Meeting Title: Engineering Leads Sync Date: 2025-09-24 Meeting participants: Samuel Roberts, Uttam Kumaran
WEBVTT
1 00:01:13.760 ⇒ 00:01:15.030 Uttam Kumaran: Hello!
2 00:01:17.300 ⇒ 00:01:19.020 Samuel Roberts: Hey, sorry, I was just finishing up lunch.
3 00:01:19.270 ⇒ 00:01:21.270 Uttam Kumaran: All good. What’s for lunch?
4 00:01:22.450 ⇒ 00:01:23.730 Samuel Roberts: Made myself a little sandwich.
5 00:01:24.190 ⇒ 00:01:24.980 Uttam Kumaran: Nice.
6 00:01:25.350 ⇒ 00:01:25.890 Samuel Roberts: Yeah.
7 00:01:27.750 ⇒ 00:01:28.840 Samuel Roberts: How’s it going today?
8 00:01:29.870 ⇒ 00:01:37.580 Uttam Kumaran: It is good, man. We… looks like we got… we got the verbal approval for the default renewal. We had also another contract just closed.
9 00:01:37.990 ⇒ 00:01:44.839 Uttam Kumaran: So, solid, solid day today. I’m very tired, but…
10 00:01:44.840 ⇒ 00:01:45.780 Samuel Roberts: I bet.
11 00:01:48.140 ⇒ 00:01:51.300 Samuel Roberts: I just had that, AI…
12 00:01:52.030 ⇒ 00:01:55.930 Uttam Kumaran: Oh, yeah, how’d it go? Damn, I was bummed I wasn’t able to go. Was it okay?
13 00:01:55.930 ⇒ 00:02:01.299 Samuel Roberts: Yeah, I think they’re gonna send out, the recording and the slides the guy had. It was really interesting, they did some…
14 00:02:01.600 ⇒ 00:02:05.800 Samuel Roberts: like, actual studies of how, like, AI is…
15 00:02:07.190 ⇒ 00:02:10.709 Samuel Roberts: How useful it is in certain scenarios.
16 00:02:10.960 ⇒ 00:02:13.020 Samuel Roberts: For engineers.
17 00:02:13.210 ⇒ 00:02:14.160 Samuel Roberts: And I think
18 00:02:15.410 ⇒ 00:02:21.950 Samuel Roberts: somewhat surprising, but also, like, if you think about it, it makes sense. Like, just, like, context building in a project in general.
19 00:02:22.060 ⇒ 00:02:24.820 Samuel Roberts: Kind of slows things down. It’s very,
20 00:02:25.760 ⇒ 00:02:30.460 Samuel Roberts: Volatile, like, week to week, they were seeing, like, how much it was helping on certain tasks.
21 00:02:31.920 ⇒ 00:02:38.939 Samuel Roberts: you know, obviously having access to, like, Beyond Free Tier was a big thing, too, because they were comparing, like, enterprise versus startup stuff, or, like, startup teams.
22 00:02:41.410 ⇒ 00:02:48.370 Samuel Roberts: I’m trying to remember. It’s not, like, overall not great with legacy code, much better for, like, new generation.
23 00:02:49.330 ⇒ 00:02:53.830 Samuel Roberts: Which I think I’ve seen a little bit, it makes sense, because you can only give it so much context of the whole thing.
24 00:02:54.150 ⇒ 00:02:55.800 Samuel Roberts: But yeah.
25 00:02:55.990 ⇒ 00:02:57.589 Samuel Roberts: Hopefully they’ll send a recording.
26 00:02:59.930 ⇒ 00:03:01.020 Uttam Kumaran: Okay, perfect.
27 00:03:01.020 ⇒ 00:03:03.669 Samuel Roberts: But I also have, like, granola-generated notes if you wanna…
28 00:03:10.020 ⇒ 00:03:10.810 Samuel Roberts: Yeah.
29 00:03:11.940 ⇒ 00:03:15.530 Uttam Kumaran: Great. Nope.
30 00:03:17.670 ⇒ 00:03:21.889 Uttam Kumaran: Yeah, I guess, like, I would love to use this meeting to maybe confirm
31 00:03:22.520 ⇒ 00:03:25.120 Uttam Kumaran: Okrs, do you want to go through that?
32 00:03:25.960 ⇒ 00:03:33.260 Samuel Roberts: Sure, yeah, I’ve been digging into some things, but, yeah, I’m…
33 00:03:34.090 ⇒ 00:03:38.620 Samuel Roberts: I was playing around with some ideas for the AI leverage, Index.
34 00:03:40.400 ⇒ 00:03:44.200 Uttam Kumaran: Yeah, and I also have some ideas on that, too, so you go first.
35 00:03:44.370 ⇒ 00:03:50.630 Samuel Roberts: Yeah, okay, so I, I mean, I had, like, a rough idea. I went back and forth with Track GPT a little bit,
36 00:03:50.990 ⇒ 00:03:59.530 Samuel Roberts: Part of my… Uncertainty right now is, like, should this be, like, a… measure of…
37 00:04:00.840 ⇒ 00:04:08.490 Samuel Roberts: The leverage thing is what’s throwing me out. Like, how to measure, like, what people are currently doing well, or, like, at… measuring that well.
38 00:04:09.000 ⇒ 00:04:12.420 Samuel Roberts: to compare, is…
39 00:04:13.760 ⇒ 00:04:18.970 Samuel Roberts: you know, getting time estimates and things, like, obviously we’re working on that with, like, linear and better, but…
40 00:04:19.250 ⇒ 00:04:22.420 Samuel Roberts: Historically, I’m not sure what we have for that kind of stuff.
41 00:04:23.450 ⇒ 00:04:28.750 Samuel Roberts: My thought is, if we want to go that route, Mapping out the workflows?
42 00:04:29.050 ⇒ 00:04:30.030 Samuel Roberts: will help.
43 00:04:32.660 ⇒ 00:04:40.370 Samuel Roberts: Because we can get some rough time estimates to do a rough calculation. I’m a little bit hesitant about the time aspect, just because I know that can be
44 00:04:41.470 ⇒ 00:04:43.829 Samuel Roberts: Noisy, probably?
45 00:04:44.870 ⇒ 00:04:48.789 Samuel Roberts: But then I was just looking… I was thinking about other ways to do it without that,
46 00:04:48.990 ⇒ 00:04:52.640 Samuel Roberts: just, like, general AI-assisted tasks,
47 00:04:53.680 ⇒ 00:04:56.660 Samuel Roberts: Like, task proportion rather than a time proportion.
48 00:04:56.780 ⇒ 00:05:03.669 Samuel Roberts: My other thought was, as we start getting more stuff into
49 00:05:04.620 ⇒ 00:05:12.659 Samuel Roberts: most talk, potentially, we can store, like, who’s making calls, at least from the platform. We can pull in the cursor data, I was playing with that API a little bit.
50 00:05:16.200 ⇒ 00:05:23.079 Samuel Roberts: I’m just not… I’m not 100% sure, like, which way to go, because I feel like there’s a few different avenues here, and all of them have their own complexities, you know?
51 00:05:23.080 ⇒ 00:05:24.699 Uttam Kumaran: I guess, like, my… I guess…
52 00:05:24.700 ⇒ 00:05:25.530 Samuel Roberts: Either, so…
53 00:05:25.530 ⇒ 00:05:32.100 Uttam Kumaran: Yeah, I guess my question would be, put yourself in your old role, Like, what do you think…
54 00:05:32.740 ⇒ 00:05:34.920 Uttam Kumaran: AKA, put yourself in my role, like, what do you.
55 00:05:34.920 ⇒ 00:05:35.919 Samuel Roberts: Yeah, of course.
56 00:05:35.920 ⇒ 00:05:36.970 Uttam Kumaran: we should do.
57 00:05:40.450 ⇒ 00:05:45.319 Samuel Roberts: Yeah, no, that’s exactly how it is. I mean, my gut is that I don’t like time estimates, because…
58 00:05:47.140 ⇒ 00:05:50.319 Samuel Roberts: it’s, again, noisy. So I’m thinking if we…
59 00:05:53.460 ⇒ 00:06:00.229 Samuel Roberts: My issue then becomes tracking tasks, so if we can use the linear tickets and do something with, like, was AIUs, that’s a…
60 00:06:00.400 ⇒ 00:06:03.999 Samuel Roberts: A signal, in addition to the cost of the platform, and…
61 00:06:04.390 ⇒ 00:06:06.350 Samuel Roberts: or… .
62 00:06:10.170 ⇒ 00:06:13.840 Uttam Kumaran: I mean, my other… my other thing you could do is, like.
63 00:06:13.990 ⇒ 00:06:15.460 Uttam Kumaran: I mean, this is where, like.
64 00:06:15.690 ⇒ 00:06:19.069 Uttam Kumaran: one of the things I was talking to folks about is, like.
65 00:06:19.650 ⇒ 00:06:26.300 Uttam Kumaran: How should… maybe we should tie the AI team to, like, financial metrics.
66 00:06:26.490 ⇒ 00:06:30.620 Uttam Kumaran: like… Maybe your optimization should be…
67 00:06:31.750 ⇒ 00:06:39.050 Uttam Kumaran: like, revenue per employee, basically. Or… or, like, or it should be the fact that
68 00:06:39.430 ⇒ 00:06:47.990 Uttam Kumaran: for example, I think AI is… AI is going to be responsible for us to get 5 or 10 points of margin on delivery.
69 00:06:48.290 ⇒ 00:06:54.620 Uttam Kumaran: like, for example, my OKR for the delivery team is, hit 40% margin.
70 00:06:54.860 ⇒ 00:07:06.860 Uttam Kumaran: But I think if you guys help them, we could probably get to 50. And here’s how. One of the conversations I had with Justin yesterday, and I’m gonna be putting up, like, a…
71 00:07:07.040 ⇒ 00:07:09.859 Uttam Kumaran: sort of write-up on this, is I talked to them about, like, how does…
72 00:07:10.150 ⇒ 00:07:13.580 Uttam Kumaran: how do, how does PM,
73 00:07:14.470 ⇒ 00:07:22.489 Uttam Kumaran: how does PM, like, what is it called, allocation work? You know, and I was like… and my question was, like.
74 00:07:22.790 ⇒ 00:07:29.950 Uttam Kumaran: Are they… are you allocating them by seniority? Are you allocating them based on, like, the number of… of,
75 00:07:30.290 ⇒ 00:07:35.740 Uttam Kumaran: clients they can handle? And his answer was, like, from his experience, people have done
76 00:07:36.210 ⇒ 00:07:40.240 Uttam Kumaran: People have allocated in sort of a way where it,
77 00:07:41.610 ⇒ 00:07:55.639 Uttam Kumaran: every client is assigned a set of points, and everybody has a max set of 100 points. For example, you have some clients that are 25 points, you have some clients that are 50 points, and some clients that are 100 points. And so you can have 4 25-point accounts.
78 00:07:55.640 ⇒ 00:08:03.350 Uttam Kumaran: you can have 250 point, or you can have 100-point, right? And the points are based on complexity.
79 00:08:03.470 ⇒ 00:08:04.829 Uttam Kumaran: Right.
80 00:08:05.200 ⇒ 00:08:11.860 Uttam Kumaran: Right? So, one of the things that I talked to him about is, like, great, so that implies that basically.
81 00:08:12.550 ⇒ 00:08:17.859 Uttam Kumaran: for the… if most of our accounts right now are… I think most of our accounts are 25 to 50 points.
82 00:08:18.180 ⇒ 00:08:23.920 Uttam Kumaran: Then what I… what I asked him is, like, okay, what would you need
83 00:08:24.120 ⇒ 00:08:31.429 Uttam Kumaran: to see, to get, to basically be able to reduce the points by 5.
84 00:08:31.550 ⇒ 00:08:32.280 Uttam Kumaran: Like, can.
85 00:08:32.289 ⇒ 00:08:32.979 Samuel Roberts: Can you see?
86 00:08:32.980 ⇒ 00:08:37.380 Uttam Kumaran: their smallest account, can it go from 25 to 20?
87 00:08:37.809 ⇒ 00:08:38.280 Uttam Kumaran: Right.
88 00:08:38.280 ⇒ 00:08:38.840 Samuel Roberts: True.
89 00:08:39.289 ⇒ 00:08:45.689 Uttam Kumaran: Like, so what would… what would you need from the AI team in order to start having people take on 5 accounts?
90 00:08:45.929 ⇒ 00:09:03.689 Uttam Kumaran: Right? Instead of just 4. Maybe someone can take on… maybe it moves from 50 to 33, or 50 to 40, right? So, like, I guess that’s where almost my… my thing is, like, maybe the AI team should be tasked with
91 00:09:04.289 ⇒ 00:09:09.589 Uttam Kumaran: helping… like, for me, if I was to play pure CFO,
92 00:09:09.639 ⇒ 00:09:18.849 Uttam Kumaran: My ROI is, it needs to either affect our ability to sell more cheaper, right? So, like, we need to see Justina, myself.
93 00:09:18.849 ⇒ 00:09:31.339 Uttam Kumaran: people on the sales team using it a lot to get us to scale, and also preventing additional sales fires. Additionally, I see it on the delivery side, right? Because those are the two areas where our spend is the highest.
94 00:09:31.379 ⇒ 00:09:43.399 Uttam Kumaran: And our opportunity, therefore, is also the highest, where if we were to get 5 points of margin on the delivery side, that is… that is huge, because I have to go hire another project manager to go do that.
95 00:09:43.489 ⇒ 00:09:49.619 Uttam Kumaran: And so, part of me is like, okay, maybe I should give the AI team an OKR around
96 00:09:49.739 ⇒ 00:09:52.209 Uttam Kumaran: Margin, which is, like.
97 00:09:53.179 ⇒ 00:10:02.709 Uttam Kumaran: Yeah, or somehow tie it to financial, because then it’s… it’s pretty pure, like, we need to find a way to attribute it, versus, like.
98 00:10:03.739 ⇒ 00:10:06.929 Uttam Kumaran: And I just sort of give you that… give you that as, like, a blanket.
99 00:10:08.199 ⇒ 00:10:13.819 Uttam Kumaran: you know, thing. Or, or again, the derivative of that is, like, if the derivative to getting the margin is…
100 00:10:14.019 ⇒ 00:10:22.289 Uttam Kumaran: making sure that PMs can take an additional client, or, like, The, for example.
101 00:10:22.409 ⇒ 00:10:32.919 Uttam Kumaran: I was also asked him, I said, okay, so then what’s the difference between junior… he had junior, senior, and principal PMs. He was like, well, the junior folks, they typically have 4 25s.
102 00:10:32.979 ⇒ 00:10:35.329 Uttam Kumaran: Like, meaning they’re taking on…
103 00:10:35.369 ⇒ 00:10:45.479 Uttam Kumaran: more, like, distinct clients, but, like, less complex ones. Whereas the mid-levels, they typically have, like, a 50 and two 25s, or, like, 250s.
104 00:10:45.479 ⇒ 00:10:58.139 Uttam Kumaran: And the principal people are typically going after the biggest things. But he’s also, like, some… some clients, for example, Insomnia Cookies right now is, like, a 25, but it could become a 50 and a 100 because of, like.
105 00:10:58.509 ⇒ 00:11:06.539 Uttam Kumaran: it could… like, the potential is there. So then, you don’t want to staff a junior person on insomnia, you want to staff a senior person.
106 00:11:06.619 ⇒ 00:11:26.209 Uttam Kumaran: But again, I think it’s… it’s… this is just one way of, like, attributing some type of, like, either we increase the max capacity, right, so now we move from, like, people have 100 points to people have 125 points, or you… or you just decrease the points per segment. Either way.
107 00:11:26.319 ⇒ 00:11:34.559 Uttam Kumaran: the goal is to get more capacity per project manager, right? So, like, that could be another thing for you guys to take on and optimize.
108 00:11:34.999 ⇒ 00:11:38.569 Uttam Kumaran: That is a little bit more… specific.
109 00:11:39.020 ⇒ 00:11:44.890 Samuel Roberts: Yeah, that, that is helpful, like, that, or even the financial idea, like, something that is a little more…
110 00:11:46.250 ⇒ 00:11:55.529 Samuel Roberts: tangible, like, not that the AI stuff isn’t measurable, but the, like… I can get, like, a raw number of, like, how much AI usage kind of thing, but that doesn’t really… it’s unanchored to something.
111 00:11:55.810 ⇒ 00:11:56.680 Samuel Roberts: bomb.
112 00:11:57.090 ⇒ 00:12:01.739 Samuel Roberts: Applying that back to… A real, like.
113 00:12:03.190 ⇒ 00:12:08.009 Samuel Roberts: real high-value KPI, like financial or the PM stuff, might…
114 00:12:09.370 ⇒ 00:12:11.100 Samuel Roberts: Push us in the right direction, then.
115 00:12:11.670 ⇒ 00:12:17.480 Samuel Roberts: more than just, like, more AI, you know? I think that might be… a better…
116 00:12:21.310 ⇒ 00:12:23.679 Samuel Roberts: Sorry, my gears are turning now. Okay.
117 00:12:24.730 ⇒ 00:12:30.419 Uttam Kumaran: So, like, I mean, let’s… let’s… let’s take a… if I was to just… if we were to do this exercise, like.
118 00:12:30.930 ⇒ 00:12:35.819 Uttam Kumaran: you know, really quick. I would… and I would sort of attribute stuff here, I would say default…
119 00:12:36.150 ⇒ 00:12:42.369 Uttam Kumaran: like, I would say Eden is, like, a 50, Insomnia’s, like, a 50, Interlude’s a 25.
120 00:12:43.530 ⇒ 00:12:44.649 Uttam Kumaran: 20th… oh.
121 00:12:46.160 ⇒ 00:12:57.799 Uttam Kumaran: 25… 25… 50… 50… 25… And let’s just say 25.
122 00:12:58.150 ⇒ 00:13:04.850 Uttam Kumaran: Right, so this is a total… 375.
123 00:13:05.390 ⇒ 00:13:07.530 Samuel Roberts: Right, so if we have…
124 00:13:08.000 ⇒ 00:13:14.460 Uttam Kumaran: And then we have our PMs, which is Justin… Amber… Rico.
125 00:13:14.940 ⇒ 00:13:16.979 Uttam Kumaran: Right? And so…
126 00:13:17.600 ⇒ 00:13:18.780 Samuel Roberts: Yeah, yeah.
127 00:13:18.780 ⇒ 00:13:23.240 Uttam Kumaran: This is some of that, then… and this is…
128 00:13:23.580 ⇒ 00:13:27.970 Uttam Kumaran: This divided by 3, it’s about $1.25 each.
129 00:13:28.060 ⇒ 00:13:29.119 Samuel Roberts: Right.
130 00:13:29.360 ⇒ 00:13:33.790 Uttam Kumaran: So if you were to look at it now, Amber is currently taking on Remo.
131 00:13:36.460 ⇒ 00:13:45.640 Uttam Kumaran: Readme… Urban Stems, ABC Home, And… Heat in.
132 00:13:47.100 ⇒ 00:13:49.550 Uttam Kumaran: What else is this,
133 00:13:53.100 ⇒ 00:13:58.980 Uttam Kumaran: Right? Rico, and then so… and then also, look, we have… we have our internal stuff, right? So if I was to do internal…
134 00:13:58.980 ⇒ 00:14:00.119 Samuel Roberts: imports, yeah.
135 00:14:00.120 ⇒ 00:14:06.330 Uttam Kumaran: So we have, AI team… Marketing team.
136 00:14:06.760 ⇒ 00:14:11.500 Uttam Kumaran: Sales team… Data team.
137 00:14:11.800 ⇒ 00:14:13.650 Samuel Roberts: Right, so I would say…
138 00:14:13.720 ⇒ 00:14:22.420 Uttam Kumaran: Probably 25, 25, 25… I would say… again, I think I’m probably, like, underestimating, but let’s just say 25 each trade.
139 00:14:22.420 ⇒ 00:14:25.960 Samuel Roberts: But yeah, these are also more variable because, at least, like, you know, the internal AI.
140 00:14:25.960 ⇒ 00:14:26.770 Uttam Kumaran: Yeah.
141 00:14:26.870 ⇒ 00:14:27.530 Samuel Roberts: Yeah.
142 00:14:27.850 ⇒ 00:14:31.789 Uttam Kumaran: And these are budget-driven, right? So, like, we have a fixed budget for these, so…
143 00:14:32.540 ⇒ 00:14:50.050 Uttam Kumaran: Yeah, and then… then Justin is basically taking on… Insomnia… Interlude… Default.
144 00:14:53.240 ⇒ 00:14:58.009 Uttam Kumaran: Right? And we should basically… oh, this is 425, so who’s doubling up?
145 00:14:58.010 ⇒ 00:15:02.469 Samuel Roberts: You didn’t add the AI teams to the… or the internal teams, rather, to the SOM, I think.
146 00:15:02.810 ⇒ 00:15:07.939 Uttam Kumaran: Oh, good call, good call. So this is actually higher, so this is equals some…
147 00:15:12.770 ⇒ 00:15:15.480 Uttam Kumaran: 475, so I’m actually missing, so…
148 00:15:15.480 ⇒ 00:15:16.230 Samuel Roberts: Right.
149 00:15:18.220 ⇒ 00:15:21.829 Uttam Kumaran: Insomnia, Interlude, Eden, so who’s taking Ellie?
150 00:15:23.420 ⇒ 00:15:26.300 Uttam Kumaran: I’m gonna just put Ellie within Justin’s name.
151 00:15:26.460 ⇒ 00:15:28.019 Samuel Roberts: Okay, yeah, I actually don’t know that one, I’m not…
152 00:15:29.680 ⇒ 00:15:38.559 Uttam Kumaran: So C… This is 450, so we’re missing one more, so C, 8, 10, 11, 7…
153 00:15:38.560 ⇒ 00:15:39.930 Samuel Roberts: Simple parts to go, maybe?
154 00:15:40.280 ⇒ 00:15:44.519 Uttam Kumaran: Oh, yeah, so pool parts, okay, pool parts is just pointing with me.
155 00:15:44.980 ⇒ 00:15:45.600 Samuel Roberts: Okay.
156 00:15:47.320 ⇒ 00:15:52.579 Uttam Kumaran: Oh, I guess, okay, let’s just put Amber, because nobody’s really managing this.
157 00:15:55.280 ⇒ 00:15:58.650 Uttam Kumaran: C16… Nope.
158 00:16:02.930 ⇒ 00:16:07.350 Uttam Kumaran: 158 each.
159 00:16:08.000 ⇒ 00:16:10.140 Uttam Kumaran: This is the split right now.
160 00:16:10.270 ⇒ 00:16:11.160 Samuel Roberts: Yeah.
161 00:16:12.860 ⇒ 00:16:19.620 Uttam Kumaran: So, ideally, what we would… what… so this is, like, let me just put a new row…
162 00:16:22.210 ⇒ 00:16:24.150 Uttam Kumaran: So this is current.
163 00:16:24.510 ⇒ 00:16:38.780 Uttam Kumaran: Our goal is probably more like… It’s… it’s sort of like… Probably, like…
164 00:16:41.100 ⇒ 00:16:46.270 Uttam Kumaran: So, 175 to 200, so this should be, like… 225…
165 00:16:47.740 ⇒ 00:16:53.790 Uttam Kumaran: So this would probably be, like, 200… 175… 100.
166 00:17:01.010 ⇒ 00:17:02.670 Uttam Kumaran: So…
167 00:17:02.670 ⇒ 00:17:07.410 Samuel Roberts: with a 50… client that could go from Amber to Justin, probably.
168 00:17:07.410 ⇒ 00:17:16.220 Uttam Kumaran: Well, like, yeah, and I guess our goal, like, we have to think about it either, like, the average team becomes easier to manage, or the capacity goes up. Like.
169 00:17:16.220 ⇒ 00:17:16.770 Samuel Roberts: Right, right.
170 00:17:16.770 ⇒ 00:17:18.759 Uttam Kumaran: I mean, it’s the same thing, but like…
171 00:17:18.760 ⇒ 00:17:19.460 Samuel Roberts: Yeah, yeah.
172 00:17:19.460 ⇒ 00:17:25.069 Uttam Kumaran: I would say, like, in an effort to… Keep wanting everything at 100?
173 00:17:25.589 ⇒ 00:17:32.550 Uttam Kumaran: our goal should instead be the small… the smallest unit, right? So we have 50… 225, 50, 100.
174 00:17:35.790 ⇒ 00:17:38.319 Uttam Kumaran: The first goal should be, can this move to 20?
175 00:17:39.090 ⇒ 00:17:39.740 Samuel Roberts: Hmm.
176 00:17:39.990 ⇒ 00:17:45.689 Uttam Kumaran: Right? So, like, for example, if this would all move to, like… if this would all move to 20s.
177 00:17:47.670 ⇒ 00:17:52.890 Uttam Kumaran: So basically, the easiest clients Are now easier to measure.
178 00:17:59.610 ⇒ 00:18:01.660 Uttam Kumaran: So this would be impact.
179 00:18:05.570 ⇒ 00:18:06.770 Uttam Kumaran: Impact.
180 00:18:08.990 ⇒ 00:18:12.540 Uttam Kumaran: So this is this accept… D, basically.
181 00:18:13.720 ⇒ 00:18:14.400 Uttam Kumaran: limited.
182 00:18:14.400 ⇒ 00:18:16.230 Samuel Roberts: If you move the goal, you just drag the…
183 00:18:16.790 ⇒ 00:18:18.109 Uttam Kumaran: Alright, I’m just gonna, I’m just gonna…
184 00:18:18.110 ⇒ 00:18:18.720 Samuel Roberts: Yeah, yeah, you’re on.
185 00:18:18.720 ⇒ 00:18:19.750 Uttam Kumaran: Okay, yeah.
186 00:18:20.990 ⇒ 00:18:24.369 Uttam Kumaran: And then… This one is also 20…
187 00:18:25.070 ⇒ 00:18:27.260 Uttam Kumaran: And this one stays the same, right?
188 00:18:48.520 ⇒ 00:18:52.549 Uttam Kumaran: Yeah, so basically what we should see here is…
189 00:18:59.440 ⇒ 00:19:01.679 Uttam Kumaran: The average moves down.
190 00:19:02.230 ⇒ 00:19:06.730 Uttam Kumaran: And… Everybody takes on less, right?
191 00:19:07.180 ⇒ 00:19:07.740 Samuel Roberts: Right.
192 00:19:07.740 ⇒ 00:19:13.429 Uttam Kumaran: It looks like, on average, Now, let’s just… I don’t care about this, actually, right now.
193 00:19:14.730 ⇒ 00:19:21.490 Uttam Kumaran: the average… per person…
194 00:19:30.000 ⇒ 00:19:32.449 Uttam Kumaran: Yeah. Oh, this is…
195 00:19:37.080 ⇒ 00:19:40.539 Uttam Kumaran: So that could be one thing, is moving 25 to 20.
196 00:19:41.290 ⇒ 00:19:46.429 Uttam Kumaran: What that allows now is… Basically, what that chain… what that change allows is, one.
197 00:19:47.160 ⇒ 00:19:50.989 Uttam Kumaran: Ricoh can now take another client without having a new person.
198 00:19:51.130 ⇒ 00:19:53.930 Uttam Kumaran: And it reduces the load on Amber and Justin.
199 00:19:54.400 ⇒ 00:19:55.050 Samuel Roberts: Yeah.
200 00:19:55.220 ⇒ 00:20:01.799 Uttam Kumaran: Another way you can do it is you could say, cool, like, now we have some… maybe the goal should be actually to move some of these 50s.
201 00:20:02.100 ⇒ 00:20:03.839 Samuel Roberts: It’s the 25.
202 00:20:03.840 ⇒ 00:20:08.040 Uttam Kumaran: And so that’s… I think this is a… could be a good framework for you to start to say, like.
203 00:20:08.150 ⇒ 00:20:15.879 Uttam Kumaran: Our business is… for example, if we were to move the 50s down 5, they’re not great, but if we move the 25s down 5,
204 00:20:16.080 ⇒ 00:20:19.389 Uttam Kumaran: We get, like, we’re getting almost, like, 20 points.
205 00:20:19.730 ⇒ 00:20:20.369 Samuel Roberts: Right, there’s a lot of.
206 00:20:20.370 ⇒ 00:20:24.479 Uttam Kumaran: 15 points in… And change, right? So…
207 00:20:24.780 ⇒ 00:20:29.780 Uttam Kumaran: Or, actually, no, no, more, like, almost, 45.
208 00:20:29.780 ⇒ 00:20:30.690 Samuel Roberts: Yeah.
209 00:20:30.690 ⇒ 00:20:35.840 Uttam Kumaran: Points of reduction, which is… which is half of a… which is, like, Half of a person.
210 00:20:36.470 ⇒ 00:20:39.340 Uttam Kumaran: I guess… or, like, I don’t know, whatever. So you kind of see.
211 00:20:39.340 ⇒ 00:20:39.700 Samuel Roberts: Like, what the.
212 00:20:39.700 ⇒ 00:20:44.219 Uttam Kumaran: optimization I’m making is, like, if we were to make our easier clients even easier to manage.
213 00:20:44.410 ⇒ 00:20:46.880 Uttam Kumaran: We have… we are… we are mostly easy clients.
214 00:20:47.070 ⇒ 00:20:49.680 Uttam Kumaran: Then… so that can be a good thing, too, is like…
215 00:20:50.680 ⇒ 00:20:57.150 Uttam Kumaran: working with Justin to look at the current allocation, because his alternative is to tell me, hey, we need more PMs.
216 00:20:57.320 ⇒ 00:20:58.040 Samuel Roberts: Right, right.
217 00:20:58.040 ⇒ 00:21:02.620 Uttam Kumaran: You know, because he would say that These should be at 100.
218 00:21:03.270 ⇒ 00:21:15.910 Uttam Kumaran: So, I guess what I’m doing is, instead of just saying these are less, I’m actually saying that both of these… all these folks are overboard. Like, they are taking on too much. So my options are to bring on another PM, or to see that, like.
219 00:21:17.180 ⇒ 00:21:18.120 Uttam Kumaran: maybe…
220 00:21:18.400 ⇒ 00:21:30.239 Uttam Kumaran: like, I need to understand for them, this is where Justin has a minimum threshold of the expectations for PMs, so he should tell me, is the… is the actual max 100, or is it, like, 150?
221 00:21:30.240 ⇒ 00:21:30.930 Samuel Roberts: 50?
222 00:21:31.400 ⇒ 00:21:33.639 Uttam Kumaran: Or is, like, should we… or should we, like.
223 00:21:33.790 ⇒ 00:21:40.460 Uttam Kumaran: Yeah, well, what is the capacity? Yeah. And then, the question for you is, like, how do you get them each more capacity, you know?
224 00:21:40.460 ⇒ 00:21:42.040 Samuel Roberts: That’s… yeah, okay.
225 00:21:46.540 ⇒ 00:21:49.200 Samuel Roberts: I guess my first thought is, like, what is a…
226 00:21:49.410 ⇒ 00:21:53.500 Samuel Roberts: Who, like, how do… how do we determine that something’s gone from 25 to 20?
227 00:21:53.690 ⇒ 00:21:57.860 Samuel Roberts: Is that just a, like, The PM’s call, Is that just a…
228 00:21:57.860 ⇒ 00:21:59.480 Uttam Kumaran: Ice… yeah.
229 00:21:59.480 ⇒ 00:22:04.689 Samuel Roberts: That’s where I’m like, we can certainly do this and, like, make moves on it, but, like, the…
230 00:22:05.120 ⇒ 00:22:08.499 Samuel Roberts: the m… The measuring aspect is, like.
231 00:22:09.030 ⇒ 00:22:15.020 Uttam Kumaran: Well, this is where, like, there’s a dis- there is a succ… there is a definition of a successful…
232 00:22:15.220 ⇒ 00:22:16.780 Uttam Kumaran: project management engagement.
233 00:22:16.930 ⇒ 00:22:26.630 Uttam Kumaran: Right? So, that’s the thing, is like, what… what is the definition of a successful PM engagement.
234 00:22:26.860 ⇒ 00:22:28.470 Uttam Kumaran: So, ideally.
235 00:22:29.240 ⇒ 00:22:37.799 Uttam Kumaran: what you’re basically looking at is a PM should be able to meet this criteria on more accounts, or the time it takes for them to meet this criteria is less.
236 00:22:37.970 ⇒ 00:22:39.890 Uttam Kumaran: Therefore, they can do it on more accounts. That’s good.
237 00:22:40.780 ⇒ 00:22:55.889 Uttam Kumaran: Right, so the… the time it takes to meet the minimum definition of success goes down. Naturally, you can then look at, okay, what are all the things you do to make… to hit the minimum?
238 00:22:56.030 ⇒ 00:22:56.540 Uttam Kumaran: And…
239 00:22:56.540 ⇒ 00:22:57.040 Samuel Roberts: Huh.
240 00:22:57.040 ⇒ 00:23:01.090 Uttam Kumaran: what options are there? And then you can ladder the SOP into this, the SOPs into this.
241 00:23:01.090 ⇒ 00:23:02.800 Samuel Roberts: Yeah, that’s good. That’s…
242 00:23:02.800 ⇒ 00:23:14.860 Uttam Kumaran: For example, to give you a sense, and Justin will tell you all of the… what the definitions are, but it’s like, you need… we need the weekly updates, we need… we need to go into stand-ups being prepared, we need all these things.
243 00:23:14.860 ⇒ 00:23:16.070 Samuel Roberts: Right, right.
244 00:23:17.730 ⇒ 00:23:21.150 Uttam Kumaran: And I guess what I’m saying is that, like,
245 00:23:21.620 ⇒ 00:23:35.300 Uttam Kumaran: I mean, again, you can then… this is a great framework for you to then do this across sales, too, and different parts of engineering. The delivery costs in our business are the highest costs, meaning the cost we’re spending on project managers cost we’re spending on engineers.
246 00:23:35.550 ⇒ 00:23:44.840 Uttam Kumaran: And so, I… my goal is actually… I’m not interested in replacing people, you know, I’ve always told you that. My goal is actually to have people take on
247 00:23:45.180 ⇒ 00:23:47.940 Uttam Kumaran: More or more complexity in the same amount of time?
248 00:23:48.730 ⇒ 00:23:57.049 Uttam Kumaran: You know? And so, partly… and that’s actually… so that’s why I think, like, partly we could…
249 00:23:57.290 ⇒ 00:24:00.300 Uttam Kumaran: This really goes to say, like, can we increase
250 00:24:00.470 ⇒ 00:24:03.470 Uttam Kumaran: Our company margin beyond 40% margin.
251 00:24:03.470 ⇒ 00:24:06.180 Samuel Roberts: That’s what I’m… yeah, the final result here is…
252 00:24:07.510 ⇒ 00:24:08.190 Uttam Kumaran: Yeah.
253 00:24:08.620 ⇒ 00:24:09.549 Samuel Roberts: This is the way to do it.
254 00:24:09.550 ⇒ 00:24:14.540 Uttam Kumaran: And so, this is where the PM team, no AI is tasked with getting to 40%.
255 00:24:15.970 ⇒ 00:24:21.020 Samuel Roberts: Right. Okay. Like, PMT plus just, like, whatever we have today, like, I’m just like… Yeah, yeah, yeah.
256 00:24:21.180 ⇒ 00:24:21.840 Uttam Kumaran: Right?
257 00:24:22.340 ⇒ 00:24:24.880 Uttam Kumaran: So my thing is, like, can you get…
258 00:24:25.180 ⇒ 00:24:32.759 Uttam Kumaran: our overall delivery margin to 50%, part of which may be PM optimization with AI,
259 00:24:32.940 ⇒ 00:24:36.740 Uttam Kumaran: Part of… part of that may be…
260 00:24:37.480 ⇒ 00:24:47.459 Uttam Kumaran: For example, if they can use AI to predict scopes of work better, and so we can move to fixed contracts, and we deliver it faster because engineers use a cursor, there’s margin.
261 00:24:47.820 ⇒ 00:24:56.659 Uttam Kumaran: Yeah, so this is where, like, there’s several angles of where to get there, but that’s where I think I could tell you that my North Star would be, can we get this beyond 40%?
262 00:24:56.960 ⇒ 00:24:58.799 Samuel Roberts: Okay, that’s… that’s…
263 00:25:00.930 ⇒ 00:25:04.160 Uttam Kumaran: And in Justin, you have someone that can answer all.
264 00:25:04.160 ⇒ 00:25:05.330 Samuel Roberts: Exactly.
265 00:25:05.330 ⇒ 00:25:06.310 Uttam Kumaran: successful PM.
266 00:25:06.490 ⇒ 00:25:08.770 Samuel Roberts: Yeah, I mean, I was already planning to chat with him
267 00:25:09.170 ⇒ 00:25:22.010 Samuel Roberts: related to, like, this mapping workflows, but we can also discuss, like, my… to get me a better understanding of what goes into this. And that’s probably related to the workflows, and then we can optimize the work… I see… okay, I see the path here.
268 00:25:26.720 ⇒ 00:25:27.510 Samuel Roberts: Alright.
269 00:25:27.760 ⇒ 00:25:28.709 Samuel Roberts: I like that.
270 00:25:31.440 ⇒ 00:25:38.940 Uttam Kumaran: Yeah, so that’s… that would be my, you know, suggestion. So looking at, like, the proposed OKRs…
271 00:25:41.160 ⇒ 00:25:47.589 Uttam Kumaran: like, this is where, like, I would say choose less, more impactful ones. Like, if you think…
272 00:25:48.130 ⇒ 00:26:02.560 Uttam Kumaran: if you… if you can… if… like, again, it’s sort of like, if you can identify, like, hey, what… what do we need to happen for us to get there, then you can change these, right? If you’re like, cool, I actually want to reduce scope to, like, 100% of
273 00:26:02.710 ⇒ 00:26:07.050 Uttam Kumaran: PMs use… use this tool, right? Or…
274 00:26:07.050 ⇒ 00:26:08.249 Samuel Roberts: Yeah, yeah, totally, totally.
275 00:26:08.250 ⇒ 00:26:13.110 Uttam Kumaran: or this, or something about grooming tickets, right? Find, like, what is limiting them right now.
276 00:26:13.460 ⇒ 00:26:14.160 Samuel Roberts: Exactly.
277 00:26:14.160 ⇒ 00:26:18.769 Uttam Kumaran: and then… Just, let’s arrive at something that’s measurable, you know?
278 00:26:19.100 ⇒ 00:26:20.599 Samuel Roberts: Yeah, that’s much better. Okay.
279 00:26:22.970 ⇒ 00:26:25.810 Uttam Kumaran: Yeah. Yeah, that’s fine.
280 00:26:26.430 ⇒ 00:26:28.120 Uttam Kumaran: Okay, cool.
281 00:26:29.480 ⇒ 00:26:32.779 Uttam Kumaran: Alright. Well, that’s all I kind of had. I mean, I think, yeah.
282 00:26:32.780 ⇒ 00:26:33.620 Samuel Roberts: This evening.
283 00:26:37.130 ⇒ 00:26:41.420 Uttam Kumaran: Yeah, just let me know if I should, anything else, I guess?
284 00:26:42.550 ⇒ 00:26:44.910 Samuel Roberts: Yeah, I don’t think there’s anything,
285 00:26:45.460 ⇒ 00:26:51.970 Samuel Roberts: But I need one blocked from you or anything right now, but… Yeah.
286 00:26:53.000 ⇒ 00:27:00.879 Uttam Kumaran: I’m gonna put together some, I’m gonna put together a product plan for role-based access control as well, for the platform, so I’ll send that over to you.
287 00:27:00.880 ⇒ 00:27:01.640 Samuel Roberts: Yes.
288 00:27:02.360 ⇒ 00:27:04.120 Samuel Roberts: Okay, yeah, so that,
289 00:27:04.240 ⇒ 00:27:08.400 Samuel Roberts: That kind of folds into, like, the bringing Google stuff into it, like, the Google Workspace stuff.
290 00:27:08.650 ⇒ 00:27:14.479 Uttam Kumaran: Kind of, yeah, there’ll be some tiered approach to it. Like, I think for some point… for some part, we’ll manage it, and then…
291 00:27:14.800 ⇒ 00:27:20.580 Uttam Kumaran: We’ll… we’ll do it, through Google, probably, through SSL.
292 00:27:20.910 ⇒ 00:27:25.910 Samuel Roberts: Yeah, I think the… Like, very first pass is just, like, you know.
293 00:27:26.170 ⇒ 00:27:28.159 Samuel Roberts: Super admin stuff that, like.
294 00:27:28.510 ⇒ 00:27:33.719 Samuel Roberts: You know, you don’t want people seeing certain things, or hiding certain things, or giving access to more tools, like we were talking about with Ricoh.
295 00:27:34.450 ⇒ 00:27:35.040 Uttam Kumaran: F.
296 00:27:35.940 ⇒ 00:27:41.180 Samuel Roberts: Which is, you know, obviously, like, you want to keep it in sync with the Google stuff, but…
297 00:27:41.550 ⇒ 00:27:48.230 Samuel Roberts: For as a first pass, like, we know Rico’s email, we can check that, we can add that to the database, we can give him a status stuff.
298 00:27:48.520 ⇒ 00:27:49.170 Uttam Kumaran: Yeah.
299 00:27:49.850 ⇒ 00:27:52.880 Samuel Roberts: Cool, that’s, yeah, that would definitely be,
300 00:27:53.120 ⇒ 00:27:56.929 Samuel Roberts: important. It’s definitely gonna be important, so I think that’s worth working on. Okay.
301 00:27:57.840 ⇒ 00:28:02.240 Uttam Kumaran: Yeah, so I’m gonna… I’m gonna work on, basically, like, PRDs for…
302 00:28:02.550 ⇒ 00:28:07.879 Uttam Kumaran: that I’ll hand to Rico and you two guys to take it out, but I’m gonna work on PRDs for,
303 00:28:08.230 ⇒ 00:28:12.640 Uttam Kumaran: Role-based access control for search, for, like, global search.
304 00:28:12.970 ⇒ 00:28:14.300 Uttam Kumaran: Yep. Yep.
305 00:28:15.460 ⇒ 00:28:20.259 Uttam Kumaran: And I think for a couple of other improvements, so I think, like, again, I think…
306 00:28:20.260 ⇒ 00:28:36.649 Uttam Kumaran: keep… let’s keep iterating on, like, how it works, because I… I think I have a lot of, like, the requirements in my head, and then I can translate them to, like, PRDs, basically, and then hand it to you guys to execute. I feel like that’s probably… it’s more about, like, what are the clear requirements from the features?
307 00:28:36.670 ⇒ 00:28:38.529 Uttam Kumaran: So I can try to deliver that.
308 00:28:42.090 ⇒ 00:28:47.090 Samuel Roberts: Yeah, that would be… that would be super helpful, because, like, there’s lots of things floating around, ideas and talk and stuff, and I’m like…
309 00:28:48.040 ⇒ 00:28:49.050 Uttam Kumaran: Yeah. Finest.
310 00:28:49.050 ⇒ 00:28:54.509 Samuel Roberts: partially that backlog, which probably has a ton of it, but, like… then the other thing was mapping it to the OKRs and getting, like, a lot more…
311 00:28:54.750 ⇒ 00:28:56.610 Samuel Roberts: Engagement of the platform, and…
312 00:28:58.370 ⇒ 00:28:59.040 Uttam Kumaran: Yes.
313 00:28:59.040 ⇒ 00:29:05.769 Samuel Roberts: There’s definitely, like, improvements to be made to the platform as from, like, a general, like, UI UX productivity improvement.
314 00:29:05.950 ⇒ 00:29:09.840 Samuel Roberts: But there’s other things, like the rule-based assets, that is, like, a little bit different than that.
315 00:29:10.000 ⇒ 00:29:12.689 Samuel Roberts: But it’s still important, you know?
316 00:29:14.060 ⇒ 00:29:21.030 Samuel Roberts: Okay, yeah, that sounds like a good plan. Yeah, if we can, you know, requirements, and then we can scope out what needs to get there, and figure out, like.
317 00:29:21.710 ⇒ 00:29:24.480 Samuel Roberts: Timelines and priorities and plans.
318 00:29:24.630 ⇒ 00:29:25.290 Uttam Kumaran: Yeah.
319 00:29:25.690 ⇒ 00:29:26.270 Samuel Roberts: Cool.
320 00:29:26.540 ⇒ 00:29:27.909 Samuel Roberts: That’d be great, yeah.
321 00:29:32.530 ⇒ 00:29:37.789 Samuel Roberts: Anything else? Like, there was… what was that project you posted? Or not project, the potential?
322 00:29:41.080 ⇒ 00:29:41.909 Uttam Kumaran: Which one?
323 00:29:42.170 ⇒ 00:29:43.920 Samuel Roberts: I’m blanking on the name already, but it…
324 00:29:46.680 ⇒ 00:29:51.289 Uttam Kumaran: Oh, shit, I forgot, which one did I send? .
325 00:29:51.550 ⇒ 00:29:52.190 Samuel Roberts: Yeah.
326 00:29:53.240 ⇒ 00:29:59.930 Uttam Kumaran: there was iCustomer, there was Marvin…
327 00:30:00.860 ⇒ 00:30:03.009 Uttam Kumaran: I can check. I don’t know, I’ve been, like.
328 00:30:03.010 ⇒ 00:30:05.940 Samuel Roberts: Oh, was it the, the underwear company.
329 00:30:06.170 ⇒ 00:30:10.010 Uttam Kumaran: Oh, Shinesty, yeah, yeah, yeah, I sent it… I sent them the proposal yesterday.
330 00:30:10.330 ⇒ 00:30:10.780 Samuel Roberts: Cool.
331 00:30:10.780 ⇒ 00:30:17.339 Uttam Kumaran: Basically, they want us to help automate, like, a manual entry process into NetSuite, all through Slack. Yeah. Yeah, cool project.
332 00:30:17.830 ⇒ 00:30:19.709 Samuel Roberts: Yeah, yeah, that seems really, like…
333 00:30:19.930 ⇒ 00:30:25.649 Samuel Roberts: doable and fun to, like, push on, now that we have… yeah, I, I think that one’s sweet.
334 00:30:27.550 ⇒ 00:30:33.130 Samuel Roberts: Okay, cool. Yeah, I just took a look at that and went through… I watched the meeting and stuff, and read the notes and stuff, or the transcript and stuff, and…
335 00:30:33.130 ⇒ 00:30:33.790 Uttam Kumaran: Yeah.
336 00:30:33.790 ⇒ 00:30:38.270 Samuel Roberts: I was like, I’m even more excited about it now. They seem, like, really eager. So it’s kinda cool, so…
337 00:30:38.470 ⇒ 00:30:41.780 Samuel Roberts: they need help. So that’s why I was like, okay, good, let’s do it.
338 00:30:42.730 ⇒ 00:30:43.430 Samuel Roberts: Cool.
339 00:30:46.090 ⇒ 00:30:49.849 Samuel Roberts: I don’t think if there was anything else I had for you right now, but I think I’m good.
340 00:30:50.600 ⇒ 00:30:51.280 Uttam Kumaran: Okay.
341 00:30:51.520 ⇒ 00:30:55.009 Uttam Kumaran: Great. Yeah, I think just, like, continue to think about, like.
342 00:30:56.180 ⇒ 00:31:03.050 Uttam Kumaran: you know, when we go to clients, you know, one thing we’re trying to show is, like, how do you actually measure ROI? And I see we’re even finding
343 00:31:03.260 ⇒ 00:31:08.289 Uttam Kumaran: difficulty in how we do it, and so I think it’s important for us to… to, like.
344 00:31:08.930 ⇒ 00:31:12.919 Uttam Kumaran: Try some unorthodox, you know, ways of, like, how do we…
345 00:31:14.150 ⇒ 00:31:19.369 Uttam Kumaran: use AI to actually, like, very specifically change, like.
346 00:31:19.570 ⇒ 00:31:22.070 Uttam Kumaran: The status quo of how businesses like this are running.
347 00:31:22.070 ⇒ 00:31:22.810 Samuel Roberts: Yeah.
348 00:31:23.080 ⇒ 00:31:32.719 Uttam Kumaran: You know, most people at the peak, like, when they crush everything, and they have every process tightened, and they’re selling a very expensive service, they get to 40%.
349 00:31:32.860 ⇒ 00:31:36.490 Uttam Kumaran: Right? And so for us, it’s like, we want to do that…
350 00:31:36.610 ⇒ 00:31:39.579 Uttam Kumaran: without any of those. You know, we want to do that.
351 00:31:40.040 ⇒ 00:31:42.750 Uttam Kumaran: With not the tightest process, like, with…
352 00:31:42.750 ⇒ 00:31:43.400 Samuel Roberts: Yeah, yeah.
353 00:31:43.400 ⇒ 00:31:51.090 Uttam Kumaran: beyond company, and we want to hit beyond 40%. So, I do think it’s important to tie to those financial metrics.
354 00:31:51.280 ⇒ 00:31:57.690 Uttam Kumaran: And everybody’s gonna sort of have a tie-in into that, you know? But I do think that that’s… That’s…
355 00:31:58.150 ⇒ 00:32:02.330 Uttam Kumaran: you know, important to think about. And then the other thing is, you know, I think we’re gonna move
356 00:32:02.740 ⇒ 00:32:06.219 Uttam Kumaran: I should send you this article about, like, sort of how…
357 00:32:06.340 ⇒ 00:32:18.840 Uttam Kumaran: like, company structures may change with AI, but, you know, a lot of people are, like, maybe there’s fewer senior people, a lot more junior people, and a lot more AI agents, or there’s, like, there’s, like, senior people, but then everybody has, like.
358 00:32:19.490 ⇒ 00:32:26.529 Uttam Kumaran: really, like, junior people are, like, almost, like, coordinators versus, like… so basically there’s, like, elimination of, like, mid-level.
359 00:32:26.860 ⇒ 00:32:27.190 Samuel Roberts: Yeah.
360 00:32:27.190 ⇒ 00:32:33.729 Uttam Kumaran: some interesting ways for us… you know, we’re not that many people, but one thing that I think me and you will work on over time is, like.
361 00:32:34.170 ⇒ 00:32:41.099 Uttam Kumaran: how does this fundamentally change the way we run a consulting business? Like, org structure and everything? Like, I think everything’s on the table.
362 00:32:41.100 ⇒ 00:32:42.380 Samuel Roberts: Given, like.
363 00:32:42.400 ⇒ 00:32:46.159 Uttam Kumaran: we, we, we implement these, these, these AI features, you know?
364 00:32:46.690 ⇒ 00:32:47.280 Samuel Roberts: Yeah.
365 00:32:47.930 ⇒ 00:32:50.270 Uttam Kumaran: So, I think something interesting to think about.
366 00:32:50.570 ⇒ 00:32:59.110 Samuel Roberts: Yeah, there was one, thing kind of related to that on the call, the presentation earlier. He was talking about teams that had
367 00:32:59.220 ⇒ 00:33:02.499 Samuel Roberts: Like, architect-level engineers versus not.
368 00:33:02.850 ⇒ 00:33:11.960 Samuel Roberts: And the idea was that, like, the pre-planning that an architecture Minded person would do, like, would…
369 00:33:12.290 ⇒ 00:33:17.389 Samuel Roberts: require much less rework from the AI, which is specifically what the AI seems worse at.
370 00:33:17.550 ⇒ 00:33:19.750 Samuel Roberts: You know? Like, it’s good at…
371 00:33:20.430 ⇒ 00:33:31.579 Samuel Roberts: the initial thing, because there’s no context to worry about, really. There’s very little context to worry about. But once you start… once you get code, like, reworking and bug fixes is actually, like, way less productive with AI, is what they found.
372 00:33:31.930 ⇒ 00:33:33.120 Uttam Kumaran: Yeah. Okay.
373 00:33:33.120 ⇒ 00:33:37.120 Samuel Roberts: Something like that, where I totally see what you’re saying, like… Enabling people to…
374 00:33:37.230 ⇒ 00:33:41.719 Samuel Roberts: have the spec feed that to Cursor, rather than…
375 00:33:42.100 ⇒ 00:33:49.090 Samuel Roberts: you know, just start working in Cursor and with a rough, rough idea of what needs to get done. I’m definitely excited for, like.
376 00:33:49.370 ⇒ 00:33:54.609 Samuel Roberts: a couple, like, some new AI projects, because I definitely want to, like, make sure… because I kind of came into a bunch of these projects, right?
377 00:33:54.950 ⇒ 00:33:56.030 Samuel Roberts: Excuse me.
378 00:33:56.720 ⇒ 00:33:57.910 Uttam Kumaran: So I’m excited to, like.
379 00:33:57.910 ⇒ 00:33:59.799 Samuel Roberts: Start one fresh.
380 00:34:00.170 ⇒ 00:34:03.370 Uttam Kumaran: Yes. Really, like… We’re a great, we’re a great…
381 00:34:03.610 ⇒ 00:34:07.120 Uttam Kumaran: we’re a great guinea pig, right? But also, I think this is where, like.
382 00:34:07.600 ⇒ 00:34:19.260 Uttam Kumaran: I’m gonna… because we’re a great guinea pig, we should dream big. Like, imagine we had a client that was like, everything’s… everything is an opportunity for automation. We’re willing to even change
383 00:34:19.380 ⇒ 00:34:21.530 Uttam Kumaran: Titles in the company, org structure.
384 00:34:21.530 ⇒ 00:34:22.070 Samuel Roberts: Sure, sure.
385 00:34:22.070 ⇒ 00:34:28.490 Uttam Kumaran: Right? So I don’t… I want us to go extremely hard in the paint on our company.
386 00:34:29.350 ⇒ 00:34:31.610 Uttam Kumaran: What works, what doesn’t for our customers.
387 00:34:31.730 ⇒ 00:34:35.250 Uttam Kumaran: But if we don’t end up challenging some of these things, like.
388 00:34:35.949 ⇒ 00:34:44.359 Uttam Kumaran: like, I feel like our word to our customers would… couldn’t… would… would just be higher if we… like, so that… I don’t know, and that’s also, I think, really where
389 00:34:44.590 ⇒ 00:34:52.940 Uttam Kumaran: we’re in such a unique spot right now, like, where we can do some of these things, and we have buy-in, you know? Yeah. I don’t think many companies…
390 00:34:53.489 ⇒ 00:34:58.920 Uttam Kumaran: have… Just sort of like, we’ll change everything if they have something for us, you know?
391 00:34:58.920 ⇒ 00:34:59.880 Samuel Roberts: Right, right.
392 00:34:59.880 ⇒ 00:35:01.269 Uttam Kumaran: So, it’s worth us, like.
393 00:35:01.400 ⇒ 00:35:08.319 Uttam Kumaran: going big and trying stuff, and… but this is the thing, like, it’s just gonna be me and you sort of talking and thinking through, like, I don’t think a lot of this…
394 00:35:08.710 ⇒ 00:35:12.089 Uttam Kumaran: people aren’t doing… I don’t think a lot of it is written or exists, you know.
395 00:35:12.410 ⇒ 00:35:15.560 Samuel Roberts: Yeah, yeah, it’s a whole new world to figure out.
396 00:35:17.520 ⇒ 00:35:18.130 Samuel Roberts: Cool.
397 00:35:18.130 ⇒ 00:35:18.790 Uttam Kumaran: Yeah.
398 00:35:24.040 ⇒ 00:35:28.270 Samuel Roberts: Yeah, thinking of, like, the re… oh yeah, definitely send me that, if there was… you mentioned an article?
399 00:35:28.740 ⇒ 00:35:31.060 Uttam Kumaran: Yes, I will send you that right now, actually.
400 00:35:31.230 ⇒ 00:35:32.160 Samuel Roberts: Cool, yeah.
401 00:35:34.220 ⇒ 00:35:38.109 Uttam Kumaran: It’s not a long article, but it basically just… just poses this, like, question, you know?
402 00:35:38.110 ⇒ 00:35:48.030 Samuel Roberts: Yeah, yeah, totally, totally. Yeah, I was watching some stuff, like, presentations from some people of different frameworks and things, and the, like, the amount… I forget what he said, but it’s, like, it’s all about, like.
403 00:35:48.340 ⇒ 00:35:53.509 Samuel Roberts: Like, theory versus practice, and, like, practice is the only thing leading right now, because, like, we just don’t…
404 00:35:54.310 ⇒ 00:35:56.400 Samuel Roberts: Have the theory for some of this stuff.
405 00:35:56.790 ⇒ 00:35:57.220 Uttam Kumaran: Yeah.
406 00:35:57.220 ⇒ 00:36:04.099 Samuel Roberts: They’re just moving so fast in so many other ways, and it’s like, if it works, it works. Even if it’s, like, the theory says it shouldn’t, kind of thing.
407 00:36:04.380 ⇒ 00:36:05.150 Uttam Kumaran: Yes.
408 00:36:05.860 ⇒ 00:36:08.009 Uttam Kumaran: But there’s no theory, like…
409 00:36:08.210 ⇒ 00:36:23.719 Uttam Kumaran: I would say I’ve… and I can send you, like… I don’t know if you… if you tend to, like, be interested in reading any business books, but I… I’ve read the theory behind the professional services firm, like… Yeah. …way that these are constructed.
410 00:36:23.850 ⇒ 00:36:37.569 Uttam Kumaran: like, there’s a great book called Managing the Professional Service Firm, it was written in, like, the 80s, and it’s basically, like, the playbook to running a company like this. Of course, back then they had no technology, so, like, everything is about being in office and phone and stuff like that, but…
411 00:36:37.570 ⇒ 00:36:38.110 Samuel Roberts: Yeah.
412 00:36:38.110 ⇒ 00:36:55.719 Uttam Kumaran: it essentially is the same thing, it’s like, it explains what the optimization problem is. So I’ve read… so that’s… that way I know… I kind of knew, like, what ceiling we were up against, but it may help you to read that, or a similar book, and then I’m happy to get… that book is, like, really the Bible on this world, because then you’ll see, like.
413 00:36:55.800 ⇒ 00:37:05.609 Samuel Roberts: you’ll read it, you’ll be like, why… that rule is completely automatable, you know? Totally, totally. Yeah, that would actually be great, because my mind is… I’m still, like, struggling a little bit, like.
414 00:37:05.900 ⇒ 00:37:07.859 Samuel Roberts: clients versus products and stuff.
415 00:37:07.860 ⇒ 00:37:08.559 Uttam Kumaran: It’s just… Yeah.
416 00:37:09.070 ⇒ 00:37:13.289 Samuel Roberts: My way of thinking, like, startup-wise is, like, similar in some ways, but…
417 00:37:13.290 ⇒ 00:37:13.960 Uttam Kumaran: Yeah.
418 00:37:14.180 ⇒ 00:37:17.079 Samuel Roberts: I definitely, like, don’t have a full grasp on all this stuff, like…
419 00:37:17.270 ⇒ 00:37:19.489 Samuel Roberts: From the business side, which would probably…
420 00:37:19.490 ⇒ 00:37:21.340 Uttam Kumaran: It’s a little bit of a heavy book, like…
421 00:37:21.690 ⇒ 00:37:24.680 Uttam Kumaran: Like, do you care? I’ll get it for you today.
422 00:37:24.930 ⇒ 00:37:30.490 Samuel Roberts: Yeah, certainly, like, I… if it’s… if I have it, and I can at least, like, start working on it, you know, if it’s a heavy book, but…
423 00:37:30.670 ⇒ 00:37:42.469 Uttam Kumaran: No, no, no, I just don’t know whether you’re, like, more of a reader, like… I don’t know if you… I guess… I just think it’s not, like, the most interesting subject, but it is very scenario-based, and it is… it almost reads like a textbook about.
424 00:37:42.470 ⇒ 00:37:43.170 Samuel Roberts: Yeah, yeah.
425 00:37:43.170 ⇒ 00:37:44.419 Uttam Kumaran: firm. But, like…
426 00:37:44.770 ⇒ 00:37:47.279 Samuel Roberts: Yeah, I’m a pretty good reader.
427 00:37:48.210 ⇒ 00:37:49.150 Uttam Kumaran: Oh, then you’ll cross it.
428 00:37:49.150 ⇒ 00:37:53.220 Samuel Roberts: is… Is there an audiobook I can listen to on, like, runs and stuff?
429 00:37:53.220 ⇒ 00:37:56.760 Uttam Kumaran: Yeah, look up… look up Managing the Professional Service Firm.
430 00:37:57.750 ⇒ 00:38:02.120 Uttam Kumaran: And then, if you want to buy it, you could just submit a… you can just submit a,
431 00:38:03.010 ⇒ 00:38:06.360 Uttam Kumaran: A receipt to… finance that.
432 00:38:08.330 ⇒ 00:38:17.199 Samuel Roberts: Okay. Yeah, yeah, because, I mean, a physical copy is always good, because it’s, like, good to reference things, but the audiobook is probably how I would digest it faster.
433 00:38:17.620 ⇒ 00:38:22.220 Uttam Kumaran: No, you should just get… just get both, and… because I use it as almost like a… like a reference.
434 00:38:22.220 ⇒ 00:38:22.900 Samuel Roberts: Yeah, exactly.
435 00:38:22.900 ⇒ 00:38:23.900 Uttam Kumaran: multiple times.
436 00:38:24.140 ⇒ 00:38:28.089 Uttam Kumaran: Amber’s read it as well, and she loves it.
437 00:38:28.290 ⇒ 00:38:29.290 Uttam Kumaran: Boom.
438 00:38:29.730 ⇒ 00:38:33.299 Uttam Kumaran: But I don’t know, that, like, that just babies… because, I don’t know, everybody, like…
439 00:38:33.600 ⇒ 00:38:38.920 Uttam Kumaran: This is actually a very simple business. It’s complica- we… it’s just complicated for… because of clients.
440 00:38:38.960 ⇒ 00:38:42.630 Samuel Roberts: Yes. From a business model standpoint, it’s actually not… we’re not, like.
441 00:38:42.630 ⇒ 00:38:45.400 Uttam Kumaran: There’s not, like… we’re not amortizing equipment, we’re not, like.
442 00:38:45.400 ⇒ 00:38:46.110 Samuel Roberts: Oh my god.
443 00:38:46.110 ⇒ 00:38:49.130 Uttam Kumaran: There’s not, like, a lot of, you know, but
444 00:38:49.540 ⇒ 00:38:56.810 Uttam Kumaran: It is, like, an optimization for… for hours… for hours build and the margin, but it’s challenging because of the… because, like, of…
445 00:38:56.990 ⇒ 00:39:00.910 Uttam Kumaran: Just the nature of getting clients, and, like, there’s high-churn business, you know, so…
446 00:39:01.210 ⇒ 00:39:02.320 Samuel Roberts: Fair, fair.
447 00:39:03.260 ⇒ 00:39:17.789 Samuel Roberts: Yeah. Oh, no, this is definitely… this is good. Yeah, I’m… I… I have been, like… I haven’t been into, like, business books in a little bit, but, like, when I was working on a haircare company, I was, like, devouring them, and, like, a few of them just, like, really popped off, and I, like, bought one for my co-founder and stuff, and we, like, really tried to…
448 00:39:19.110 ⇒ 00:39:20.569 Samuel Roberts: Like, you know, every once in a while, there’s a good.
449 00:39:20.570 ⇒ 00:39:23.430 Uttam Kumaran: All I used to read is only business books.
450 00:39:24.270 ⇒ 00:39:28.299 Uttam Kumaran: And so now I read, like, business… like, business stories, like.
451 00:39:28.630 ⇒ 00:39:32.759 Uttam Kumaran: About crime… about white-collar crime, like, stuff like that.
452 00:39:33.300 ⇒ 00:39:39.520 Uttam Kumaran: Or, like, yeah, like, I don’t know, I like these books. I learned everything from them, like…
453 00:39:39.610 ⇒ 00:39:40.700 Samuel Roberts: Yeah. You know?
454 00:39:41.740 ⇒ 00:39:46.939 Samuel Roberts: I don’t remember what that book was, I just popped up. I think it was as we were growing, it was a book about, like, company culture kind of thing, and I…
455 00:39:46.940 ⇒ 00:39:48.010 Uttam Kumaran: Oh, nice.
456 00:39:48.010 ⇒ 00:39:50.229 Samuel Roberts: And I was just, like, something about it was like, this is…
457 00:39:50.440 ⇒ 00:39:56.470 Samuel Roberts: this is… this has touched on everything I’ve been thinking about and trying to figure out. Which…
458 00:39:56.750 ⇒ 00:40:01.780 Samuel Roberts: is a big problem in and of itself, company culture, actually, growing, like, super fast anyway, but…
459 00:40:03.590 ⇒ 00:40:07.720 Samuel Roberts: A lot of things went wrong that way, but the book was interesting and helpful.
460 00:40:09.110 ⇒ 00:40:15.010 Samuel Roberts: But yeah, I’ll take a look, and then, yeah, I’ll totally submit a receipt, pro- yeah, for both, then. That sounds great.
461 00:40:15.340 ⇒ 00:40:15.980 Uttam Kumaran: Okay.
462 00:40:16.550 ⇒ 00:40:17.150 Samuel Roberts: Sweet.
463 00:40:18.010 ⇒ 00:40:22.019 Uttam Kumaran: Okay, perfect. Alright. Thank you, dude. Talk to you soon.
464 00:40:22.220 ⇒ 00:40:23.560 Samuel Roberts: Thank you, yeah, peace.