Meeting Title: Brainforce x Interlude Project Sync Date: 2025-08-06 Meeting participants: Rafay’s Circleback.ai Notes, Mustafa Raja, Giselle Agot, Sam Roberts, Andrew Del Rizzo, Uttam Kumaran, Matthew Good, Rafay Iqbal
WEBVTT
1 00:01:43.380 ⇒ 00:01:44.440 Uttam Kumaran: Hey, everyone.
2 00:01:47.480 ⇒ 00:01:47.920 Andrew Del Rizzo: Hello!
3 00:01:48.770 ⇒ 00:01:49.500 Mustafa Raja: Hey!
4 00:01:49.824 ⇒ 00:01:51.119 Uttam Kumaran: Nice to meet you.
5 00:01:52.430 ⇒ 00:01:54.969 Andrew Del Rizzo: Nice to meet you, too. I think we’re just waiting on Rafai to join.
6 00:01:55.220 ⇒ 00:01:56.010 Uttam Kumaran: Yes.
7 00:01:56.920 ⇒ 00:01:58.240 Andrew Del Rizzo: I’m just messaging him right now.
8 00:01:58.610 ⇒ 00:02:00.920 Uttam Kumaran: Is Matt in or no.
9 00:02:02.998 ⇒ 00:02:08.829 Andrew Del Rizzo: He should be. I think we’re just wrapping up another meeting moment. Yeah.
10 00:02:19.560 ⇒ 00:02:25.139 Uttam Kumaran: Have they? Kind of give you an insight to what we’re working on or we can also go through it. So.
11 00:02:25.961 ⇒ 00:02:33.599 Andrew Del Rizzo: No, honestly, I I was actually just asking him. I I just got put into this call. So I’m not 100% sure. I think it’s just an intro for me as well.
12 00:02:33.780 ⇒ 00:02:34.839 Uttam Kumaran: Great great tip.
13 00:02:37.280 ⇒ 00:02:40.790 Uttam Kumaran: So yeah, maybe I’ll give a brief introduction. So
14 00:02:42.000 ⇒ 00:02:46.745 Uttam Kumaran: we are the brainforce team. We’re basically working on a
15 00:02:47.450 ⇒ 00:02:52.719 Uttam Kumaran: integration. For everyone, for
16 00:02:53.454 ⇒ 00:03:04.470 Uttam Kumaran: your team on streamlining, the process of going from sort of transcripts and onboarding docs into actually like finished
17 00:03:04.836 ⇒ 00:03:24.430 Uttam Kumaran: sort of storyboards and notion. For decks. That’s kind of the 1st automation we’re tackling, I think, in today’s meeting. I’ll let Mustafa probably drive a lot of it. But you also have on our side just a couple of people that are starting to shadow. Matthew. So Giselle is onboarding as a project manager in our team.
18 00:03:24.667 ⇒ 00:03:35.102 Uttam Kumaran: So she’s gonna be helping some with some stuff in the background. Just get things organized and Sam is coming on our broader AI team, so he’s just floating across a bunch of meetings as he’s sort of seeing what we’re doing for clients.
19 00:03:35.730 ⇒ 00:03:38.867 Uttam Kumaran: But we made a lot of great progress.
20 00:03:39.710 ⇒ 00:03:45.849 Uttam Kumaran: I don’t know if you had a chance, Matthew, to see any of the loom videos, and you kind of get a sense, for where?
21 00:03:45.850 ⇒ 00:03:53.300 Matthew Good: I looked through. Yeah. I looked through one the audio was a little bit tough, though I think the one that Mustafa sent over. I looked through it yesterday.
22 00:03:54.410 ⇒ 00:04:02.354 Matthew Good: but not honestly, not in its entirety. I’m I’m moving this week, so like my apartment. So I I have a general sense. But
23 00:04:04.340 ⇒ 00:04:11.509 Uttam Kumaran: Yeah. So basically, Mustafa, do you wanna sort of share the n 8 n flow. And I can give everyone like.
24 00:04:11.680 ⇒ 00:04:21.793 Uttam Kumaran: it’s a nice visual. So I give everyone a basically just an overview of where we are with the integration today. I think we want to talk a little bit about
25 00:04:22.250 ⇒ 00:04:25.130 Uttam Kumaran: how we want to evaluate some of these.
26 00:04:25.950 ⇒ 00:04:32.409 Uttam Kumaran: and we want to start to decide on where this is gonna land in notion, and like, what are the human in the loop steps
27 00:04:32.550 ⇒ 00:04:34.870 Uttam Kumaran: but myself feel free? Yeah.
28 00:04:35.290 ⇒ 00:04:36.770 Mustafa Raja: Yeah, let me share my screen.
29 00:04:44.870 ⇒ 00:05:11.830 Mustafa Raja: Yeah, so this is this is the Dex agent that we have right now. So this is an this is the main orchestrator agent, and we have some child agents. That it lets do their work. So we have a summarizer. We have insight analyst this. The job of this insight analyst is that it looks through the files that we would give it.
30 00:05:11.850 ⇒ 00:05:36.630 Mustafa Raja: and it will fetch all of the facts that are present in there, so we can make sure that there’s no fact that is wrong mentioned in the deck. And this is our narrative architect, and this agent is responsible to correctly format
31 00:05:36.800 ⇒ 00:05:38.500 Mustafa Raja: each of the slide.
32 00:05:39.410 ⇒ 00:05:45.370 Mustafa Raja: And this is just to make sure that we have nothing wrong in the yep.
33 00:05:45.370 ⇒ 00:05:51.819 Uttam Kumaran: Yeah, so automatically, if you just zoom out for this, just just this view. So or maybe
34 00:05:51.960 ⇒ 00:06:03.889 Uttam Kumaran: yeah, maybe let’s walk kind of through from the left to right. And I can sort of narrate as you go through. So basically, this is the platform ending end that we use to build our our agents to great orchestration kind of a
35 00:06:04.030 ⇒ 00:06:18.479 Uttam Kumaran: advertises like a low code tool, but we actually do a lot of coding within it. But we trigger web hooks, integrations all from here. So basically, the way this works is like when a chat message received, we execute a bunch of actions. So we
36 00:06:18.480 ⇒ 00:06:42.010 Uttam Kumaran: bring in a bunch of files. We move them all into a really like a 1 piece of context. And then we go into sort of our AI agent mode. We’re using your Claude, you know. Api key for this and then we sort of move into this next phase, which is all the different, you know, agents that are actually acting together to create the final doc, I think.
37 00:06:42.335 ⇒ 00:06:48.844 Uttam Kumaran: If you want to just zoom in and Mustafa just on like having all these agents in one view.
38 00:06:49.240 ⇒ 00:06:54.670 Uttam Kumaran: and we can just pause here. I think one of the things that I described in the beginning of the project is
39 00:06:54.700 ⇒ 00:07:18.049 Uttam Kumaran: kind of the benefits of having a multi agent system. Matthew, before what you were doing is sort of shoving all the contacts into one flow and then kind of like one shotting it, or few shotting it and expecting for the output right works and gets close. But it’s not reliable. And it’s also hard to debug when it doesn’t work.
40 00:07:18.050 ⇒ 00:07:33.559 Uttam Kumaran: And so what do we do when we build great engineering systems. We try to make sure that there’s gets more deterministic. And then we have mechanisms to actually measure and improve. So each of these are, if you want to make it boil it down very, basically just series of prompts. That kind of run
41 00:07:33.560 ⇒ 00:07:50.739 Uttam Kumaran: on top of outputs from each other. So certain agents are just in charge of doing certain things. So we have the ability to add more reduce but our goal is actually to start to measure using both qualitative and quantitative feedback.
42 00:07:50.870 ⇒ 00:08:07.137 Uttam Kumaran: How these agents are performing versus baseline and the baseline is really those examples that we’ve gotten from your team before. So I’ll just pause here. This is like sort of a little bit of into the you know architecture of of how we’re building this
43 00:08:07.600 ⇒ 00:08:35.710 Uttam Kumaran: And you know, hopefully, what you can kind of see is we’re decomposing the problem, building out different components. And what it does is it just allows the improvement of the accuracy. Additionally, the way where you’ll find is that the more narrow use case you have for AI agent the better performing it is of course there’s diminishing returns like we could have literally one agent for every single slide, and that may be overkill, so we will figure out like
44 00:08:35.990 ⇒ 00:08:39.990 Uttam Kumaran: we will find that balance. But this was our sort of 1st crack at the the problem.
45 00:08:40.460 ⇒ 00:08:44.869 Matthew Good: Cool, awesome. Okay. Summarize your insight.
46 00:08:46.170 ⇒ 00:08:47.980 Matthew Good: Got it so.
47 00:08:48.130 ⇒ 00:08:52.040 Uttam Kumaran: Yeah, one thing, Mustafa, that could be helpful, too, is we can get the prompts
48 00:08:52.420 ⇒ 00:08:55.184 Uttam Kumaran: in notion, or we can make sure that
49 00:08:56.020 ⇒ 00:08:57.970 Uttam Kumaran: Matthew and Team can go in here and poke around.
50 00:08:57.970 ⇒ 00:08:58.390 Mustafa Raja: Oh, yeah.
51 00:08:58.741 ⇒ 00:09:02.959 Uttam Kumaran: That way. They can see the prompts. That’s probably the most tangible.
52 00:09:02.960 ⇒ 00:09:06.227 Matthew Good: I was, gonna say, like, What’s what questions should I be asking.
53 00:09:06.500 ⇒ 00:09:12.259 Uttam Kumaran: Yeah. So let’s let’s let’s actually, let’s actually go to the next piece. Mustafa, do you want to go to the stuff we’re doing on evals?
54 00:09:12.960 ⇒ 00:09:15.869 Mustafa Raja: Yeah, let me what we did.
55 00:09:16.993 ⇒ 00:09:20.760 Uttam Kumaran: So so the probably the next best question is like, Okay, like.
56 00:09:20.860 ⇒ 00:09:24.370 Uttam Kumaran: let’s see an example of like what the output is, and then
57 00:09:24.940 ⇒ 00:09:35.249 Uttam Kumaran: your your 1st instance may be okay. This looks pretty decent. But then, what is the process of like getting feedback from you as you try to use the system is what we’re we typically use just
58 00:09:35.420 ⇒ 00:09:37.595 Uttam Kumaran: program called Brain trust.
59 00:09:38.250 ⇒ 00:09:47.180 Uttam Kumaran: basically, it’s an evaluation software for AI responses. So feel free to pull up any example.
60 00:09:48.580 ⇒ 00:09:54.820 Uttam Kumaran: If there’s a clear example of like the side by sides, or any scores perfect.
61 00:09:54.820 ⇒ 00:10:01.669 Mustafa Raja: Yes, this is an example, input, we have the question over here, and then we have the transcript.
62 00:10:02.487 ⇒ 00:10:06.180 Mustafa Raja: And then we have our expected output.
63 00:10:06.480 ⇒ 00:10:08.119 Mustafa Raja: And this is our output.
64 00:10:08.930 ⇒ 00:10:10.680 Mustafa Raja: This is the output that
65 00:10:11.870 ⇒ 00:10:17.470 Mustafa Raja: these are, yeah. This is the output that we gave. And then this is the expert one.
66 00:10:18.380 ⇒ 00:10:26.199 Uttam Kumaran: So the one in the expected is like our control. Right? It’s what you guys gave us is what you guys said, this is like what we should expect.
67 00:10:26.270 ⇒ 00:10:50.310 Uttam Kumaran: We also now get an output from the AI agent, and so we want to start to measure the distance between those on several different modes. We have, like some common evaluation steps that are a little bit less that are more scientific. We also have some that are going to be using what’s called like Llm. As a judge, we’re basically another Llm. On top compares and is like, here’s how far off they are
68 00:10:50.390 ⇒ 00:10:53.790 Uttam Kumaran: over time. We want to be able to produce results
69 00:10:53.890 ⇒ 00:11:01.305 Uttam Kumaran: that are as close to the outputs that are deemed great are right. And so
70 00:11:01.870 ⇒ 00:11:27.029 Uttam Kumaran: one step beyond, just like, okay, generally, that it feels good is actually running scores. So every output that comes out from the agents that we build we score. And you know you’ll be able to actually see. Okay, how far are we? And and that’s when the improvements kind of come in. So there’s a couple of opportunities for improvements. And I don’t know, Mustafa, do we have a an example of an output end to end that we can just
71 00:11:27.190 ⇒ 00:11:32.900 Uttam Kumaran: throw the screen and then maybe everyone can sort of just read through really quickly.
72 00:11:34.840 ⇒ 00:11:36.559 Uttam Kumaran: Did we put something in motion.
73 00:11:38.680 ⇒ 00:11:44.280 Mustafa Raja: We can do we let’s just pull and output.
74 00:11:44.780 ⇒ 00:11:49.450 Uttam Kumaran: Yeah output, and then we can throw it in notion. So everybody can see it.
75 00:11:52.230 ⇒ 00:11:54.349 Mustafa Raja: Yeah, yeah, perfect. This one.
76 00:11:54.520 ⇒ 00:12:00.159 Uttam Kumaran: Do you want to just take this copy, paste it into our interlude notion, and then I’ll send that link to everybody here.
77 00:12:00.470 ⇒ 00:12:01.250 Mustafa Raja: Yep.
78 00:12:17.960 ⇒ 00:12:19.760 Uttam Kumaran: yeah. Feel free to just to throw it at
79 00:12:20.860 ⇒ 00:12:23.029 Uttam Kumaran: at the bottom of this, for now it’s fine, soft.
80 00:12:23.030 ⇒ 00:12:23.600 Mustafa Raja: Okay.
81 00:12:26.590 ⇒ 00:12:35.909 Rafay Iqbal: And after we’re done going through this as well, I wanted energy. Give some of his input. Since he’s been working on some updates for our notion side of things as well like the project management side. So.
82 00:12:35.910 ⇒ 00:12:36.380 Uttam Kumaran: Oh, great!
83 00:12:36.400 ⇒ 00:12:41.479 Rafay Iqbal: I’ll let him ask to see if he wants to have your insight on anything as well. After you’ve done this.
84 00:12:42.480 ⇒ 00:12:47.039 Uttam Kumaran: Cool so maybe, Mustafa, I could just share my. I’ll just share my view of notes.
85 00:12:47.040 ⇒ 00:12:47.700 Mustafa Raja: Yeah.
86 00:12:51.940 ⇒ 00:13:03.589 Uttam Kumaran: So if everyone has looked at this, or you can just follow along. Basically, this is the output from the agent for Arctures. So I don’t know who. I don’t know, Matthew. If this is a recent account, or
87 00:13:03.750 ⇒ 00:13:05.500 Uttam Kumaran: it’s it’s live right now.
88 00:13:05.500 ⇒ 00:13:10.769 Uttam Kumaran: So I would love even now for you to just like give like off the cuff feedback
89 00:13:10.910 ⇒ 00:13:14.150 Uttam Kumaran: on, like what you’re seeing and like.
90 00:13:14.420 ⇒ 00:13:16.999 Uttam Kumaran: I think I have some immediate feedback.
91 00:13:17.210 ⇒ 00:13:22.091 Uttam Kumaran: But the formatting is not gonna be like perfect. But
92 00:13:23.090 ⇒ 00:13:41.649 Uttam Kumaran: I would love to kind of hear a little bit about what you think of the output, and then we probably have a few tweaks to make to kind of improve this. But also I also want to hear things that you would have loved to have had as part of your slide overviews that maybe, now that we have this AI solution, we can add
93 00:13:41.750 ⇒ 00:13:46.600 Uttam Kumaran: whether it’s broader designs, feedback or or anything else.
94 00:13:47.980 ⇒ 00:13:52.189 Matthew Good: Yeah, I’m just looking at this right now. Yeah, I think definitely the formatting in notion.
95 00:13:52.860 ⇒ 00:13:53.640 Uttam Kumaran: Yeah.
96 00:13:53.640 ⇒ 00:13:57.690 Matthew Good: Is it? Gonna take more time now I have to go through individually like bullet these out.
97 00:13:57.690 ⇒ 00:14:02.809 Uttam Kumaran: No? Well, we can. We’ll we’ll modify the formatting to. Do that. Yeah.
98 00:14:02.810 ⇒ 00:14:08.262 Matthew Good: Okay, cool. And and I changed this like font sizing and stuff. So it’s clear like, what’s an eyebrow text? Or what’s a header?
99 00:14:08.490 ⇒ 00:14:09.650 Uttam Kumaran: Okay. Great. Yeah.
100 00:14:11.680 ⇒ 00:14:14.979 Matthew Good: Slow solution overview performance metrics.
101 00:14:19.450 ⇒ 00:14:24.690 Matthew Good: I guess I’m showing future. Yes, it’s pretty good technical balance mix.
102 00:14:26.630 ⇒ 00:14:36.233 Matthew Good: yeah, definitely font sizing and formatting. And I I’m still like, I’m okay with doing that like, there might just be like a world which I have to do a little bit of it manually.
103 00:14:37.730 ⇒ 00:14:40.570 Matthew Good: performance metrics, tax gains.
104 00:14:42.160 ⇒ 00:14:48.420 Matthew Good: Yeah. And then rationale. Just kind of putting that on the bottom like underneath. Is how we’re doing.
105 00:14:48.420 ⇒ 00:14:53.350 Uttam Kumaran: Can you talk to me about the rationale like? What else would you want to see there? Because right now it is pretty short.
106 00:14:53.540 ⇒ 00:14:54.590 Matthew Good: Yeah.
107 00:14:57.000 ⇒ 00:14:58.592 Uttam Kumaran: Like, would you want to see? Like,
108 00:14:59.810 ⇒ 00:15:01.509 Uttam Kumaran: yeah, it’s all it’s almost like.
109 00:15:02.110 ⇒ 00:15:08.639 Uttam Kumaran: imagine you were to have an intern do this like, what would you want them to say like, why did we decide on these things right like.
110 00:15:08.640 ⇒ 00:15:09.968 Matthew Good: I think right now, it’s like
111 00:15:10.670 ⇒ 00:15:14.039 Matthew Good: it’s not less. This is like on Slide 7.
112 00:15:14.040 ⇒ 00:15:14.450 Uttam Kumaran: Yeah.
113 00:15:14.450 ⇒ 00:15:17.550 Matthew Good: Rational is not a rationale. It’s just like a
114 00:15:17.850 ⇒ 00:15:26.779 Matthew Good: right like, of course, it’s going to be clear competitive differentiation, because that’s the title of the slide, right competitive landscape as opposed to. I would want like, Oh.
115 00:15:27.250 ⇒ 00:15:31.159 Matthew Good: you know, differentiating from XY. And Z.
116 00:15:31.270 ⇒ 00:15:34.449 Matthew Good: For this reason, or something like that like highlight, like a.
117 00:15:34.450 ⇒ 00:15:34.900 Uttam Kumaran: Yeah.
118 00:15:34.900 ⇒ 00:15:38.089 Matthew Good: Flushed out as opposed to just like, what is this slide?
119 00:15:40.700 ⇒ 00:15:51.449 Matthew Good: And then, in terms of like, Yeah, rationale, I also, I’ve been doing because I’m I was doing some more work on insight health, which is like another client from yesterday, and one of the stuff one of the things I’m starting to have
120 00:15:51.940 ⇒ 00:16:00.190 Matthew Good: Claude do is just include, like a summary of like narrative flow, and like why each piece is ordered the way it is.
121 00:16:00.190 ⇒ 00:16:00.720 Uttam Kumaran: Great.
122 00:16:01.870 ⇒ 00:16:09.523 Matthew Good: so that could be helpful as well. So like, you know. Why are we? Why are we? Why is the story ordered in this way,
123 00:16:10.520 ⇒ 00:16:15.383 Matthew Good: and that can be like it doesn’t be at after each specific slide. But at the bottom.
124 00:16:18.470 ⇒ 00:16:27.050 Matthew Good: yeah, those are the 2 main things that come to mind. But all the pieces are here right like, you gotta have vision, opportunity segments, team, and eventually, like, there’s gonna be a world in which.
125 00:16:27.230 ⇒ 00:16:35.987 Matthew Good: you know, we’re gonna probably have one for like seed decks that looks different than like a series. B deck, you’re just like you just are talking about very different things.
126 00:16:37.070 ⇒ 00:16:48.779 Matthew Good: and I have to. I’m I’m like as I’m like also still in clawed like I’ve now built out different prompts, for, like seed. Vc. Deck versus series B versus series a. As we have more and more data.
127 00:16:48.960 ⇒ 00:16:49.340 Uttam Kumaran: Cool.
128 00:16:49.340 ⇒ 00:16:50.649 Matthew Good: That’s get completed.
129 00:16:52.020 ⇒ 00:16:58.870 Uttam Kumaran: Okay, that’s exactly where I also see this going is like, now that we have the framework. And again, the meat of this is like.
130 00:16:59.040 ⇒ 00:17:18.990 Uttam Kumaran: can we pull the files in? Can we write to notion? And so those were, we’re basically in the clear on like, I think that that one. I wanna well, we could talk right after this about where we want to land these stuff in notion. The second piece is even hearing your feedback. I think my next question is going to be like.
131 00:17:19.109 ⇒ 00:17:27.690 Uttam Kumaran: let’s say you were like hypothetically, you’re interacting with this bot via slack. And this is a 1st draft of what you get into slack.
132 00:17:28.400 ⇒ 00:17:37.769 Uttam Kumaran: I want. The the thing I want to avoid is that you press one button, and it goes here, and we all expect it to be perfect on the 1st try like. That’s not what
133 00:17:37.940 ⇒ 00:17:44.269 Uttam Kumaran: I don’t. I think there’s always gonna be some need for you to tweak and add some direction. So
134 00:17:44.860 ⇒ 00:17:51.970 Uttam Kumaran: give me a sense of like of what that direction could look like. You know. Could it be as wide as like.
135 00:17:52.290 ⇒ 00:17:57.150 Uttam Kumaran: just like, give me a Co. Like, try it again. Could it be more of like.
136 00:17:57.330 ⇒ 00:17:58.190 Matthew Good: Yeah, okay.
137 00:17:58.190 ⇒ 00:18:03.729 Uttam Kumaran: Like 7 is wrong, like what were what are some examples of feedback that you can see yourself giving.
138 00:18:03.900 ⇒ 00:18:05.816 Matthew Good: Yeah, I I see what you’re saying.
139 00:18:06.320 ⇒ 00:18:15.870 Matthew Good: I would probably say, give me a little bit more detailed rationale, or like, based on this like, and your knowledge as like a.
140 00:18:15.870 ⇒ 00:18:16.480 Uttam Kumaran: Yeah.
141 00:18:16.480 ⇒ 00:18:27.611 Matthew Good: You know, as a C stage investor, or whatever. Give me a couple of bullets on the rationale for this order, right? So I wanna be able to send that to the client. And then,
142 00:18:30.780 ⇒ 00:18:34.280 Matthew Good: I’m just looking back to this. Go to market market entry, plan, focus, drug
143 00:18:34.952 ⇒ 00:18:40.279 Matthew Good: and then, yeah, expand on the rationale for each slide. And then sometimes, if I like.
144 00:18:40.470 ⇒ 00:18:58.622 Matthew Good: honestly have more time, I’ll say, like, Okay, now go back through this and like strong steel man, the case against like this company, right or or like, come up with questions that like, if you were a venture investor that you’d be asking right like, Where where are the holes in this? So definitely I’ll do that as well.
145 00:18:59.170 ⇒ 00:19:01.830 Matthew Good: And then I’ll I’ll present those to the client, and be like, hey, like.
146 00:19:01.830 ⇒ 00:19:02.310 Uttam Kumaran: Perfect.
147 00:19:02.310 ⇒ 00:19:05.119 Matthew Good: Take a look at that. These are 5 counter questions.
148 00:19:05.600 ⇒ 00:19:06.179 Uttam Kumaran: Oh, yeah.
149 00:19:09.620 ⇒ 00:19:16.240 Matthew Good: solution performance. But yeah, like all the core pieces are here. That I think the main thing is
150 00:19:16.969 ⇒ 00:19:31.490 Matthew Good: yeah, just like the formatting that I kind of play around with it and and showcase because people are very visual. So when they see a notion page, they’re like, what am I looking at? But if they can see visually the hierarchy of like, okay, that’s the headline. That’s the yeah. The bullets. Okay? Great
151 00:19:32.600 ⇒ 00:19:43.650 Uttam Kumaran: Do you care about citations or references? Like in 2 ways like, do you? We can do citations back to like where we sourced it from, for example.
152 00:19:44.119 ⇒ 00:19:48.310 Uttam Kumaran: If there’s if there’s something like this, and it’s from the transcript.
153 00:19:48.450 ⇒ 00:19:52.399 Uttam Kumaran: And would you want to see like this was sourced from
154 00:19:52.640 ⇒ 00:19:55.820 Uttam Kumaran: this line of the transcript like, would that be helpful.
155 00:19:55.820 ⇒ 00:20:06.970 Matthew Good: Yeah, wherever it’s sourced from, whether it’s like transcript old deck or whatever. So I can go back through and check like, oh, is this number, actually, right? Or whatnot, even like 13, right? Like
156 00:20:07.820 ⇒ 00:20:11.269 Matthew Good: seeking series, a funding, clear use of funds and milestones like
157 00:20:11.520 ⇒ 00:20:19.889 Matthew Good: that’s just like we need to actually talk about what that is and be like. If that they don’t know yet we could just say like, this is a placeholder, and I could tell them like, you gotta figure that stuff out, and then we’ll add.
158 00:20:19.890 ⇒ 00:20:23.610 Uttam Kumaran: That’s a good example, like, let’s say it’s like, my guess is that
159 00:20:24.110 ⇒ 00:20:32.290 Uttam Kumaran: I mean, we’ll see like my guess is is that it probably wasn’t there. And it probably just like, build this out. And maybe there isn’t. Wasn’t a number discussed or like.
160 00:20:32.290 ⇒ 00:20:32.820 Matthew Good: Yeah.
161 00:20:32.820 ⇒ 00:20:33.830 Uttam Kumaran: And so
162 00:20:34.000 ⇒ 00:20:42.320 Uttam Kumaran: would you rather like in this situation? Would you rather the AI almost say these? The reason why this is kind of
163 00:20:42.520 ⇒ 00:20:45.080 Uttam Kumaran: crappy is because we don’t have information.
164 00:20:45.220 ⇒ 00:20:45.760 Matthew Good: Yeah.
165 00:20:46.570 ⇒ 00:20:57.459 Matthew Good: yeah, totally. That that’s helpful for me to go through and flag. Okay, this is good. This is good. Now let me go like, dig back through the source material and double check what they said here, and or flag it to the client.
166 00:21:02.360 ⇒ 00:21:05.319 Uttam Kumaran: So almost like one idea is also like.
167 00:21:07.470 ⇒ 00:21:12.420 Uttam Kumaran: what information should you like? What are? What are things that you should go? Ask the client for next
168 00:21:13.420 ⇒ 00:21:19.080 Uttam Kumaran: like that. That may be missing, or could be stronger right? If if you take the other way.
169 00:21:19.420 ⇒ 00:21:24.180 Uttam Kumaran: someone would be like, Hey, I produce this, but it would. It would could be much stronger if I had Xyz.
170 00:21:25.870 ⇒ 00:21:33.509 Matthew Good: Yeah, you’re saying, like for the for the AI to say like, well, if we had information, we could beef up these slides, yeah.
171 00:21:33.870 ⇒ 00:21:34.810 Matthew Good: totally.
172 00:21:38.460 ⇒ 00:21:39.940 Matthew Good: Variations.
173 00:21:40.060 ⇒ 00:21:41.180 Matthew Good: Stage.
174 00:21:42.880 ⇒ 00:21:52.999 Uttam Kumaran: Yeah. And if you think about the variations, I, basically, we will either build another agent or we will have sort of prompt injections that change the prompts
175 00:21:53.110 ⇒ 00:22:01.879 Uttam Kumaran: again. Just make it much tighter towards the objective. So as much of that information as we can have. We will start to route, you know a lot more effectively.
176 00:22:02.110 ⇒ 00:22:02.660 Matthew Good: Yeah,
177 00:22:05.490 ⇒ 00:22:32.839 Matthew Good: awesome. And as we, this is like a question as we go forward like, we’re wrapping decks as well right now that we’ve gotten good feedback. And we’re we’re constantly getting better at what we do, is it? What’s the best format to give you like? Basically, hey? Here’s a completed deck that we just did. That’s great. We’d love to like. Add this into the you know. I don’t know what the correct noun is, but, like the benchmark the AI is using as like the right thing, you know.
178 00:22:32.840 ⇒ 00:22:38.941 Uttam Kumaran: Yeah. So the so basically what that the name of it is called is called golden data sheet.
179 00:22:39.610 ⇒ 00:22:48.059 Uttam Kumaran: this is like the the term basically right now, do we have a Google Doc for this Mustafa? Or are we just using this.
180 00:22:51.213 ⇒ 00:22:58.140 Mustafa Raja: so for for this I simply imported the Csv into the brain. Trust.
181 00:22:58.450 ⇒ 00:23:02.665 Uttam Kumaran: Okay, okay, so yeah, this would be the best place to put them.
182 00:23:02.990 ⇒ 00:23:10.280 Matthew Good: Is it? Is it? Okay? If, like, some of these are like, just P, like even like female founders fund, we just wrapped up with them today. The deck was great.
183 00:23:11.216 ⇒ 00:23:12.590 Matthew Good: Should I?
184 00:23:12.590 ⇒ 00:23:20.689 Uttam Kumaran: Yeah. So so my point was that, any of these? This cuts a little bit Swiss cheese right now as much we we can get all
185 00:23:21.120 ⇒ 00:23:22.330 Uttam Kumaran: 3
186 00:23:23.074 ⇒ 00:23:43.400 Uttam Kumaran: and ideally like an example of you. Maybe working with Claude, if that exists, would be great like for part one of the questions I had last week was just like, Hey, we got like some for some of them. We got the questionnaire for someone. We got the Transcript. It would be great to have both. And then yes, the output would be amazing. You can literally just throw in here.
187 00:23:43.400 ⇒ 00:24:02.369 Matthew Good: Okay, perfect. So for so, okay, great. So I’ll add this for an action item for me is to like, fill this out because we have a couple that are a couple of these that are wrapping like I can go get. I can go get like kind of the initial questionnaire transferred with the call is easy to get, and then just the output, if that’s like, okay, and like, Pdf.
188 00:24:02.370 ⇒ 00:24:08.900 Uttam Kumaran: Yeah. But maybe the only other thing I could ask is a why, it’s it.
189 00:24:09.120 ⇒ 00:24:11.210 Uttam Kumaran: It hit. Basically. Yeah.
190 00:24:11.250 ⇒ 00:24:12.130 Matthew Good: Yeah.
191 00:24:12.130 ⇒ 00:24:29.309 Uttam Kumaran: Whether that’s like the client really loved it because it did this, this or it. We loved it because of this, this and this. That’s what’s gonna give us the narrative. To then start to say, like, this thing is good for this reason. And then when we start to judge the responses.
192 00:24:29.310 ⇒ 00:24:29.680 Uttam Kumaran: Yeah.
193 00:24:29.680 ⇒ 00:24:36.730 Uttam Kumaran: that’s like the way we’re gonna do it. So if you can get me those that’s great, we don’t need like as many as you can give us. But
194 00:24:37.230 ⇒ 00:24:39.330 Uttam Kumaran: yeah, that’d be helpful.
195 00:24:39.330 ⇒ 00:24:42.379 Matthew Good: And I’ll start sorting these 2 by like this was seed.
196 00:24:42.610 ⇒ 00:24:43.860 Matthew Good: This is.
197 00:24:43.860 ⇒ 00:24:50.959 Uttam Kumaran: Oh, yeah, great, great, great. Yeah, you can. You feel free to add whatever call up? Or, yeah.
198 00:24:50.960 ⇒ 00:24:54.485 Uttam Kumaran: that’s gonna like, significantly determine. Like, yeah.
199 00:24:55.830 ⇒ 00:25:07.720 Uttam Kumaran: So you can see you can see my point in that. Now that we start to break this process up, we can only see these variations now, right like, that’s why it’s because you don’t. You may not think of
200 00:25:08.230 ⇒ 00:25:13.260 Uttam Kumaran: these like routes as potential possibilities. But now that we’re breaking down the problem.
201 00:25:13.440 ⇒ 00:25:19.390 Uttam Kumaran: As much dimensionality that we can get allows us to build more use case specific agents.
202 00:25:20.340 ⇒ 00:25:21.290 Matthew Good: Got it?
203 00:25:22.538 ⇒ 00:25:34.269 Uttam Kumaran: So I think probably the last. So I think it’s clear on what kind of things we would need in terms of human in the loop. I think, Mustafa. What I’m hearing is that we may have some
204 00:25:34.733 ⇒ 00:25:41.299 Uttam Kumaran: loops that we do, and at some point the team needs to be able to say, cool, this is approved. Let’s get into notion.
205 00:25:42.390 ⇒ 00:25:45.610 Uttam Kumaran: That seems accurate. Right.
206 00:25:45.970 ⇒ 00:25:46.520 Mustafa Raja: Yep.
207 00:25:47.458 ⇒ 00:25:52.870 Uttam Kumaran: And then basically, the AI should return with a the notion link of of where it landed.
208 00:25:53.720 ⇒ 00:25:54.560 Mustafa Raja: Yeah.
209 00:25:57.970 ⇒ 00:26:06.126 Uttam Kumaran: So I feel good about this so far, I think maybe if we want to spend the rest of the time talking about notion, and where
210 00:26:06.770 ⇒ 00:26:07.420 Matthew Good: Yeah.
211 00:26:07.730 ⇒ 00:26:09.440 Uttam Kumaran: Yeah, where? Where we wanna.
212 00:26:09.440 ⇒ 00:26:13.809 Rafay Iqbal: Yeah, I know Andrew has some input as well. So I guess feel free to ask for Andrew.
213 00:26:14.020 ⇒ 00:26:15.400 Andrew Del Rizzo: Yeah, so
214 00:26:17.037 ⇒ 00:26:20.779 Uttam Kumaran: Any questions about the project. Too. Sorry, I know we just jumped into this, so in case you.
215 00:26:20.780 ⇒ 00:26:26.846 Andrew Del Rizzo: No, no, it’s okay. I think all that is like, mostly on like Matt side like that was like a lot of foreign stuff to me.
216 00:26:27.860 ⇒ 00:26:44.939 Andrew Del Rizzo: I don’t really deal with all that. But the project management side, so I like I haven’t I haven’t really changed anything on the notion itself. But I was like thinking about some workflows that could be good. I was gonna share my screen. It might not make sense. I’ll try my best to explain. Like.
217 00:26:45.850 ⇒ 00:26:48.259 Andrew Del Rizzo: how I could imagine this working.
218 00:26:48.735 ⇒ 00:26:52.449 Andrew Del Rizzo: I I’m the only one that uses this like timeline view right now.
219 00:26:52.570 ⇒ 00:26:56.819 Andrew Del Rizzo: But I just wanna show it like this, because the main thing is
220 00:26:57.286 ⇒ 00:27:14.590 Andrew Del Rizzo: setting up dependencies for projects like. For when we get a project like we intake it. So say, for example, I set this up right now for this one project called like this Chat feature video. And this is like the the roots of the like. The project. So I guess, like the
221 00:27:14.710 ⇒ 00:27:20.289 Andrew Del Rizzo: the hub that encases, like all the subtasks beneath it, so like within, that you have, like
222 00:27:20.430 ⇒ 00:27:45.379 Andrew Del Rizzo: the storyboards that come first, st then the version one of the video, and then the final video. And I set this all up manually like, just because, like one’s basically depend on the other. So like one has to finish. And then we can get to the next. And then we can get to the next. And this works well, because even with like other people’s tasks, say, for example, this one here, it’s a branding task that comes before we do a website.
223 00:27:45.380 ⇒ 00:27:45.760 Uttam Kumaran: Yes.
224 00:27:45.760 ⇒ 00:27:54.359 Andrew Del Rizzo: And again, like this is all set up manually. But I was just wondering if there’d be like a way to automate this where, when we get like a client, come in.
225 00:27:54.850 ⇒ 00:28:20.169 Andrew Del Rizzo: we can establish. Okay, this is a web design project, which means, then, if we are approved to do brand, and we start with mood boards, and then for mood boards we move to like a second round of mood boards, and then we move to a 1st round of brand, then a second round of brand, and then a 1st round of web design, basically like stack it all up, but have them be subtask rather than be tasks that aren’t linked to each other because that’s kind of what we had before was.
226 00:28:20.250 ⇒ 00:28:30.069 Andrew Del Rizzo: We’d have individual tasks for each thing, but they wouldn’t link together so like so like sometimes it would get missed if, like one thing connects to each other like not missed, but like it’s nice to know.
227 00:28:30.070 ⇒ 00:28:34.690 Uttam Kumaran: You need the dependencies like, you need to know that one thing blocks another thing. And so exactly, yeah.
228 00:28:34.690 ⇒ 00:28:37.424 Uttam Kumaran: So I guess my question is
229 00:28:40.020 ⇒ 00:28:46.580 Uttam Kumaran: like you. You mentioned that it’s fairly structured. What the subtasks are based on the type of project.
230 00:28:46.950 ⇒ 00:28:49.870 Andrew Del Rizzo: Yeah, yeah, yeah, we we right now.
231 00:28:50.250 ⇒ 00:29:02.600 Andrew Del Rizzo: and Rafai or Matt, you guys can chime into. But right now, I’m just basing them off of the Project board that I believe this gets auto generated. I’m not sure if these are like fixed due dates or anything. But yeah, basically, we’re
232 00:29:04.160 ⇒ 00:29:05.149 Andrew Del Rizzo: sorry. Say, again.
233 00:29:05.150 ⇒ 00:29:09.040 Matthew Good: That gets auto generated this timeline situation. Yeah, right?
234 00:29:09.040 ⇒ 00:29:17.089 Andrew Del Rizzo: Yeah. But but it’s pretty much this, it’s basically having this project. Timeline be separated into tasks where we have
235 00:29:17.310 ⇒ 00:29:27.069 Andrew Del Rizzo: like, I don’t know if this one would have to be a task. But we have, like the kickoff, mean discovery. But then the next task is mood boards, and then we have brown brand round one brand round 2, brand round, 3.
236 00:29:27.070 ⇒ 00:29:31.010 Uttam Kumaran: And are these tasks right now? Like if you click on name or those pages.
237 00:29:31.572 ⇒ 00:29:41.239 Andrew Del Rizzo: I think they’re just they’re just names. Yeah, I don’t think they lead to anything. It’s just it’s just to for these to be set up then, in their own task, so like.
238 00:29:41.240 ⇒ 00:29:47.899 Uttam Kumaran: The dependencies are all fairly easy, like r, 2 comes before comes after r, 1.
239 00:29:47.900 ⇒ 00:30:10.670 Andrew Del Rizzo: Yeah, there’s there’s no like. It’s not like 2 things need to be done before one thing is done. It’s it’s literally just like a linear flow. We’re just this thing needs to be completed before this before this, before this. It’s just one line pretty much like it’s it’s not. It’s nothing crazy or anything. That’s why. But yeah, it’s basically just I, I want to know if we can get this
240 00:30:10.870 ⇒ 00:30:16.770 Andrew Del Rizzo: into tasks like, just cut out that, like middle portion, return this into a task.
241 00:30:16.900 ⇒ 00:30:23.119 Uttam Kumaran: So the ideal state is that a new client gets created in like if you go back
242 00:30:23.600 ⇒ 00:30:30.779 Uttam Kumaran: one page and maybe 2, if you go like down this tree, like wherever to the main client area is.
243 00:30:30.780 ⇒ 00:30:38.919 Andrew Del Rizzo: Oh, that would that would be I believe it’s here from I’m not 100% sure how it works. I I know Matt and Rafael like set this up, but.
244 00:30:38.920 ⇒ 00:30:42.016 Uttam Kumaran: My question is like, if you go back here
245 00:30:42.980 ⇒ 00:30:48.480 Uttam Kumaran: and like, if you click these 3 dots, is there like a main clients? I’m like, up here. Yeah.
246 00:30:48.670 ⇒ 00:30:49.830 Andrew Del Rizzo: Oh, oh, up there. Yeah.
247 00:30:51.043 ⇒ 00:30:54.670 Uttam Kumaran: Yeah, if you go to like interlude client board.
248 00:30:55.170 ⇒ 00:30:58.749 Uttam Kumaran: So my thinking is like you add a client here.
249 00:30:58.960 ⇒ 00:31:01.410 Uttam Kumaran: and then you put the type. And then
250 00:31:01.620 ⇒ 00:31:04.239 Uttam Kumaran: there’s probably a bunch of things that kick off. Is that correct?
251 00:31:04.240 ⇒ 00:31:04.870 Andrew Del Rizzo: Yes.
252 00:31:05.298 ⇒ 00:31:08.300 Matthew Good: Built this automation. I I’m pretty sure.
253 00:31:08.300 ⇒ 00:31:15.329 Rafay Iqbal: Yeah, it’s like, if I have a sales. I have a sales funnel sheet. So once a sales client closes, and I have like a deck
254 00:31:15.747 ⇒ 00:31:20.470 Rafay Iqbal: attribution applied to it. It’ll automate into this client board, then this will automate into another board.
255 00:31:20.470 ⇒ 00:31:20.980 Uttam Kumaran: Yeah, yeah.
256 00:31:20.980 ⇒ 00:31:23.070 Rafay Iqbal: Of automations I have built into this. Yeah.
257 00:31:23.070 ⇒ 00:31:27.080 Uttam Kumaran: Okay, did you try to use that to also do the subtasks? I guess.
258 00:31:27.372 ⇒ 00:31:28.250 Rafay Iqbal: No, not yet.
259 00:31:28.250 ⇒ 00:31:28.859 Uttam Kumaran: It’s a lot of.
260 00:31:28.860 ⇒ 00:31:33.180 Rafay Iqbal: That’s where it got a bit more complex. And they kept breaking. So I was like, Oh, I gotta get my hands off this.
261 00:31:33.560 ⇒ 00:31:43.049 Uttam Kumaran: So I think I think 2 things for us to look at one. We’ll look through that second I need to think about. I don’t know if we can do the
262 00:31:43.200 ⇒ 00:31:46.649 Uttam Kumaran: like. I’ll try to. We’ll try to use the notion automations.
263 00:31:46.810 ⇒ 00:31:57.689 Uttam Kumaran: but I’m not sure whether you can do task dependencies. We’ll have to check, I guess, Mustafa, we can play around but ideally like, if generally these are like
264 00:31:57.890 ⇒ 00:32:01.390 Uttam Kumaran: at least standard based on type.
265 00:32:01.790 ⇒ 00:32:13.259 Uttam Kumaran: And then what we can do is try to see whether we can have some default dependencies based on the task. So that’s something we can look at. Is there Andrew, is there a good one to like
266 00:32:13.780 ⇒ 00:32:17.759 Uttam Kumaran: that? I can clone and like try to play around with like is architect, health.
267 00:32:19.380 ⇒ 00:32:19.830 Uttam Kumaran: Like.
268 00:32:19.830 ⇒ 00:32:25.380 Andrew Del Rizzo: I mean, you can clone the the raspberry one. I did, because that’s closest to what I imagine it being like once
269 00:32:26.040 ⇒ 00:32:28.989 Andrew Del Rizzo: is finalized. Yeah.
270 00:32:29.550 ⇒ 00:32:32.831 Uttam Kumaran: Basically, the the reason why is we’ll clone it. And then,
271 00:32:33.830 ⇒ 00:32:38.950 Uttam Kumaran: we will build the automation and then test like what it looks like versus
272 00:32:41.060 ⇒ 00:32:44.029 Uttam Kumaran: What exists in the in real for for raspberry.
273 00:32:44.300 ⇒ 00:32:54.729 Andrew Del Rizzo: Yeah, I guess for context, like that, one would be a video one. Because, like I, I handle most of the video stuff. So I just know like how the flow works for that. Like I did, I did kind of summarize like.
274 00:32:55.370 ⇒ 00:33:08.830 Andrew Del Rizzo: like the storyboards would actually probably be. Sometimes it gets. It’s just one round, or sometimes it’s 2. So I I think I think it would be good to base it off. Also like that, like this, architects project board, just because.
275 00:33:08.990 ⇒ 00:33:14.159 Andrew Del Rizzo: even if we don’t use all 3 rounds, it would just be good if it auto populates those 3 rounds.
276 00:33:14.160 ⇒ 00:33:15.360 Uttam Kumaran: You can remove it.
277 00:33:15.360 ⇒ 00:33:17.390 Andrew Del Rizzo: Exactly. Yeah, exactly. Yeah.
278 00:33:17.680 ⇒ 00:33:27.110 Uttam Kumaran: Okay, yeah, I mean, in typical project management, like in for for our work with our clients, this is this is typically what’s called just like reusable epics where you just have.
279 00:33:27.430 ⇒ 00:33:43.629 Uttam Kumaran: like, for example, building an Ios app. Okay, you need to do this, and you do this, and you do this. And then it’s sort of like a big, reusable chunk of work with dependencies and built based on a type. I think. Don’t think this should be hard for us to do so. We’ll take a look this week.
280 00:33:44.030 ⇒ 00:33:53.049 Andrew Del Rizzo: Yeah, I I it really does look like like Rafey like you you saw. I think everything is like pretty like it’s almost set up like 100%. There. It’s just like getting those.
281 00:33:53.250 ⇒ 00:33:53.730 Rafay Iqbal: Yeah.
282 00:33:53.730 ⇒ 00:33:59.189 Andrew Del Rizzo: On that project board into a subtask. But when I was trying to myself like last night, I was getting ahead.
283 00:33:59.190 ⇒ 00:34:00.020 Uttam Kumaran: No, it’s also.
284 00:34:00.020 ⇒ 00:34:00.560 Andrew Del Rizzo: Really had a head.
285 00:34:00.560 ⇒ 00:34:03.719 Uttam Kumaran: Notion automations is like not the best.
286 00:34:03.720 ⇒ 00:34:04.310 Andrew Del Rizzo: Oh, yeah.
287 00:34:04.310 ⇒ 00:34:09.699 Uttam Kumaran: But it’s like you can do a lot of stuff but you will end up kind of being like.
288 00:34:10.090 ⇒ 00:34:11.420 Uttam Kumaran: I don’t know what I’m doing.
289 00:34:12.499 ⇒ 00:34:14.389 Matthew Good: Yeah, what do you guys
290 00:34:14.389 ⇒ 00:34:19.299 Matthew Good: out of curiosity also want to be? I know we’re over time. if you guys gotta hop no worries.
291 00:34:19.650 ⇒ 00:34:20.420 Uttam Kumaran: Go ahead!
292 00:34:20.429 ⇒ 00:34:24.709 Matthew Good: What do you guys use for 2 things? One. What do you guys use for your own? Like sort of internal?
293 00:34:24.979 ⇒ 00:34:26.629 Matthew Good: This equivalent.
294 00:34:26.630 ⇒ 00:34:28.509 Uttam Kumaran: We use linear
295 00:34:30.130 ⇒ 00:34:32.839 Uttam Kumaran: We were using notion for a while for.
296 00:34:32.840 ⇒ 00:34:33.959 Rafay Iqbal: Is there a bird?
297 00:34:34.600 ⇒ 00:34:35.150 Uttam Kumaran: Right.
298 00:34:35.150 ⇒ 00:34:36.040 Rafay Iqbal: Oh, sorry!
299 00:34:36.710 ⇒ 00:34:37.090 Uttam Kumaran: We.
300 00:34:37.090 ⇒ 00:34:39.323 Rafay Iqbal: Someone’s walking by me. I was like, Yeah, you can go ahead.
301 00:34:39.826 ⇒ 00:35:01.049 Uttam Kumaran: We we we. I wanted to use notion for everything, mainly because I wanted to have one area where I can run AI on like everything, but we found it really complicated for project management. I would say it’s our work is different, though in that I think your work is probably a lot of writing and a lot of linking to stuff ours. Work
302 00:35:01.370 ⇒ 00:35:03.310 Uttam Kumaran: does not happen in
303 00:35:03.763 ⇒ 00:35:25.406 Uttam Kumaran: notion like it’s either code or it’s in another system. So notion is just what we use for project management, basically like, what is the client? And then what are the tickets? Who’s assigned? When does it do? How? What’s the estimation? And then we run standups based on that. Linear is pretty good. I’ve seen people have a lot of success with Clickup as well.
304 00:35:26.180 ⇒ 00:35:29.890 Uttam Kumaran: The thing is, notion is like decent at everything.
305 00:35:30.120 ⇒ 00:35:31.560 Matthew Good: Right? Right? It’s like.
306 00:35:31.560 ⇒ 00:35:46.842 Uttam Kumaran: And that’s busting and a curse as they they’re like, oh, you could do database you could do project manager, and then it kind of just gets to be a lot. But I think, like it’s not like a big grass is greener situation.
307 00:35:47.450 ⇒ 00:35:57.709 Uttam Kumaran: there are other project management tools that that’s what they specialize in linear being one of them. It’s what we use. But we don’t do. We don’t do writing in linear. It’s like we still use notion for a ton of stuff.
308 00:35:59.110 ⇒ 00:36:02.489 Uttam Kumaran: We then link back to the notion, Doc, once once it gets written.
309 00:36:02.960 ⇒ 00:36:18.409 Matthew Good: Okay, got it. I was just curious. Cause. Yeah, there is like that writing and and commenting element. It will play around with linear. The other thing, too, is just as we’re talking about the notion stuff like. I want to be mindful of scope creep so just like shoot me a ping like if if we’re getting out of scope. Here.
310 00:36:18.410 ⇒ 00:36:25.529 Uttam Kumaran: Yeah, I don’t think I think we’ll probably spend an hour or 2 to just look at it. And it’s like, gonna be something massive. I don’t think it should be that bad.
311 00:36:25.530 ⇒ 00:36:28.112 Matthew Good: Okay, awesome. I just wanted to call that out.
312 00:36:28.600 ⇒ 00:36:31.750 Matthew Good: great, I will. I have my homework in terms of filling that stuff out. I’ll get that.
313 00:36:31.750 ⇒ 00:36:48.950 Uttam Kumaran: Yeah. And then, in terms of on our side, we’re gonna we’ll go ahead and we’re gonna get the notion integration set up. So this will start landing in notion, and then we’ll we’ll set up the slack bot, and you’ll see us get Test start testing with that this week. That’ll be great because we can do all that testing Async, in our joint channel.
314 00:36:51.480 ⇒ 00:36:55.360 Uttam Kumaran: And yeah, we’ll follow up on the the notion automation piece.
315 00:36:55.360 ⇒ 00:36:57.909 Matthew Good: Okay, awesome. Thank you. Guys, I know we went over really appreciate it.
316 00:36:57.910 ⇒ 00:37:01.029 Uttam Kumaran: Yeah, yeah, appreciate it. Okay, thank you. Everyone.
317 00:37:01.030 ⇒ 00:37:02.630 Matthew Good: Yeah. Chat soon. Bye, bye.