Meeting Title: Uttam-Kumaran’s-Personal-Meeting-Room Date: 2024-09-06 Meeting participants: Ryan Luke Daque, Abigail Zhao, Uttam Kumaran


WEBVTT

1 00:00:35.330 00:00:36.440 Uttam Kumaran: Yes.

2 00:00:36.510 00:00:40.740 Uttam Kumaran: apologies just like in in person meetings.

3 00:00:41.120 00:00:41.650 Uttam Kumaran: See?

4 00:00:41.650 00:00:42.680 Ryan Luke Daque: Yeah, it’s what.

5 00:00:43.750 00:00:44.598 Uttam Kumaran: Hey, how are you?

6 00:00:45.720 00:00:46.540 Uttam Kumaran: Well.

7 00:01:21.750 00:01:24.209 Uttam Kumaran: let me just see if anyone else is gonna join.

8 00:01:58.940 00:02:01.090 Uttam Kumaran: Okay, I think it might just be

9 00:02:01.670 00:02:02.580 Uttam Kumaran: us.

10 00:02:04.585 00:02:10.650 Uttam Kumaran: Maybe, Ryan, I could. Maybe if we just want to talk through that AI thing I mean, I just wanna see what

11 00:02:12.320 00:02:14.889 Uttam Kumaran: what you were, what you were thinking, or even just like

12 00:02:15.468 00:02:18.801 Uttam Kumaran: what you read in that thing. It’s been a while since I read that.

13 00:02:19.040 00:02:22.159 Ryan Luke Daque: Yeah, sure. Yeah, let me do that.

14 00:02:22.530 00:02:23.540 Ryan Luke Daque: So

15 00:02:23.880 00:02:27.230 Ryan Luke Daque: I guess. Let me share my screen.

16 00:02:33.160 00:02:34.830 Ryan Luke Daque: Let me open that

17 00:02:35.240 00:02:36.860 Ryan Luke Daque: blog post.

18 00:02:50.710 00:02:52.800 Ryan Luke Daque: Can you see my screen.

19 00:02:54.200 00:02:55.030 Uttam Kumaran: Yes.

20 00:02:58.510 00:03:01.353 Ryan Luke Daque: Yeah. So basically, this blog post is

21 00:03:02.030 00:03:06.649 Ryan Luke Daque: showing us how to like integrate chat, gpt with github actions.

22 00:03:06.970 00:03:08.030 Ryan Luke Daque: So

23 00:03:08.380 00:03:16.619 Ryan Luke Daque: they do have a github repository that already has like this integration setup.

24 00:03:17.050 00:03:18.833 Ryan Luke Daque: So it has all the

25 00:03:20.320 00:03:26.010 Ryan Luke Daque: basically, it has all the like prompts and stuff like that. So this is

26 00:03:27.120 00:03:38.209 Ryan Luke Daque: this is for code review, though. So basically, what this does is like if we do, if we create the the workflow file, and then, just use the

27 00:03:38.920 00:03:49.299 Ryan Luke Daque: the repo. This repo that they created. This is going to do a code review on any Pr that gets opened right or reopened. So basically.

28 00:03:50.165 00:03:54.540 Ryan Luke Daque: chat, Gpt will basically read the code

29 00:03:54.940 00:03:57.629 Ryan Luke Daque: something like this. Yeah, it’s

30 00:03:58.650 00:04:00.430 Ryan Luke Daque: not clear, but

31 00:04:00.640 00:04:02.439 Ryan Luke Daque: something like this. It

32 00:04:04.890 00:04:07.850 Ryan Luke Daque: chat. Gpt will read all the files changed

33 00:04:08.020 00:04:09.430 Ryan Luke Daque: and the Pr.

34 00:04:09.470 00:04:17.010 Ryan Luke Daque: And does a code review, and does like add all the code comments here, and like potential bugs and cold smells and everything like that.

35 00:04:17.050 00:04:18.269 Ryan Luke Daque: So this is

36 00:04:18.450 00:04:28.089 Ryan Luke Daque: Chat Gbt, actually doing this. Open. AI, right? This is cool. So this is what I was trying to. That was what I’m currently trying to

37 00:04:28.260 00:04:30.990 Ryan Luke Daque: ado at the moment, just to see if this

38 00:04:31.130 00:04:32.490 Ryan Luke Daque: actually works

39 00:04:32.890 00:04:36.140 Ryan Luke Daque: just using their repository to do that.

40 00:04:36.290 00:04:38.620 Ryan Luke Daque: And once that works. We can like.

41 00:04:38.750 00:04:43.730 Ryan Luke Daque: try to fork this and like, create our own prompts, like, whether

42 00:04:44.230 00:04:48.510 Ryan Luke Daque: we want open AI AI to

43 00:04:49.030 00:04:49.800 Ryan Luke Daque: like

44 00:04:50.120 00:04:52.400 Ryan Luke Daque: check the jobs in the workflow.

45 00:04:52.480 00:05:03.480 Ryan Luke Daque: like specifically the intraday job, run for Dbt. And if ever there are errors, especially like the anomaly detection thing, where we set everything to warning at the moment. So if

46 00:05:04.130 00:05:09.870 Ryan Luke Daque: I’m not sure at the moment, but like what, what, where I was thinking is we might have

47 00:05:10.090 00:05:13.830 Ryan Luke Daque: Openai look into those warnings to see if they’re really

48 00:05:14.090 00:05:17.600 Ryan Luke Daque: something that need to be actioned upon. Or

49 00:05:17.870 00:05:20.061 Ryan Luke Daque: it’s just it’s a valid

50 00:05:20.690 00:05:22.560 Ryan Luke Daque: pointing, or I don’t know what.

51 00:05:23.060 00:05:24.560 Ryan Luke Daque: like something like that.

52 00:05:24.640 00:05:27.230 Ryan Luke Daque: Or if there’s any error like it, can.

53 00:05:28.640 00:05:30.909 Ryan Luke Daque: it can do like some stuff about it.

54 00:05:31.290 00:05:37.100 Ryan Luke Daque: But yeah, which is cool. So I already tested our like Api stuff. I’m a code.

55 00:05:37.717 00:05:42.530 Ryan Luke Daque: Yeah, it it pretty much works the our open AI Api.

56 00:05:42.750 00:05:46.470 Ryan Luke Daque: the get the Github token. I was able to create. As well.

57 00:05:46.720 00:05:47.339 Ryan Luke Daque: Okay, cool.

58 00:05:47.340 00:05:48.270 Uttam Kumaran: Yeah.

59 00:05:48.270 00:05:52.379 Ryan Luke Daque: I’m just working on this at the moment to try to see if we can

60 00:05:52.970 00:05:55.190 Ryan Luke Daque: at least have the code review.

61 00:05:56.620 00:05:57.740 Ryan Luke Daque: work. So.

62 00:05:57.740 00:06:01.609 Uttam Kumaran: Yes, maybe honestly, it could be nice if you want to try it, even in the real

63 00:06:02.164 00:06:08.889 Uttam Kumaran: repo, or something that maybe like, yeah, I don’t know. Pool parts may be like a little bit crazy

64 00:06:09.130 00:06:15.359 Uttam Kumaran: like if you try it in the real repo that way, you could test really quickly and like you can merge into whatever you need.

65 00:06:15.780 00:06:16.380 Ryan Luke Daque: Yeah.

66 00:06:16.820 00:06:20.169 Uttam Kumaran: And then once we get it working there, we can start to move it into client stuff.

67 00:06:20.880 00:06:25.850 Ryan Luke Daque: Yeah, that sounds great. Yeah, I’ll I’ll see what I can do with the the real repo.

68 00:06:28.260 00:06:28.850 Uttam Kumaran: Cool. Okay.

69 00:06:29.600 00:06:35.629 Uttam Kumaran: yeah. The other only other update from my end is, I have like a string of meetings I’m like in

70 00:06:36.180 00:06:43.169 Uttam Kumaran: West Austin right now, and just like had a bunch of meetings, and then need to drive home but

71 00:06:43.290 00:07:06.839 Uttam Kumaran: had a good meeting. We have one proposal out for an AI deal. That we’re working on. It’s it’s for this client who needs to scrape like target websites actually, and scrape to see if up one of their products is in stock. And so we have a little bit of an AI process that we develop that goes for each of the target websites like target.com.

72 00:07:06.950 00:07:21.949 Uttam Kumaran: and then looks for the product and can like basically pick a Zip code and tell you whether it’s in stock or not. So working on that. We have another. We got a verbal yes, from Javi coffee, which is like we’re a coffee tick tock coffee brand.

73 00:07:22.273 00:07:27.730 Uttam Kumaran: So I’m not sure when that’s gonna start yet, but still have to sign a contract for them.

74 00:07:28.632 00:07:32.187 Uttam Kumaran: Stella decided not to renew this

75 00:07:32.960 00:07:38.980 Uttam Kumaran: week. But they may bring us on for, like some sort of retainer hours

76 00:07:39.985 00:07:40.880 Uttam Kumaran: for

77 00:07:41.955 00:07:42.890 Uttam Kumaran: the

78 00:07:42.920 00:07:46.420 Uttam Kumaran: next few months, and then we’ll probably loop back around with them in q. 1

79 00:07:49.810 00:08:12.949 Uttam Kumaran: I think, Abigail, the one thing on your side is, I still owe you some stuff. But I added a notion. Like a little bit of a document around like sales goals, and I created a little bit of like a sales forecast and kind of looking at the conversion rates. I think that’ll be helpful for some conversations next week. Basically looking at like, okay, we if we want to hit these sort of revenue goals. Here’s how many clients we need to convert.

80 00:08:12.950 00:08:22.708 Uttam Kumaran: Here’s how many meetings we need to have. Here’s how many email opens we need. And here’s how many emails we get sent. I think I sent a little bit of like an insight into

81 00:08:23.000 00:08:29.380 Uttam Kumaran: what that looks like. And then basically that will back that into like. Here’s how many leads we need to move from Apollo into Hubspot.

82 00:08:29.570 00:08:34.030 Uttam Kumaran: and then move from Hubspot into instantly the email.

83 00:08:34.059 00:08:45.189 Uttam Kumaran: And we have, like 24,000 credits. Right? We have more than enough. I think the biggest thing will be. We’re gonna have minimum numbers on emails that need to go out right. So I think the nice thing is.

84 00:08:45.210 00:08:49.699 Uttam Kumaran: we’re really we have a lot of credits that we can use. We’re going to use the Hubspot scoring

85 00:08:49.720 00:08:58.880 Uttam Kumaran: to shorten that even more. And then our goal will be looking every day and see like what’s like a thousand divided by 30 like 300.

86 00:08:59.770 00:09:11.879 Uttam Kumaran: Is that right? Yeah, like No. 30. So we should see if, like 30 emails, you’re a math, major. You should approach me on that. But that’s we’ll see if 30 emails a day on average go out right?

87 00:09:12.280 00:09:15.290 Uttam Kumaran: Or like about 200 250 emails a week.

88 00:09:15.330 00:09:20.269 Uttam Kumaran: So this is how we set minimums right? We need to know that we need. We are doing that much. Outreach

89 00:09:20.623 00:09:27.480 Uttam Kumaran: and then that’ll. And then the second piece of that is, once we have that much outreach going, we’ll start to get the engagement store going.

90 00:09:27.980 00:09:51.759 Uttam Kumaran: So once once the outreach is going out, we expect people to come to the website, and then we can start looping back that engagement score and then ideally we have both engagement that we didn’t poke people on that people are coming naturally. And then we have engagement where we are poking people on. And then basically, once we see that people opening our emails and that they’re looking at the site, we’ll basically get understanding that they’re super interested.

91 00:09:52.244 00:10:03.689 Uttam Kumaran: They have more propensity. We’ll figure out quickly, like after a few 1,000 emails and a few 1,000 clicks on the site, like what people who are actually interested in on do on the site. And then that’ll make us make the site better. So

92 00:10:04.554 00:10:10.230 Uttam Kumaran: that’s a bunch of stuff I want to try to kick off next week. And

93 00:10:10.510 00:10:17.010 Uttam Kumaran: yeah, there’s probably like 10 more things. Oh, I met with the CTO of rail just right before this.

94 00:10:17.494 00:10:23.335 Uttam Kumaran: So, Ryan, I think you might be interested. And we’re gonna start to do some projects with them.

95 00:10:23.920 00:10:29.270 Uttam Kumaran: And I want to connect you and Patrick to some people on their team. Maybe next week

96 00:10:29.280 00:10:34.659 Uttam Kumaran: that whole conversation went that went super. Super. Well, he’s really great guy. That’s a pretty technical guy.

97 00:10:37.460 00:10:38.970 Uttam Kumaran: yeah.

98 00:10:40.220 00:10:41.313 Uttam Kumaran: think that’s

99 00:10:44.740 00:10:47.840 Uttam Kumaran: That’s the biggest stuff from this week.

100 00:10:49.530 00:10:53.810 Ryan Luke Daque: Yeah, I just added the discord countless. I mean like

101 00:10:54.280 00:10:55.790 Ryan Luke Daque: discord server for.

102 00:10:55.790 00:10:56.400 Uttam Kumaran: Cool.

103 00:10:56.770 00:10:59.049 Ryan Luke Daque: For real as well. So yeah, that’s cool.

104 00:10:59.330 00:11:03.299 Uttam Kumaran: Yeah. Feel free to be in there and chat and like, ask questions.

105 00:11:04.210 00:11:05.040 Ryan Luke Daque: Yeah, sure.

106 00:11:08.100 00:11:22.449 Uttam Kumaran: Yeah, I think that’s it. I think maybe next week or the week after we’re gonna start having some dedicated conversations on the AI side. I think Miguel is working on some automations that like, for example, we’re all in this meeting. After this meeting you should get hit with a slack bot with like

107 00:11:22.540 00:11:41.359 Uttam Kumaran: action items. Basically your action items. Actually, so we’re working on some stuff like that. That’s like, really interesting, really fun. And I wanna kind of have like a weekly meeting where we basically be like, Hey, that sucks, or this is working out like, I wonder if we can add this kind of like, and we’re the client in the situation, right? Right? And it kind of like.

108 00:11:41.520 00:11:54.768 Uttam Kumaran: If this AI client signs, then I’m actually more inclined to kind of start talking more about how we all get up to speed and can start doing AI services as well. So I’m speaking to a few other firms related to that.

109 00:11:57.370 00:11:59.789 Uttam Kumaran: yeah, I think that’s it. Anything

110 00:12:00.150 00:12:02.880 Uttam Kumaran: Abigail, are gonna talk about.

111 00:12:05.431 00:12:10.080 Abigail Zhao: I’m good on my end. I’ll just take a look at the notion and like stuff like that and see what

112 00:12:11.510 00:12:14.349 Abigail Zhao: should be. Fine. Yeah, okay.

113 00:12:16.200 00:12:17.390 Ryan Luke Daque: Yeah. I’m also good.

114 00:12:17.710 00:12:28.930 Uttam Kumaran: Okay, yeah, Ryan, let me know. I would live. Blog, how the tragically stuff going. I’m really interested. You could still throw that into engineering or throw that into the automations channel. But I think everybody

115 00:12:29.130 00:12:32.550 Uttam Kumaran: would be really interested to how like what you’re seeing in progress. So.

116 00:12:33.230 00:12:34.650 Ryan Luke Daque: Yeah, cool. Sure.

117 00:12:34.870 00:12:35.570 Ryan Luke Daque: Cool.

118 00:12:36.910 00:12:41.030 Uttam Kumaran: Okay, great nico is out today. So if any questions for Nico just message me.

119 00:12:41.570 00:12:42.400 Ryan Luke Daque: Sounds good.

120 00:12:42.830 00:12:44.460 Ryan Luke Daque: Thanks for them. Thanks.

121 00:12:44.460 00:12:45.460 Uttam Kumaran: Thanks guys.

122 00:12:46.240 00:12:47.599 Ryan Luke Daque: Have a nice rest of your day.

123 00:12:47.700 00:12:48.340 Ryan Luke Daque: Yeah, bye, bye.

124 00:12:48.340 00:12:49.000 Uttam Kumaran: You too, bye.