Meeting Title: Uttam <> Emily Date: 2025-05-15 Meeting participants: Amber Lin, Uttam Kumaran, Emily Chan
WEBVTT
1 00:02:29.050 ⇒ 00:02:30.759 Uttam Kumaran: Hey, Emily, how are you?
2 00:02:31.160 ⇒ 00:02:33.020 Emily Chan: Hi, Tom, how are you?
3 00:02:33.020 ⇒ 00:02:39.092 Uttam Kumaran: Good. Sorry. I’m just just running back from a meeting. That’s why I’m video off. But
4 00:02:39.770 ⇒ 00:02:41.470 Uttam Kumaran: how’s life? How’s everything?
5 00:02:41.860 ⇒ 00:02:46.499 Emily Chan: Good good much the same, which is good, I guess. How about you?
6 00:02:46.870 ⇒ 00:02:59.900 Uttam Kumaran: Things are like every week. It’s it’s different around here. So it’s good. It’s a team has expanded a lot. We’re we’re at 15 people where we sort of
7 00:03:00.310 ⇒ 00:03:24.369 Uttam Kumaran: plugged a lot of the the people holes that. You know, we were. We were dealing with less about like who more about getting like a strong core group of of people that are full time at the company. So that’s been really good. We have a really strong group of AI and data engineers now, and, like, you know, basically moving everything to to be a process.
8 00:03:24.941 ⇒ 00:03:34.419 Uttam Kumaran: and some sort of repeatable process versus something at me or partner handling. So it’s getting a lot better. Yeah.
9 00:03:35.520 ⇒ 00:03:36.969 Emily Chan: Sounds like you’re scaling up.
10 00:03:37.630 ⇒ 00:03:55.630 Uttam Kumaran: Yes, yes, but a lot slower than I’m I’m used to a lot more painful than I’m using, you know, when, when you don’t have unlimited Vc money things go a lot slower. But we’re making do with what we can. And I would say, even exactly like for not having any
11 00:03:55.910 ⇒ 00:04:03.180 Uttam Kumaran: any cash and like sort of just reinvesting. It’s it’s it’s definitely working. So.
12 00:04:04.326 ⇒ 00:04:12.390 Emily Chan: That’s interesting. Can you tell me more? What? What is a pain point when you say it is slow and maybe needing some money.
13 00:04:13.144 ⇒ 00:04:19.179 Uttam Kumaran: Yeah, I mean, we like when when we’re like a cash in cash out business. So
14 00:04:19.300 ⇒ 00:04:41.965 Uttam Kumaran: I can only reinvest the money that you know, we get from clients that we make a margin on right? So I don’t we? We didn’t. We don’t have like a million dollars in the bank where I can go hire 10 people today and then start everything. It’s sort of all incremental. So we have to increase the margin slowly before we can actually go. Do recruitment, you know.
15 00:04:42.410 ⇒ 00:04:53.209 Uttam Kumaran: but yeah, interested even in your thoughts, like hearing that. And yeah, it’s difficult.
16 00:04:54.750 ⇒ 00:05:03.429 Emily Chan: Of you know, bringing on Robert and scaling that way, because maybe that is a good way to expand your margins. Without Vc funding.
17 00:05:07.664 ⇒ 00:05:18.379 Uttam Kumaran: Sorry. I’m just just hopping out and over. Yeah, I mean, that’s exactly what we did. We? We. We tried our best to scale through people, but also through a lot of
18 00:05:18.500 ⇒ 00:05:37.919 Uttam Kumaran: a lot of favors like I, you know, I tapped every every resource I have, and we also use an AI ton. So I’m able to do things a lot less than I think people expect. We have a global workforce. So although we do have people here in the Us. We also have people in many other places. So that helps keep our costs low.
19 00:05:38.270 ⇒ 00:05:57.870 Uttam Kumaran: And if you look at our materials, if you look at the pace at which we’re doing things that other consultancies aren’t doing right, finding ways to differentiate ourselves through marketing, through relationships, we’re punching a lot higher above our weight. You know, like. And those are things that even if you have a hundred
20 00:05:57.940 ⇒ 00:06:21.226 Uttam Kumaran: people are 500 people. If you have no relationships, no presence, no sales motion, you know. It’s it’ll be very, very difficult for you. So it’s it’s just like, I think the biggest cost. Is that what I? What we don’t pay for in price? We pay for an effort right? And then we pay for in, you know, 1012, 15 h days for for a long time. But
21 00:06:21.590 ⇒ 00:06:40.670 Uttam Kumaran: it’s it’s all you know. It’s it’s it’s working out. And it’s we have a lot of really great repeat clients. We’ve signed our longest deal just yesterday. With a really large local home service client here. In Austin for for AI work. And
22 00:06:41.031 ⇒ 00:06:46.739 Uttam Kumaran: yeah, we’re fine. We’re putting out case studies. We’re we’re doing all the right things. I think it just took a
23 00:06:46.850 ⇒ 00:06:56.380 Uttam Kumaran: it just all takes time, and it takes it takes people, you know. So. But then the lovely thing is, when you have constraints, you’re forced to innovate, and we’ve done a lot of that. So it’s been good.
24 00:06:56.850 ⇒ 00:06:59.230 Emily Chan: Congrats on that long-term contract.
25 00:06:59.780 ⇒ 00:07:11.119 Uttam Kumaran: Thanks. Thanks. Yeah. We’re we’re signing like now, moving from signing 3 to 6 month contracts. And the next thing we’ll be looking for is to sign 9 and 12 month contracts. So that’s like what keeps us
26 00:07:11.910 ⇒ 00:07:28.629 Uttam Kumaran: keeps the business, you know, alive and allows us to work from. You know, when I when I when I used when I started the business, I was just like, Okay, are we gonna be alive next month? Right now, I’m like, Okay, and what’s our 2, 3, 4 Year Plan? I can actually
27 00:07:35.620 ⇒ 00:07:40.480 Uttam Kumaran: thank you for that verses.
28 00:07:44.130 ⇒ 00:07:50.270 Uttam Kumaran: You know. Not a very, you know.
29 00:07:51.180 ⇒ 00:07:52.850 Emily Chan: Hey, Utam, you’re breaking up.
30 00:07:54.920 ⇒ 00:07:55.840 Emily Chan: Hello!
31 00:07:56.950 ⇒ 00:08:02.650 Amber Lin: Yeah. His audio is also very choppy for me. By the way, Emily, Hello! I’m amber nice to meet you.
32 00:08:02.650 ⇒ 00:08:03.799 Uttam Kumaran: Can you hear me? I’m just.
33 00:08:03.800 ⇒ 00:08:04.530 Emily Chan: Thank you.
34 00:08:05.070 ⇒ 00:08:05.780 Emily Chan: Yes, you’re good.
35 00:08:08.090 ⇒ 00:08:12.969 Uttam Kumaran: Okay, maybe maybe maybe just do intros amber. I’m just getting out of this parking garage. So I’ll.
36 00:08:12.970 ⇒ 00:08:14.990 Amber Lin: Okay, okay, sounds good.
37 00:08:15.210 ⇒ 00:08:43.297 Amber Lin: Hi, Emily, I’m amber. I have no context of this call. Sent me an invitation. I was like, cool. Let me join I’m a project manager at Brainpage. I joined about almost 3 months, 2 months, 3 months ago, and I went from project manager, and I was helping with some internal initiatives. And right now I’m helping with setting the strategy, organizing our different initiatives. So I was very curious to see what this call was about.
38 00:08:43.919 ⇒ 00:08:57.379 Emily Chan: Yeah, I think. I just you know, I met with them through a a mutual connection. And then we just have catch up calls. And this just happens to be one of them. Maybe I tell you a bit about me, and then that will explain a bit of the context. I think.
39 00:08:57.380 ⇒ 00:09:18.110 Uttam Kumaran: Yeah, that that’d be that’d be lovely, Emily. And yeah, I think, Amber, as you know, we we, you know, when when smart people come our way, and for me the best thing I can do is like steward of the companies. Try to find ways to work with people. But I you know, I’ve learned a lot talking to Emily and I I just, you know, thought you’d be an interesting connection for for you to have. And also, I think
40 00:09:18.448 ⇒ 00:09:43.759 Uttam Kumaran: Emily Amber could give you another perspective on our project execution. Because, you know, my perspective these days went from really in the weeds. Now, I’m I’m really focused on the macro of the company, but she’s really in the day to day of like, okay, assigning engineers to clients executing how we communicate sprint schedules like, it’s nice to get multiple perspectives. So yeah, feel free to maybe give a brief intro, Emily, and yeah.
41 00:09:44.100 ⇒ 00:10:09.000 Emily Chan: Yeah, definitely. So I’m exploring, you know, data science consulting part-time role right now, which is how autumn and I got connected my background is that I have over 15 years of experience in data science. And it’s pretty much, you know, anything that use data to help a business make better decisions. So it could be like, okay, what is the business problem? What does that mean?
42 00:10:09.010 ⇒ 00:10:28.889 Emily Chan: For what kind of data do we need? What kind of, you know, data modeling infrastructure do we need? And then how do we use the analysis? And it doesn’t stop there. Right? It’s like, okay, really, working with the client and the executives to figure out, okay, how do they use this analysis to make decision? Can we bring the data into the boardroom, into meetings, into planning.
43 00:10:28.890 ⇒ 00:10:40.509 Emily Chan: And then the cycles goes again. It’s like, okay. Now that we are in this meeting with the business stakeholders, the data say something, it inspires a solution. So then, okay, that solution will probably
44 00:10:40.510 ⇒ 00:10:42.318 Emily Chan: create more data project.
45 00:10:42.820 ⇒ 00:10:55.869 Emily Chan: my my experience is, I work at spotify previously. And then before that I was at matter, and then some like really ancient history. I started my career at Mckinsey, and then I did finance for a little while.
46 00:10:56.316 ⇒ 00:11:06.449 Emily Chan: I can talk about maybe an example from my spotify days that may sort of like, you know, really solidify what I do. So I led a team at spotify that
47 00:11:06.450 ⇒ 00:11:30.400 Emily Chan: expanded spotify into India. There was no spotify in India, and my team actually did all the data related to that project we launched spotify in India. So that means, you know, at the beginning of the project, we go out there. Source all the data that we need to understand. You know what is the market opportunity, what other regions we want to go after? What language do we want to support? So all that data that goes into.
48 00:11:30.420 ⇒ 00:11:52.229 Emily Chan: you know, helping the product manager define requirements for that product and also helping our business leaders plan like revenue and cost structure for this project. And then, as we goes into like product development. My team does a lot of Ab testing. Okay, once we have a a features, you know, we need to test it, make sure that it is performing up to standards.
49 00:11:52.230 ⇒ 00:12:17.659 Emily Chan: And then, as we go towards launch, you know, there’s this huge effort into like building, this end to end customer view. So we go out there. We take data from external marketing sources. We take data from the app store. We take data from our internal. You know, music app, so that we can see like, okay, this user is coming in from, you know, a marketing campaign on Facebook. And then they go through an android app store. They come to
50 00:12:17.660 ⇒ 00:12:30.040 Emily Chan: our platform. Oh, they have a problem logging in what’s going on so like my team. Make sure that we have data coming in from every places. They all talk to each other, and we can then surface them in like a seamless end, to end manner.
51 00:12:30.160 ⇒ 00:12:57.739 Emily Chan: And and then we bring in the business stakeholders right? Like my team, also run this weekly task force where we have people from like all the functions. Right? So if we have data from marketing, someone from marketing is there and then at the end of the user journeys, we have data from customer service. Someone is mad that their user playlist isn’t showing them the right place. We make sure that, you know call center people are in that meeting. So if we have data showing us like, okay, how is the user performing which part of the journey is not really right?
52 00:12:57.740 ⇒ 00:13:04.649 Emily Chan: So we have the data to tell us what we need to focus on. And we also make sure that the right people is looking at the right data
53 00:13:04.929 ⇒ 00:13:28.909 Emily Chan: and that’s also where the fun happens. Because then, like, you know, 1 point is that, oh, we realize that people have problem logging in, or we’re not providing the right lock in method, they would prefer to log in using phone number. So that actually creates a whole project for engineering and for us to build a new features. So you know again, it’s that flywheel of like, what is the business problem, figuring out how to like use data to make that business and the product better.
54 00:13:28.910 ⇒ 00:13:36.580 Emily Chan: And then the data will give you more inspiration to again create that flywheel of always making a better business and product.
55 00:13:36.580 ⇒ 00:13:52.910 Emily Chan: So that is kind of what I do. Personally, I’m a mom with 2 young kids. That’s why I left the workforce a little over 2 years ago. My kids are now older now. And so I’m exploring, you know, consulting or data science work on the part time side. That’s about me.
56 00:13:53.130 ⇒ 00:14:13.649 Uttam Kumaran: Yeah, I think you guys have, you know, something common? Because Amber also worked at Ui briefly, and part of the reason I think she’s excelled. Here is exactly like what you described, which is like, you know, commonly for engineers on our team, and even for typical project managers. It’s almost hard sometimes to just tell the narrative
57 00:14:14.122 ⇒ 00:14:19.379 Uttam Kumaran: what internally, and again, when you’re part of the company. But when we’re coming in as outside consultants.
58 00:14:19.450 ⇒ 00:14:34.389 Uttam Kumaran: you know they expect a lot from us. So what I like about Amber’s approaches she does a lot on like sort of the what you could expect from like management consulting, which is like, Hey, here’s like a 3 we’ve we’ve dug through all this data. Here’s like 3 slides on like improvements. You should make.
59 00:14:34.711 ⇒ 00:14:55.109 Uttam Kumaran: I think similar. Also, Emily Amber asked a lot of really good questions. So one thing I enjoy about our conversations is, you always ask like, can you explain that? Or what do you mean? Not a lot of people do that. And as I I just if you ask me about the business, I’ll I’ll say I’ll tell whatever. But some people just listen. So yeah, I’m glad that I can connect everyone here.
60 00:14:55.790 ⇒ 00:14:57.309 Emily Chan: Yeah, thanks for connecting us.
61 00:14:57.450 ⇒ 00:15:14.109 Amber Lin: Yeah, thank you. I didn’t know this was for that purpose. I’m really glad I joined. And, Emily, I I really like your trajectory or the background because for for my background, I went to a school program. That was a collaboration between
62 00:15:14.110 ⇒ 00:15:30.290 Amber Lin: 3 schools. It was Usc. A school in Hong Kong, and then a school in Italy, and everybody in that program. My cohort is like around 50 people each year everybody goes into consulting and finance. And when I 1st started I was like, Okay, I’ll go there, too.
63 00:15:30.400 ⇒ 00:15:53.439 Amber Lin: But then I was like, I don’t think I I want to stay there forever. Like, if I like, I want to have, as you said, actual impact of like you working at spotify, you identify a problem. You can immediately, wanna, not immediately, but soon see it come to effect and see a new project come up. You can. You can have a say in what’s actually done.
64 00:15:53.510 ⇒ 00:16:04.200 Amber Lin: and that’s why I was interning at a lot of random different places. And I worked at also many different functions. I worked in sales, finance.
65 00:16:04.678 ⇒ 00:16:09.460 Amber Lin: a bit of data, and then ended up in consulting. And ey.
66 00:16:09.690 ⇒ 00:16:19.320 Amber Lin: And then I was like, you know what? I think. I actually wanna do something. If, apart from just giving advice. And hence now, I’m at a
67 00:16:19.690 ⇒ 00:16:35.139 Amber Lin: AI and data consultancy. But still we do a lot more implementation. And that’s what I really like is that we can find problems. And then we actually do something about it, whether it’s client project or our internal projects as well, and that that just so interesting.
68 00:16:36.590 ⇒ 00:16:54.980 Emily Chan: Yeah, that really resonate with me. I think it’s like the data part like to me is, it’s like an experiment. It’s almost like cooking right? I want to like, see the impact, and then I can adjust it in the next cycle when I was like in finance or as a consultant it. Just you leave the deck, and you don’t know what happens right? But then.
69 00:16:55.750 ⇒ 00:17:00.660 Emily Chan: Part of the whole process. You don’t just leave it. It makes it more fun when you can see, like the results.
70 00:17:00.660 ⇒ 00:17:28.539 Uttam Kumaran: Way, more fun way, more fun. And I and I think you know, in in our business like our, we’re getting into the some of our engagements. Finally, where they’re actually maturing to the point where we not only come in, we’ve established data warehouse. We landed data. It’s all modeled in Dbt, we have data marks. There’s definitions. There’s dashboards. And then finally, we get to the point where we’re like, let’s go find. Let’s go dig right? You know, and I think.
71 00:17:28.560 ⇒ 00:17:50.080 Uttam Kumaran: of course, like my expectations, for how fast we can do that are extremely high. I think we are. And I say this all the time about how, for the value that we deliver, and being able to deliver on every single part of that stack, it’s not very common for consulting firms to do both. In fact, I would say most firms do.
72 00:17:50.500 ⇒ 00:17:56.369 Uttam Kumaran: I would say there’s a lot of firms that do the engineering piece, and then there’s some small set of some of the firms that do
73 00:17:56.490 ⇒ 00:18:09.379 Uttam Kumaran: the data piece poorly, like the insights piece, because they they don’t. They’re like surface level using bad data, and they sort of make. They just kind of leave you with a deck. We’re incredibly underpriced, and we do everything
74 00:18:09.795 ⇒ 00:18:27.774 Uttam Kumaran: like. Our offering is very wide, you know. That’s that’s helped us and hurt us. In fact, it’s helped us in that we can come into any engagement, and and we find a way to be valuable, I would say, where? Where? It’s where it could be, you know, I would say, challenging sometimes, is that
75 00:18:28.240 ⇒ 00:18:39.859 Uttam Kumaran: we we find ourselves having to do all that work which some companies don’t see ex see the value in on until the point at which we can actually
76 00:18:40.340 ⇒ 00:18:59.379 Uttam Kumaran: you know, like basically demonstrate to them that there is a that there is going to be an Roi because of because of the insights piece. So yeah, that’s a little bit about like kind of where we’re. I don’t know amber, even in your in your time, reflecting now, like, if when you join the company like, if this is actually like
77 00:18:59.550 ⇒ 00:19:10.769 Uttam Kumaran: what you thought it was gonna be, or if you now have a different idea, I don’t remember really what I explained about what we do. Did anything change, you know? I think, like even I’m just curious.
78 00:19:11.160 ⇒ 00:19:23.979 Amber Lin: Hmm! Well, when I joined, when I 1st talked to Robert I got the connection to this job because me and Robert went to the same program, and Robert wanted to hire me as a data analyst.
79 00:19:23.980 ⇒ 00:19:42.360 Amber Lin: And then I talked to Utam, and Uta was talking about. Okay, we want to improve the efficiency of this business. We want to have process in place. We want to automate things like cool. I like that. And I was like, please, please, please. I’ve never been a project manager. But can you please make me a project manager? And so I became a project manager.
80 00:19:43.000 ⇒ 00:19:57.070 Amber Lin: So 1st off the valley. Everything was kind of new for me, but I do think within the time that I joined the reason why it has so much fun was because we’re setting a lot of processes to transform from
81 00:19:57.445 ⇒ 00:20:16.960 Amber Lin: a group of freelancers to a company. So there’s a lot of company building components that happened in the past 3 months. Whether it’s bringing structure to climb projects. Whether it’s bringing structure to starting off on the initial internal initiatives. We have AI initiatives, and we have data, initiatives
82 00:20:17.020 ⇒ 00:20:26.959 Amber Lin: and and overall to restructure how we do sales and allowing Usham and Robert to have more time to do sales. So I really saw the whole
83 00:20:27.070 ⇒ 00:20:29.420 Amber Lin: company transform.
84 00:20:29.600 ⇒ 00:20:31.159 Amber Lin: And for me.
85 00:20:32.080 ⇒ 00:20:38.440 Amber Lin: that’s a big part of what I what I like being involved in and setting the initial
86 00:20:38.600 ⇒ 00:20:51.469 Amber Lin: setting, the initial systems and looking at. Okay, these are all the initiatives. Why we should and why they’re important. And seeing seeing that get pushed along. So that’s for that for me, was
87 00:20:51.620 ⇒ 00:20:58.890 Amber Lin: what was interesting and what was different since when I joined, because to me, like, everything’s just happened within these 3 months.
88 00:20:59.330 ⇒ 00:21:01.450 Amber Lin: so I don’t know what happened before.
89 00:21:02.400 ⇒ 00:21:06.449 Uttam Kumaran: It’s just it’s same slower.
90 00:21:07.640 ⇒ 00:21:24.319 Emily Chan: That’s that’s really exciting to hear that the company is going, basically scaling up and becoming a company. Can you maybe like talk about an example of that I know you talk about like internal AI initiative, internal data initiative like better sales process. Do you mind digging into one of those.
91 00:21:24.320 ⇒ 00:21:27.590 Uttam Kumaran: Yeah, is there one of those that’s interesting and and
92 00:21:27.930 ⇒ 00:21:32.059 Uttam Kumaran: maybe amber? I would love for you to go, and I maybe I could fill in any gaps.
93 00:21:32.590 ⇒ 00:21:38.619 Amber Lin: Okay, yeah. Totally. So recently, something I’ve been working on aside of
94 00:21:38.790 ⇒ 00:22:03.529 Amber Lin: client projects is internal strategy right? And it kind of aligns for me. It defines all the stuff of why and what we’re doing. So as a service business, we have, we need to have great client service. We need to also 2, we need to have employee satisfaction because people is our product. And lastly, to tile things, everything together, we need good financial performance.
95 00:22:03.921 ⇒ 00:22:23.370 Amber Lin: And I guess, since I’m also a project manager, I’ll talk about the 1st part, about great client service. And for that it’s okay. What makes what does it mean to have great client service? What it makes a client happy? And ultimately we found that it’s actually about a psychology
96 00:22:23.750 ⇒ 00:22:52.750 Amber Lin: of how a client feels versus a lot of the engineering work, you actually you you would be surprised to know that actually, it’s 20% of the work which is communication. Making sure the client feels safe, feels heard knows that that they know that what’s gonna happen and knows that this is essentially, that you have their back, and all of that determines a huge part of 80% of what the client feels to
97 00:22:52.950 ⇒ 00:23:16.239 Amber Lin: feel like they they got value from. You felt that this is a great experience and willing to pay even more. But that only takes 20% of the work. 80% of like 20% of the work mostly is on me to communicate, to set those rituals, to make sure we have decks. We have communication guidelines, like all that stuff takes like 20, but has 80% impact.
98 00:23:16.420 ⇒ 00:23:29.079 Amber Lin: like our engineers work their ass off to do all this work. But then, if we don’t do that, 20%, the clients will leave because they they don’t see the value in this because ultimately we’re client service business.
99 00:23:29.530 ⇒ 00:23:48.249 Amber Lin: and that’s to set. That’s like the reasoning behind some of the initiatives that we want to start of. Okay, how frequently do we communicate to clients? How frequently do we meet with clients? How? What stakeholders should we get in touch with do we send meeting summaries every single day, or every single week?
100 00:23:48.370 ⇒ 00:24:11.610 Amber Lin: And how do we? What else? How do we onboard them. So we have the scoping document, making sure that everything is addressed. How do we create tickets, and how involved should the client be on our work, even of getting maybe getting their internal data team to collaborate with us as well. So all of these initiatives are
101 00:24:12.360 ⇒ 00:24:21.009 Amber Lin: sort of based on the core value of okay, we want the client to be happy psychologically.
102 00:24:23.340 ⇒ 00:24:24.340 Emily Chan: That’s important.
103 00:24:24.550 ⇒ 00:24:31.259 Amber Lin: Yeah, I hope that answers, it’s somewhat similar to like.
104 00:24:32.670 ⇒ 00:24:36.319 Uttam Kumaran: Yeah, it’s similar to managing up when you’re in a company except
105 00:24:36.874 ⇒ 00:24:41.100 Uttam Kumaran: as a consultant, we’re we’re always in chopping block, meaning like,
106 00:24:42.910 ⇒ 00:24:48.489 Uttam Kumaran: People view us with a lot of disdain just from what they probably have experience working with consultants.
107 00:24:48.760 ⇒ 00:24:58.979 Uttam Kumaran: First, st job is to change the narrative to them, for them to trust us. And then it’s to really demonstrate that like, Oh, my! Gosh! Like these guys are are really the smartest folks in the room.
108 00:24:59.341 ⇒ 00:25:04.879 Uttam Kumaran: And we we haven’t done a good job at that at times, like, you know, that’s this is something that we’re learning
109 00:25:05.261 ⇒ 00:25:23.630 Uttam Kumaran: and I think the way, Amber articulated it is, is exactly like some of the things we’re trying to do. And I would say, that’s just related to the client engagement. Right? We’re. I’m we’re thinking about, how do employees think about the company? Internally? I’m thinking about, how do prospects and potential leads? Think about the company and brand building
110 00:25:24.030 ⇒ 00:25:49.510 Uttam Kumaran: and then the I would say, across, everything is AI like we are, gonna stretch AI to the absolute limits. Right? And I feel very lucky that I can run the company, and I can set the vision, and that I can force us to use AI in every nook and cranny of the business and we’re gonna run, you know, one of the few really purely like AI, native or AI in the DNA
111 00:25:49.570 ⇒ 00:26:00.840 Uttam Kumaran: consulting companies that exist. You know you’re gonna you. You will read a lot about other companies sort of trying to do what we’re doing. There will be companies like Ey and Accenture and Deloitte, who will
112 00:26:01.140 ⇒ 00:26:08.079 Uttam Kumaran: roll out. Oh, we rolled out, chat, gpt for all our employees. They’ll stop there. There’ll be also small freelancers who will say cool. I can like
113 00:26:08.240 ⇒ 00:26:36.049 Uttam Kumaran: I can. I’m now like I’m able to handle multiple clients. Now. It’s it’s a different game. When it’s like, I want to use AI across sales, employees, marketing delivery operations, finance right? And every single one of our team gets asked about how how they can use AI to scale themselves. Because for me, that’s the only way we’re gonna survive and and continue to build a lead, you know.
114 00:26:36.856 ⇒ 00:26:40.919 Uttam Kumaran: Past all the other sort of consultants in our in our world.
115 00:26:42.560 ⇒ 00:26:57.899 Emily Chan: I’m curious. How do you think about like using an external AI tool versus like building something an AI agent on your own, because every of one of your team member is smart, right? You cannot obviously just build an AI agent like, How how do you balance that.
116 00:26:58.610 ⇒ 00:27:15.809 Uttam Kumaran: Yeah, so we we we do both, you know. So 2 things we do one. And originally my thesis was like, Okay, I’m going to build AI agents for everybody, and then people will start to use them. But what I actually realized, it’s like, actually, part of it is just getting people just to even use chat. Gpt to do basic stuff
117 00:27:15.870 ⇒ 00:27:41.099 Uttam Kumaran: right. Use granola for meeting notes, use chat, Gpt. Use loom and use AI to trim your meetings. It’s like, Use AI to get through use as AI as a 1st escalation point for your problems. So part of it was just purely education, like, we’re gonna get 10 to 20 to 30% efficiency gains over the competitor just based on that alone. The next piece is any process that we’re doing multiple times or that
118 00:27:41.160 ⇒ 00:27:57.270 Uttam Kumaran: we are. We have the urge to hire, for we will try to use either an off the shelf AI tool or build our own. We’re not a product company. So I actually don’t have much interest in building products. But there are a lot of the things that we need to do at the price that we can afford do not exist.
119 00:27:57.350 ⇒ 00:28:19.596 Uttam Kumaran: So like as as a great engineering team, we’ll build it for ourselves, but building it for ourselves is not like building off building ui. These are slack bots or quick integrations, or zapier. You know, agentic workflows like this is just whatever it takes to get it to work. We’re using open source tech. We’re self hosting a lot, you know.
120 00:28:20.190 ⇒ 00:28:41.569 Uttam Kumaran: so yeah, we’re just sort of scrappy, because these are all internal tools, like none of this are the the thing is not to build something cool for coolness sake. It’s for our clients to win. So I want our clients to get a faster product. I want them to get a better product. I want us to be able to serve more clients, and I wanted to do best while running like a healthy P. And L like, that’s that’s it.
121 00:28:42.160 ⇒ 00:28:54.510 Emily Chan: Yeah, yeah, I hear you. Do you think you know, maybe you can like kind of playbook some of this like, Hey, this is how my agency use AI throughout all process. Mr. Client, maybe you can use this. Use. AI, this.
122 00:28:54.510 ⇒ 00:29:00.449 Uttam Kumaran: You’re you’re totally right. Yeah. And that’s exactly what we’ve done. So I I don’t know when the last time we spoke, but we
123 00:29:00.460 ⇒ 00:29:19.990 Uttam Kumaran: where we are. We’re both. We’re doing both data and AI, you know, services for clients. So for some clients, we’re purely building AI agents. But there is a data component where we’re measuring evaluations. We’re measuring response times. You know, accuracy. So exactly right, the reason we got into doing AI work for clients
124 00:29:19.990 ⇒ 00:29:30.879 Uttam Kumaran: is because we were applying AI internally, and we found that it was working. But I also found that I was having a challenge getting our team to adopt it. I was having a challenge
125 00:29:30.880 ⇒ 00:29:55.020 Uttam Kumaran: finding out what tools actually work. And I was like, if I’m on the edge and I’m having trouble, then your average company is gonna have a real struggle. So that’s exactly what we did. So you know, we hired our 1st full time AI engineer in like in September. And now we we just we’re gonna bring on one more. So we’ll have 3 full time. AI people on the team that all they do is build
126 00:29:55.260 ⇒ 00:29:57.870 Uttam Kumaran: AI workflows for clients or for us internally.
127 00:29:58.420 ⇒ 00:30:03.829 Emily Chan: That’s exciting. It’s a really exciting time to be testing things out, pushing things out in AI.
128 00:30:03.990 ⇒ 00:30:13.279 Uttam Kumaran: Yeah, yeah, I. So I actually have to. We have a client meeting coming up right now. I have to actually run to
129 00:30:13.686 ⇒ 00:30:22.849 Uttam Kumaran: but would love to continue this. I know this is short, like I feel like we didn’t get to talk about anything, so I don’t know would love, to catch up again.
130 00:30:23.010 ⇒ 00:30:28.830 Emily Chan: Yeah, I just want to, you know, stay connected. See what my availability is and what you need.
131 00:30:28.830 ⇒ 00:30:29.480 Uttam Kumaran: Yeah.
132 00:30:29.630 ⇒ 00:30:30.340 Emily Chan: How I can help.
133 00:30:30.340 ⇒ 00:30:52.590 Uttam Kumaran: I didn’t even get to ask anything about how yeah, how stuff is going. Maybe, like, I wanna send a couple of notes over over email. And then I mean my time, I’m just like running around between things. But like, even if I can chat on the phone quickly, or something would love to hear and just see, like kind of like where you’re at and just tell you like. And any questions any more questions you have about how things are going happy to answer.
134 00:30:52.590 ⇒ 00:30:58.920 Emily Chan: Yeah, yeah. So why don’t we, you know, continue this over email, and we can jump on the phone to let me know what time works.
135 00:30:59.180 ⇒ 00:31:00.610 Uttam Kumaran: Okay, okay, perfect.
136 00:31:00.610 ⇒ 00:31:01.150 Emily Chan: Alright!
137 00:31:01.150 ⇒ 00:31:04.500 Amber Lin: Okay, CC, me, on the email connected.
138 00:31:04.500 ⇒ 00:31:06.170 Uttam Kumaran: Definitely definitely.
139 00:31:08.710 ⇒ 00:31:10.810 Amber Lin: That’s good got a hop. Thank you.
140 00:31:11.810 ⇒ 00:31:12.340 Emily Chan: Oh!
141 00:31:12.340 ⇒ 00:31:13.330 Uttam Kumaran: Thanks. Everyone.
142 00:31:13.720 ⇒ 00:31:14.460 Amber Lin: Bye.