Shivani Amar: Cool.
15 00:03:31.940 ⇒ 00:03:34.600 Shivani Amar: Okay.
16 00:03:35.910 ⇒ 00:03:40.380 Shivani Amar: Where’s the latest version of the document that I can reference? Is it the one that we were in yesterday?
17 00:03:40.840 ⇒ 00:03:43.810 Robert Tseng: Yeah, I mean, I’ll reshare it.
18 00:03:44.050 ⇒ 00:03:45.409 Robert Tseng: In this chat.
19 00:03:45.720 ⇒ 00:03:46.570 Shivani Amar: Okay.
20 00:03:47.430 ⇒ 00:03:51.629 Shivani Amar: Thank you. It’s helpful to, like, look at it with fresh eyes, so…
21 00:04:28.880 ⇒ 00:04:30.250 Shivani Amar: Okay?
22 00:04:31.450 ⇒ 00:04:32.330 Shivani Amar: Alright.
23 00:04:32.850 ⇒ 00:04:35.999 Shivani Amar: So we’ve got Data Foundation and Governance, which is really about, like.
24 00:04:36.610 ⇒ 00:04:41.120 Shivani Amar: Documentation, and making sure, like, consistent data quality.
25 00:04:50.270 ⇒ 00:04:54.029 Shivani Amar: the deliverables… Data dictionary formula signed off.
26 00:04:54.870 ⇒ 00:04:57.020 Shivani Amar: And then this new scorecard thing.
27 00:04:57.130 ⇒ 00:05:01.119 Shivani Amar: And a QA process for each new domain. Okay.
28 00:05:01.850 ⇒ 00:05:06.829 Shivani Amar: Then, the deliverables, we’re talking about commercial data marts and QA,
29 00:05:07.150 ⇒ 00:05:21.230 Shivani Amar: And then, ultimately, that translates into, like, some… dashboards, for… Walmart and Target, Retail sales…
30 00:05:22.020 ⇒ 00:05:27.279 Shivani Amar: Product category performance, growth trends, okay, this is just, like, give a feel for, like, the types of dashboards.
31 00:05:27.730 ⇒ 00:05:30.760 Shivani Amar: Encompass ingestion, which is for DSD.
32 00:05:31.350 ⇒ 00:05:32.350 Shivani Amar: Great.
33 00:05:44.480 ⇒ 00:05:50.009 Shivani Amar: probably revenue reconciliation for a side-by-side comparison.
34 00:05:51.660 ⇒ 00:05:52.650 Shivani Amar: Oh fair.
35 00:06:52.860 ⇒ 00:07:01.489 Shivani Amar: It’s interesting, like, the… We had this, prompt this week, which I’m still sort of, like, working through.
36 00:07:01.710 ⇒ 00:07:07.389 Shivani Amar: And it was like, everybody should come up with one way of using AI.
37 00:07:07.950 ⇒ 00:07:22.680 Shivani Amar: like, this week, to improve the workflows. And you can see that somebody from Supply Chain wrote, ingest historical retail order data from Emerson and inventory data from Geotis to paint a picture of weekly order trends by customer, by SKU,
38 00:07:23.570 ⇒ 00:07:35.210 Shivani Amar: inventory coverage, can we… can we fill the current week’s orders in 100%… 100% in full? This is, like, giving you a feel for, like, people are hungry to start connecting AI to the data.
39 00:07:35.700 ⇒ 00:07:36.390 Robert Tseng: Huh.
40 00:07:36.390 ⇒ 00:07:37.880 Shivani Amar: Even if it’s, like, manually.
41 00:07:38.430 ⇒ 00:07:41.800 Shivani Amar: Okay, let me go back…
42 00:07:50.030 ⇒ 00:07:57.260 Shivani Amar: So, another thing that I wanted to show you was this document that I was working on with…
43 00:07:58.890 ⇒ 00:08:05.230 Shivani Amar: For the COO of… element, and… like…
44 00:08:06.410 ⇒ 00:08:20.089 Shivani Amar: basically the way we’re phrasing this, framing the work is like, okay, we just did a vendor pilot. That was what we did through March, right? Like, basically to select, yeah, like, we’re good to work with Brainforge, now ongoing.
45 00:08:20.280 ⇒ 00:08:26.599 Shivani Amar: And then the phases, the way we’re calling them, it’s, like, build infrastructure, which is, like.
46 00:08:26.760 ⇒ 00:08:31.770 Shivani Amar: The set of data sources, ingesting, modeling, QA, like…
47 00:08:32.340 ⇒ 00:08:36.389 Shivani Amar: connecting to BI, developing dashboards, right? So this is, like.
48 00:08:36.559 ⇒ 00:08:38.780 Shivani Amar: This is the… the thing here.
49 00:08:38.990 ⇒ 00:08:53.130 Shivani Amar: And then this is 2B. And then internal rollout will be kind of the end of the year. Like, once we’ve got a lot of, like… it’s like, we’ll work with the stakeholders, make the dashboards, but then at the end of the year, it’s almost like Omni goes live for a lot more people.
50 00:08:53.800 ⇒ 00:08:54.700 Shivani Amar: Right?
51 00:08:54.890 ⇒ 00:09:02.489 Shivani Amar: And then we’re saying what 2A is, is external connections. And, like, I don’t… this one’s TBD here, but…
52 00:09:02.980 ⇒ 00:09:03.770 Shivani Amar: like…
53 00:09:04.140 ⇒ 00:09:18.590 Shivani Amar: I was, like, just reflecting on the language about NetSuite’s not happening, NetSuite’s not happening, but we’re, like, in implementation with NetSuite. I don’t think there’s gonna be a ton for us to do on the Brainforge side, but I think it’s, like, a success measure is, like.
54 00:09:18.590 ⇒ 00:09:29.770 Shivani Amar: September 1st, did it get integrated well? Does it have the data it needs? Are we ready with the pipes to pull what we need from NetSuite? So, if I think about our contract going through, like.
55 00:09:29.810 ⇒ 00:09:41.539 Shivani Amar: August 31st, and then NetSuite’s go-live date is theoretically September 1st, we’ll want, like, a mapping of which pipes we want, which things we want to pull from NetSuite as time progresses.
56 00:09:41.670 ⇒ 00:09:47.970 Shivani Amar: So… Similar with Atomic, while we haven’t selected the vendor, it’s kind of like…
57 00:09:48.260 ⇒ 00:09:53.090 Shivani Amar: It’s like, when they’re ready, we’re gonna have to give them access to Snowflake, right?
58 00:09:53.270 ⇒ 00:09:54.040 Shivani Amar: And say, like.
59 00:09:54.040 ⇒ 00:09:54.470 Robert Tseng: Yeah.
60 00:09:54.470 ⇒ 00:09:59.860 Shivani Amar: These are the things that you can take. So, I’m… I don’t know if it necessarily needs to be, like, a separate
61 00:10:00.760 ⇒ 00:10:17.370 Shivani Amar: like, if it can be layered in somewhere, or it can be part of, like, you know, you kind of say, for example, you have demand planning validation, reporting, point of sales velocity trends versus distributor order, blah blah blah blah blah. Like, I’m like, cool, it could also be, like.
62 00:10:17.570 ⇒ 00:10:24.990 Shivani Amar: Could, like, a deliverable here could be, like, connect…
63 00:10:36.160 ⇒ 00:10:41.850 Shivani Amar: Something like that. So it’s just like, hey, when they want access to Snowflake, we, like, make sure they get what they need.
64 00:10:42.020 ⇒ 00:10:59.989 Shivani Amar: which hopefully isn’t, like, a huge time suck, because they’re very, like, oh, you’re already pulling so much of this, this is great, but it will be sometime. So I just kind of want to make sure, like, when I sign this document, that I’m, like, we’re being comprehensive about the work at hand. And then when I look at how we’re talking about NetSuite, right?
65 00:11:00.350 ⇒ 00:11:06.139 Shivani Amar: Out of scope, NetSuite implementation expected Q4, ingestion dashboards dependent on Go Live.
66 00:11:08.040 ⇒ 00:11:10.100 Shivani Amar: I’m trying to figure out, like, how to f-
67 00:11:10.460 ⇒ 00:11:12.770 Shivani Amar: how to… do you get what I’m saying? It’s like…
68 00:11:13.910 ⇒ 00:11:21.680 Shivani Amar: kind of say, like, yeah, we’re aware and, like, we’ll play a supporting role as needed, but it’s, like, generally not owned by the Brainforge team.
69 00:11:25.070 ⇒ 00:11:26.080 Robert Tseng: Yeah…
70 00:11:26.660 ⇒ 00:11:28.599 Shivani Amar: Because otherwise, it makes me feel like it’s, like.
71 00:11:29.160 ⇒ 00:11:40.190 Shivani Amar: we’re kind of like, no, we don’t touch anything related to NetSuite. And I’m like, I don’t… I don’t know, like, when NetSuite’s ready to go live, we’ll want to say, like, okay, these are the codes we want to start pulling in pretty immediately, we want to have a plan.
72 00:11:41.570 ⇒ 00:11:55.109 Robert Tseng: Yeah, well, I mean, I think there’s a reason why we kept NetSuite out of scope. I feel like this has kind of changed in the understanding. Well, because NetSuite, and I don’t know, we’ve not plugged into Atomic, I’m sure we could do it, but NetSuite is a…
73 00:11:55.420 ⇒ 00:12:11.759 Robert Tseng: like, if you go to any data connector tool, NetSuite prices a premium that’s probably, like, 5X what you would get for any other data connector. Even if we rely on Polytomic to do it, which Polytomic did not do it for us on a previous client, we have to actually build the custom data connector. So, it is a big lift.
74 00:12:11.760 ⇒ 00:12:13.860 Shivani Amar: What do you mean by that?
75 00:12:14.660 ⇒ 00:12:31.000 Robert Tseng: Because just the way that their API is set up, they… I mean, they have an… they have a REST API, but then they also have something called SuiteScript, or whatever, it’s, like, their own version, so, like, compared to, like, Google Ads, which is a pretty standardized plug-and-play API connector, every, like.
76 00:12:31.200 ⇒ 00:12:36.960 Robert Tseng: data connector platform out there can connect to Google Ads pretty easily, but like NetSuite.
77 00:12:37.200 ⇒ 00:12:40.909 Robert Tseng: Requires custom dev work in order to
78 00:12:41.210 ⇒ 00:12:43.599 Robert Tseng: Be able to go and grab, like.
79 00:12:43.780 ⇒ 00:12:52.650 Robert Tseng: custom fields that you’ve set up in NetSuite. I mean, whatever they have in their REST API is just, like, generic fields, but…
80 00:12:52.990 ⇒ 00:13:02.049 Robert Tseng: it’s likely that whoever is running NetSuite, especially with the migration plan, you’ve set up all these, like, custom workflows, custom fields in it, in order
81 00:13:02.440 ⇒ 00:13:06.089 Robert Tseng: get it, it’s not going to be… it’s not going to be a plug-and-play.
82 00:13:08.330 ⇒ 00:13:15.220 Shivani Amar: You think… you think Jason has that, like, internalized?
83 00:13:16.890 ⇒ 00:13:22.149 Robert Tseng: I… I mean, I’m assuming if you’re migrating from QBO to… to NetSuite, like.
84 00:13:22.470 ⇒ 00:13:30.540 Robert Tseng: those… whatever’s kind of informing that migration probably already has some of those requirements. But yeah, I mean…
85 00:13:30.540 ⇒ 00:13:45.630 Shivani Amar: Atomic wouldn’t build connections to NetSuite. When I was at Brave Health, we were connected to NetSuite with Fivetran. So is it that, like, Fivetran just has connections to NetSuite kind of built out for a lot of customers? Because I imagine those are, like, big companies.
86 00:13:45.990 ⇒ 00:13:47.100 Shivani Amar: using Netflix.
87 00:13:47.100 ⇒ 00:13:55.559 Robert Tseng: Trend probably has… I mean, it does have a NetSuite connector. I’m sure it doesn’t have every custom field that you want, like, probably has a bigger, like.
88 00:13:55.870 ⇒ 00:14:05.579 Robert Tseng: base than what Polyatomic has, but I’m sure that Fivetran charges you at least 5x what it would cost a normal connector, or to plug into it.
89 00:14:05.580 ⇒ 00:14:17.600 Robert Tseng: That, and you also need NetSuite… you’re paying NetSuite a fee to get their API. They don’t… they don’t let you get it out of the box, either. So, one of our other clients, they pay, like, five grand a month just to, like, let NetSuite
90 00:14:17.640 ⇒ 00:14:25.519 Robert Tseng: API be accessible via, we are using a mix of polyatomic plus, like, custom,
91 00:14:25.800 ⇒ 00:14:28.329 Robert Tseng: cron jobs that we’ve set up to pull data out of there.
92 00:14:38.210 ⇒ 00:14:41.279 Shivani Amar: Yeah, I’m having trouble, like, kind of.
93 00:14:41.560 ⇒ 00:14:48.289 Robert Tseng: like, not all data sources are made equal. Like, this is considered, like, a premium, like, connector for… if you look at the productized platform.
94 00:14:48.290 ⇒ 00:14:55.030 Shivani Amar: I hear you on that, but I… I know that, like… There’s a component of, like.
95 00:14:55.440 ⇒ 00:15:12.730 Shivani Amar: for NetSuite to go live, like, we need to be able to get data from NetSuite. And so I’m like, is that… I think I just want to touch base with Jason, be like, is that really fully being owned by the tech team, and that we’re building something ourselves? Or will any work come to Brainforge, and let’s just, like, make sure we’re all super aligned?
96 00:15:13.070 ⇒ 00:15:13.660 Robert Tseng: Yeah.
97 00:15:14.020 ⇒ 00:15:24.780 Robert Tseng: I agree with you that it is… I mean, this is not from this call, but when you were saying, hey, maybe this is a separate worksheet, I agree, like, NetSuite, like, just pulling data out of NetSuite and
98 00:15:25.080 ⇒ 00:15:28.009 Robert Tseng: is… is it so in separate workstream? Like, I… I think…
99 00:15:28.610 ⇒ 00:15:40.009 Robert Tseng: like, I don’t think we would… we would be able to kind of handle it with the… with the timeline that we have here. So, I mean, we could add another… maybe we just have to pull in another data engineer beyond, kind of.
100 00:15:40.560 ⇒ 00:15:45.160 Robert Tseng: I mean, like, I have a guy whose half his time is literally just doing NetSuite.
101 00:15:45.280 ⇒ 00:15:50.139 Robert Tseng: data, like, maintenance on one of our clients. Gotcha.
102 00:15:50.420 ⇒ 00:16:03.240 Robert Tseng: Yeah, like, I mean, in a world where that ends up becoming a workstream, I would basically pull them off as other two clients and just be like, alright, you’re just doing NetSuite maintenance for two clients, and, like, that’s how I imagine that staffing working.
103 00:16:03.240 ⇒ 00:16:04.269 Shivani Amar: Yeah, that makes sense.
104 00:16:04.270 ⇒ 00:16:04.630 Robert Tseng: Yeah.
105 00:16:04.630 ⇒ 00:16:17.740 Shivani Amar: So, Jason looks like he’s in a huddle right now, and I think that I would love to just huddle the three of us when he’s free, so let me see when he frees up. You’re… you’re kind of available for the next 15, right?
106 00:16:18.210 ⇒ 00:16:20.920 Robert Tseng: Yeah, I can be up until 3 Eastern.
107 00:16:20.920 ⇒ 00:16:27.100 Shivani Amar: Up until 3? Okay, so I’ll… if he’s free, then I’ll just, like, create a Slack huddle so we can just connect and…
108 00:16:27.100 ⇒ 00:16:27.420 Robert Tseng: Sure.
109 00:16:27.420 ⇒ 00:16:35.719 Shivani Amar: finalize this. I just… I’m, like, nervous to… I just want to make sure that he and you are aligned on this, versus… versus me, like, trying to translate it, okay?
110 00:16:35.720 ⇒ 00:16:36.330 Robert Tseng: Yeah.
111 00:16:36.330 ⇒ 00:16:37.679 Shivani Amar: Okay, thank you.
112 00:16:39.330 ⇒ 00:16:40.090 Robert Tseng: See ya.