Meeting Title: AI Migration Project Check-in Date: 2026-02-17 Meeting participants: Greg Stoutenburg, Mustafa Raja


WEBVTT

1 00:00:36.010 00:00:36.500 Mustafa Raja: A.

2 00:00:36.500 00:00:38.029 Greg Stoutenburg: Hey, Mustafa, how’s it going?

3 00:00:38.030 00:00:39.860 Mustafa Raja: How’s it? Yeah, good, how are you?

4 00:00:40.600 00:00:41.630 Greg Stoutenburg: Doing alright.

5 00:00:42.020 00:00:45.749 Greg Stoutenburg: Glad this, migration work feels like it’s moving along so quickly.

6 00:00:46.690 00:00:52.930 Mustafa Raja: Yeah… Okay, so, wrote this script.

7 00:00:53.350 00:00:57.020 Mustafa Raja: That’s just, that’s this automating cursor.

8 00:00:57.250 00:01:16.160 Mustafa Raja: For every single table and topic we have, right? And what you’re doing is just, adding this sort of AI context to all of those files, just as Max showed us, you know? So, all of these have it now, so I guess I could merge the branch.

9 00:01:16.840 00:01:19.019 Mustafa Raja: Let me take a look…

10 00:01:20.550 00:01:27.099 Mustafa Raja: Let me take a look. It just finished, just now. I’ve been running since last night.

11 00:01:27.690 00:01:29.019 Greg Stoutenburg: That’s a big one.

12 00:01:29.570 00:01:30.949 Mustafa Raja: Yeah, that’s a big one.

13 00:01:33.050 00:01:40.329 Mustafa Raja: Yeah, looks… Looks good to me. I guess, as we go, we could improve this, right?

14 00:01:40.480 00:01:44.320 Mustafa Raja: I’ll keep this script, saved, so… Okay.

15 00:01:44.520 00:01:46.669 Mustafa Raja: No one has to do it manually, ever.

16 00:01:47.050 00:01:53.560 Greg Stoutenburg: Okay, so now we’ve got, you’ve got the merge topics set up.

17 00:01:53.780 00:01:54.540 Greg Stoutenburg: Right.

18 00:01:54.810 00:02:07.810 Mustafa Raja: Yeah, yes. So, you’d see, you’d see, we have revenue details, and then we have revenue summary, right? So, revenue summary would be the summary, summary table that’s coming directly from DBT.

19 00:02:07.880 00:02:10.580 Mustafa Raja: And then the revenue details would be…

20 00:02:10.580 00:02:33.140 Mustafa Raja: oh, we have all of these joins, and these joins are actually coming from dbt that build up the summary table itself, you know? So AI has both of those contexts, you know? So it’s AI’s choice if it wants to be more granular in its chart building, or if it just wants to use the summary table.

21 00:02:33.140 00:02:51.099 Mustafa Raja: It has lots of those choices, right? And I want you to keep the summary table for myself, too, because I’m going to be replicating those, you know, dashboards, so I’d want to use the exact same approach that we would have used in Tableau. So let me know if that works for you.

22 00:02:51.420 00:02:53.320 Greg Stoutenburg: That sounds great. Yep, that sounds great.

23 00:02:53.670 00:02:57.370 Mustafa Raja: Okay, so, so similar here for, for channel performance.

24 00:02:57.620 00:03:07.509 Mustafa Raja: Similar… okay, this order fulfillment and pharmacy SLA, these are… these… pharmacy SLA is the summary.

25 00:03:07.520 00:03:17.969 Mustafa Raja: And then order fulfillment is just us breaking this one down, right? AI changed its name, it said that, yo, this isn’t a good name. Change it.

26 00:03:17.970 00:03:18.560 Greg Stoutenburg: Okay.

27 00:03:20.010 00:03:22.439 Greg Stoutenburg: So which one is the new name? Is Order Fulfillment is the.

28 00:03:22.440 00:03:29.860 Mustafa Raja: Yeah, so if you would see the details at the end, it’s just going to mean that, okay, we have broken down a summary.

29 00:03:31.420 00:03:33.439 Greg Stoutenburg: I see. Okay, I get what you did.

30 00:03:33.780 00:03:36.479 Greg Stoutenburg: So, for all of these, there’s details and there’s summary. Got it.

31 00:03:36.480 00:03:49.209 Mustafa Raja: Yeah, yeah, yeah. So summary means, yeah, it’s going to be one single table that’s going to be coming from dbt, and then for details, it’s going to be breakdown of a summary table, yeah?

32 00:03:49.380 00:04:01.780 Greg Stoutenburg: Got it. Okay, and the summaries of the tables that you created initially, right, where you were just taking the table that existed in BigQuery and sending that in, because that table was already merged with some other tables, and you were sending that in as saying.

33 00:04:01.780 00:04:02.330 Mustafa Raja: Yeah, yeah.

34 00:04:02.330 00:04:02.720 Greg Stoutenburg: table.

35 00:04:02.720 00:04:03.340 Mustafa Raja: Right.

36 00:04:03.340 00:04:05.310 Greg Stoutenburg: Okay, got it. Okay, yep.

37 00:04:05.310 00:04:05.860 Mustafa Raja: Yeah, sure.

38 00:04:05.860 00:04:07.570 Greg Stoutenburg: Up to speed on the approach. Okay.

39 00:04:07.570 00:04:10.209 Mustafa Raja: Yeah, similar, similar here.

40 00:04:10.440 00:04:17.759 Mustafa Raja: Yeah, you see? Yep. But now, I met with Demilade, and he’s saying, break these down too.

41 00:04:19.320 00:04:25.849 Mustafa Raja: These aren’t exact summaries, but, Demilade is saying that cohorts are also a type of summary.

42 00:04:26.220 00:04:33.110 Mustafa Raja: So, they’re powering some, retention tables, so these… these should be broken down too, so these two…

43 00:04:33.470 00:04:35.879 Mustafa Raja: I’ll add details for, also.

44 00:04:36.190 00:04:43.310 Mustafa Raja: And I think this should be good to merge. I might just… Merge this, to be honest.

45 00:04:45.020 00:04:45.580 Mustafa Raja: Boy.

46 00:04:45.580 00:04:46.060 Greg Stoutenburg: Okay.

47 00:04:46.470 00:04:48.649 Mustafa Raja: Productivity march.

48 00:04:53.980 00:04:57.020 Mustafa Raja: I might remove these caveats, what do you think?

49 00:04:57.710 00:05:03.479 Mustafa Raja: These are, these are coming directly from, transcripts, right? So these might not be up to date.

50 00:05:05.380 00:05:11.480 Greg Stoutenburg: Yeah, I would worry about just relying on transcripts, you know, regardless of when

51 00:05:11.880 00:05:15.280 Greg Stoutenburg: those conversations happen, so let’s see, what are the caveats? So…

52 00:05:15.420 00:05:19.019 Greg Stoutenburg: LTV is a lagging indicator, part of something incomplete.

53 00:05:19.300 00:05:24.019 Greg Stoutenburg: Yeah, we don’t need to give them advice, we just want them to be able to find the right information.

54 00:05:24.580 00:05:27.849 Mustafa Raja: Yeah, so examples… I think it’s good, right?

55 00:05:28.070 00:05:28.680 Greg Stoutenburg: Yeah.

56 00:05:29.420 00:05:45.029 Mustafa Raja: I should keep this, I should… I agree. Yeah, so I’m going to quickly write a script that’s going to go through all of these, remove this particular caveat section, and we’re going to keep everything else, but not this.

57 00:05:45.300 00:05:45.670 Greg Stoutenburg: Sounds good.

58 00:05:45.670 00:05:50.109 Mustafa Raja: I didn’t see it last night, I just saw it, and I was like, bro, what? Yeah.

59 00:05:50.380 00:05:52.319 Greg Stoutenburg: Yeah, yeah, okay. No, good call.

60 00:05:52.320 00:06:02.920 Mustafa Raja: But it shouldn’t be… it shouldn’t be a long task. It should be, you know, pretty quick. So, for my to-dos, break down… break these two down.

61 00:06:02.920 00:06:04.530 Greg Stoutenburg: Break down the pivots, yup.

62 00:06:04.660 00:06:13.710 Mustafa Raja: yeah, get rid of these caveats. We have them everywhere, so I’ll get rid of these everywhere, and then what else?

63 00:06:14.870 00:06:17.349 Mustafa Raja: And then…

64 00:06:19.060 00:06:35.919 Mustafa Raja: Maybe Demilade, I remember I told you about this one chart that’s, you know, a lot complex, so Demilade asked me to just break it down into 3 different charts, because it’s 3 different charts merged into one. So I’m going to be doing that also…

65 00:06:37.970 00:06:41.369 Mustafa Raja: Yeah, this… this might be because… yeah.

66 00:06:41.560 00:06:51.690 Mustafa Raja: Okay, I changed the name of, topics. It’s now Channel Performance Summary. Guys, I’ll just need to update that here too, and these will work.

67 00:06:52.110 00:06:56.729 Greg Stoutenburg: Okay, sounds good. So, once those are working, I mean, how long do you think that will take?

68 00:06:59.060 00:07:01.410 Mustafa Raja: Once these are working…

69 00:07:02.910 00:07:08.209 Greg Stoutenburg: Yeah, I mean, how long do you think it’ll take between now and when this dashboard is stood up?

70 00:07:08.970 00:07:12.089 Mustafa Raja: An hour and a half, to be honest.

71 00:07:12.320 00:07:31.660 Greg Stoutenburg: Okay, great. Okay, so, and at that time, let’s get in and just poke around, and, ask it some questions, verify that Blobby is delivering information that seems relevant, so we know that our context and semantic layer is functioning as intended. And then when we feel good about that, we can start bringing in the rest of the dashboards.

72 00:07:32.580 00:07:36.479 Mustafa Raja: Okay, yeah, yeah, let’s meet, let’s meet again today, then.

73 00:07:36.480 00:07:39.960 Greg Stoutenburg: Yeah, we’ll meet again today, that sounds good. And,

74 00:07:40.380 00:07:45.769 Greg Stoutenburg: Have you had a conversation with you, Tom, about, you know, building what he’s calling the Golden Data Set?

75 00:07:46.800 00:07:53.789 Mustafa Raja: No, and that is because, I’m still… this… this semantic layer is still under… under work, right? I need to get to the KV.

76 00:07:53.790 00:07:54.160 Greg Stoutenburg: Yeah.

77 00:07:54.470 00:08:13.650 Mustafa Raja: Sure. But I understand what he’s saying, and how, how, he’s asking me to do it. I understand that, so, that I should be able to do. And then he also wants to take a track at it, so I’ll be connecting this Omni instance that we have to a GitHub repo.

78 00:08:13.890 00:08:17.629 Greg Stoutenburg: Okay. So, you know, he can work with Kersher around it, you know?

79 00:08:18.040 00:08:27.610 Greg Stoutenburg: Yeah, that sounds good. And then, finally, I think before we let any Omni stakeholders, I mean, sorry, any Eden stakeholders into this.

80 00:08:27.610 00:08:42.239 Greg Stoutenburg: let’s look through the topic… the topic contexts that have been created, just to make sure that if there’s anything that’s, like, sensitive that came up in a transcript, that they’re not seeing it. You know, so one thought I had was, if we’re sending in call transcripts, sometimes before

81 00:08:42.289 00:08:52.790 Greg Stoutenburg: a stakeholder gets on, there’s a conversation, like, alright, how are we gonna handle this objection? You know what I mean? Like, some kind of, like, internal discussion. Let’s just skim through and make sure that we don’t see anything like that standing out.

82 00:08:53.470 00:08:55.270 Mustafa Raja: Okay, yeah, I agree.

83 00:08:55.950 00:09:00.080 Mustafa Raja: I guess that sort of thing might appear in caveats.

84 00:09:00.280 00:09:07.790 Mustafa Raja: We’ll remove that, and then… Yeah. It might be smart to still, you know, have a manual check over it.

85 00:09:07.970 00:09:09.169 Greg Stoutenburg: Yeah, yep.

86 00:09:10.000 00:09:10.430 Greg Stoutenburg: Like, I mean.

87 00:09:11.290 00:09:27.989 Greg Stoutenburg: Yeah, exactly, right. Like, someone looks through, like, we let someone in, they’re like, where did you say… did you imply I’m difficult? Because that showed up in the AI contest. We don’t want that. Not that I’m saying anyone’s, you know, bad-mouthing, you know, clients, but, just to be sensitive to that sort of thing.

88 00:09:27.990 00:09:29.430 Mustafa Raja: Yeah, yeah, yeah, yeah, yeah.

89 00:09:29.710 00:09:30.840 Greg Stoutenburg: Cool. Okay.

90 00:09:30.840 00:09:35.209 Mustafa Raja: If you have a good relationship going on, let’s keep… Yeah, we wanna…

91 00:09:35.210 00:09:39.230 Greg Stoutenburg: Wanted to stay that way, yeah. Yeah, you know, it’s… I mean…

92 00:09:39.230 00:09:39.730 Mustafa Raja: Thank you so much.

93 00:09:39.730 00:09:48.899 Greg Stoutenburg: I guess a funny example, I think of that partly because I remember this story where, this made it into news where someone went to pick up a prescription.

94 00:09:48.920 00:10:03.610 Greg Stoutenburg: And on the label, it said something like, you know, like, she’s a loudmouth. So, like, the client… the customer goes to pick up this prescription, and they see basically a mean note that had been written by a pharmacist that was intended to be in, like, some confidential section.

95 00:10:03.650 00:10:08.789 Greg Stoutenburg: But then, you know, like, they see it, it’s like, my pharmacist thinks I’m a jerk.

96 00:10:09.930 00:10:11.940 Greg Stoutenburg: Not great, so we don’t want to do that.

97 00:10:12.350 00:10:18.590 Mustafa Raja: Okay, cool, thanks for this update. Cool, so you’re gonna take out the caveats, you’re going to split up the pivot tables.

98 00:10:20.030 00:10:27.039 Greg Stoutenburg: You’re gonna connect the data sources so that this dashboard is filled in, and then let me know, and we’ll review together.

99 00:10:27.040 00:10:27.830 Mustafa Raja: Yeah.

100 00:10:28.240 00:10:31.459 Greg Stoutenburg: I’m about to start a call with Mixpanel’s AI.

101 00:10:31.630 00:10:36.460 Greg Stoutenburg: product manager, and then after that, I don’t have anything for 4 hours, so.

102 00:10:36.460 00:10:37.300 Mustafa Raja: Okay, yeah, then that’s.

103 00:10:37.300 00:10:38.469 Greg Stoutenburg: Ping me anytime.

104 00:10:38.700 00:10:39.460 Mustafa Raja: Yeah.

105 00:10:40.090 00:10:42.069 Mustafa Raja: Cool. Thank you. Alright. Have a good day.

106 00:10:42.070 00:10:43.189 Greg Stoutenburg: Sure, Mustava. See ya.

107 00:10:43.190 00:10:44.519 Mustafa Raja: Yeah, see ya.