Meeting Title: Brainforge Interview w- Greg Date: 2026-02-18 Meeting participants: Advait Nandakumar Menon, Greg Stoutenburg
WEBVTT
1 00:02:52.900 ⇒ 00:02:54.110 Greg Stoutenburg: Hello?
2 00:02:54.110 ⇒ 00:02:55.820 Advait Nandakumar Menon: Hey, Grant, hello!
3 00:02:55.820 ⇒ 00:02:57.490 Greg Stoutenburg: Hey, great to meet you, Edvate.
4 00:02:57.900 ⇒ 00:03:00.189 Advait Nandakumar Menon: to meet you as well. How are you doing?
5 00:03:00.490 ⇒ 00:03:02.300 Greg Stoutenburg: I’m doing well today, how are you?
6 00:03:03.030 ⇒ 00:03:05.300 Advait Nandakumar Menon: I’m doing good, I’m doing good, thanks for asking.
7 00:03:05.300 ⇒ 00:03:07.700 Greg Stoutenburg: Yeah. Where are you located?
8 00:03:08.420 ⇒ 00:03:10.219 Advait Nandakumar Menon: I’m in Cincinnati.
9 00:03:10.390 ⇒ 00:03:12.730 Greg Stoutenburg: Okay, alright, great. I’m from Detroit, so…
10 00:03:12.730 ⇒ 00:03:16.659 Advait Nandakumar Menon: Oh, okay, okay, yeah, it’s a little further up, yeah? Yeah. Yeah.
11 00:03:16.850 ⇒ 00:03:27.119 Greg Stoutenburg: I live in Central Pennsylvania now, but from Detroit originally, so, anytime I hear, you know… Now, it used to be, you know, Ohio-Michigan rivalry, but now it’s like Ohio, like, oh, that’s sort of near home.
12 00:03:27.120 ⇒ 00:03:30.299 Advait Nandakumar Menon: Heh Yeah, yeah, right, right, yeah.
13 00:03:31.880 ⇒ 00:03:38.889 Greg Stoutenburg: Before we jump in, I just wanted to confirm. So, you… have you… you’ve applied for the Senior Associate for Data Insights position?
14 00:03:39.740 ⇒ 00:03:45.239 Greg Stoutenburg: reading the right thing? Okay, alright, I just want to make sure I was looking at the right thing here. Okay, so…
15 00:03:45.790 ⇒ 00:03:47.240 Greg Stoutenburg: One second, please.
16 00:04:07.430 ⇒ 00:04:08.300 Greg Stoutenburg: Okay.
17 00:04:11.570 ⇒ 00:04:17.809 Greg Stoutenburg: All right, cool. Okay. Yeah, let’s, so, yeah, let’s jump in. So,
18 00:04:18.350 ⇒ 00:04:24.319 Greg Stoutenburg: I guess maybe just take, like, 2 minutes and just tell me, you know, how did you end up here, talking to me on this interview?
19 00:04:25.140 ⇒ 00:04:38.709 Advait Nandakumar Menon: Sure. So, a little intro about me, I’ll keep it brief. So, I have about four and a half to five years of experience in BI, data analytics, engineering, and consulting in general.
20 00:04:38.710 ⇒ 00:04:58.340 Advait Nandakumar Menon: So, I started at Tarta Consultancy Services, majorly working on dashboards and customer analytics, and later moved into a data engineering role where I realized, I love analytics more, and even though I was focusing on pipelines, migrations, and performance improvements, I realized I want to do business-facing analytics.
21 00:04:58.340 ⇒ 00:05:10.299 Advait Nandakumar Menon: So, most recently, I worked as a data analyst at a consulting startup. So, there, I was supporting clients across, insurance, SaaS, manufacturing, and non-profit sectors.
22 00:05:10.300 ⇒ 00:05:27.529 Advait Nandakumar Menon: A big part of my role was basically taking unclear questions from the business and turning them into something, really, that they could use daily. So, whether that was dashboards, automation, or reporting tied to real business decisions, basically.
23 00:05:27.530 ⇒ 00:05:37.449 Advait Nandakumar Menon: So, an example of that would be a pipeline visibility dashboard that I worked on, which helped improve their overall forecast accuracy for that insurance client, which I mentioned.
24 00:05:37.450 ⇒ 00:05:54.369 Advait Nandakumar Menon: And I also built reporting systems that replaced manual processes and saved teams significant amount of time. So, I have found that I really enjoy work where there’s ambiguity and the end goal is to create something practical that gets adopted in the long run.
25 00:05:54.510 ⇒ 00:06:12.980 Advait Nandakumar Menon: So, going forward, I am really looking to continue growing in those roles, and really take ownership of the problems, and work very closely with the stakeholders, and basically build practical data and solutions, which this role at Brainforge is what looks like it.
26 00:06:13.830 ⇒ 00:06:18.970 Greg Stoutenburg: Yeah, okay, great, yeah, excellent, yeah, excellent, yeah, and in this interview, so this is round two for you, I believe, right?
27 00:06:18.970 ⇒ 00:06:19.549 Advait Nandakumar Menon: Huh.
28 00:06:19.550 ⇒ 00:06:31.310 Greg Stoutenburg: Yeah, okay, yeah, this is, so this is really just, like, focused on, on the role specifically, rather than, like, a more general, like, you know, background, kind of role, but yeah, so I’ll just ask questions that are specific to the role.
29 00:06:31.310 ⇒ 00:06:40.809 Greg Stoutenburg: And just, you know, anywhere you can give specific answers and tell a story about your experience, you know, always good, so go for it.
30 00:06:40.810 ⇒ 00:06:41.140 Advait Nandakumar Menon: Right.
31 00:06:41.140 ⇒ 00:06:48.679 Greg Stoutenburg: So, let’s go for, like, problem structuring, because you kind of just mentioned this sort of thing. How do you turn a vague question into a plan for analysis?
32 00:06:49.830 ⇒ 00:07:03.930 Advait Nandakumar Menon: Okay, so how I approach this is any way question. I start with splitting it into different simple parts so that we can really, tackle it one by one. So, an example of this is when,
33 00:07:03.930 ⇒ 00:07:14.059 Advait Nandakumar Menon: when the insurance client that I was working with, they said they want better visibility into their pipeline, and they didn’t really expand much more than that, and
34 00:07:14.060 ⇒ 00:07:32.229 Advait Nandakumar Menon: that’s when I sat down with each of the stakeholders. So, this dashboard I worked on didn’t just, work for the RevOps team, it was for the end sales reps as well, so the managers, the sales executive, and everyone. So everyone had, like, different definitions of the key metrics, what they were trying to look at.
35 00:07:32.230 ⇒ 00:07:46.589 Advait Nandakumar Menon: So, I really had one-on-one sessions with them, and tried to really understand what their pain points and what they’re trying to do with their data, basically, and turn those vague metrics or definitions they had into structured ones by working closely with them.
36 00:07:46.590 ⇒ 00:08:01.399 Advait Nandakumar Menon: And that basically ended up in a clean dashboard, wherein the definitions and the metrics were aligned with all the stakeholders, and they were really able to get to the numbers they wanted to see. So, that’s how I would approach a vague business question and structure it.
37 00:08:01.400 ⇒ 00:08:02.120 Greg Stoutenburg: Yeah, cool.
38 00:08:02.250 ⇒ 00:08:18.469 Greg Stoutenburg: Yeah, and I mean, have you found in your experience that there’s, that stakeholders tend to align on, you know, once you’ve clarified the question and showed them what you think they want to see, that they’ve tended to align at that time, or do you have to do more work to get everyone on the same page?
39 00:08:18.790 ⇒ 00:08:33.689 Advait Nandakumar Menon: Yeah, I mean, that really depends on the client and the stakeholders. So, usually what I have seen earlier in my career is, I am not really focused on asking the right questions at first, so…
40 00:08:33.760 ⇒ 00:08:46.239 Advait Nandakumar Menon: this resulted, even though I gave a technically correct dashboard or automation or system, they were really confused about the numbers not matching or disagreeing on some of the stuff, so…
41 00:08:46.260 ⇒ 00:08:56.039 Advait Nandakumar Menon: Now that I have, like, recently matured in my career, I have seen that working with them closely at first, really getting to know their pain points and asking the right questions.
42 00:08:56.110 ⇒ 00:09:09.229 Advait Nandakumar Menon: In the long run, it avoids many of the confusions or follow-ups, but there can be instances, like I said, where they can come back to us and maybe not happy with the number, but yeah, it depends on the client, I would say.
43 00:09:09.400 ⇒ 00:09:14.939 Greg Stoutenburg: Yeah, yeah, yeah, okay. Talk me through a messy problem that you structured well.
44 00:09:15.950 ⇒ 00:09:25.900 Advait Nandakumar Menon: Okay, so one messy problem, that I structured well would be, with another client we were working with, it’s a manufacturing plant.
45 00:09:25.910 ⇒ 00:09:38.999 Advait Nandakumar Menon: And they were in the supply chain space as well. So, over there, I was really working with the CFO and the CEO, and they wanted, basically, to, look at their,
46 00:09:39.710 ⇒ 00:09:53.369 Advait Nandakumar Menon: data which was sitting in their ERP system as well as QuickBooks. So, the overall data set was really messy in the sense that there was no real connection between the ERP system and the QuickBooks data. So, they…
47 00:09:53.370 ⇒ 00:10:03.629 Advait Nandakumar Menon: I had… I had to step in, basically, to create a connection between, the ERP dataset and the QuickBooks Finance, so that included data mapping and
48 00:10:03.630 ⇒ 00:10:26.090 Advait Nandakumar Menon: mapping the right fields, and of course, before all of this, I had to sit down with both of them and understand what they’re trying to automate. So, they were really focused on automating their manual Excel spreadsheet into a dashboard which they can just look at and get insights from. So, the messy data is dealing with really, like I said, the ERP and the QuickBooks data.
49 00:10:26.090 ⇒ 00:10:31.030 Advait Nandakumar Menon: I had to structure it into a rigid schema and bring it over
50 00:10:31.030 ⇒ 00:10:41.119 Advait Nandakumar Menon: Power BI dashboard is what I used, so bring it over to Power BI service, and, really that, cut away, like, 12 plus hours on a,
51 00:10:41.120 ⇒ 00:10:50.050 Advait Nandakumar Menon: monthly basis, from their queue, that they were spending to, create manual reports. So, that’s one of the messy data sets I worked with.
52 00:10:50.620 ⇒ 00:10:52.680 Greg Stoutenburg: Yeah. How did,
53 00:10:53.150 ⇒ 00:10:59.339 Greg Stoutenburg: At the end of that project, what were you able to do to make that messy data a little cleaner?
54 00:10:59.450 ⇒ 00:11:04.109 Greg Stoutenburg: Or rather, you know, the multiple data sources and data sort of all over the place, what were you able to do to make.
55 00:11:04.110 ⇒ 00:11:04.490 Advait Nandakumar Menon: Yeah.
56 00:11:04.490 ⇒ 00:11:06.149 Greg Stoutenburg: Or effective for stakeholders.
57 00:11:06.400 ⇒ 00:11:13.340 Advait Nandakumar Menon: Yeah. Do you, mean from a technical standpoint, or, like, how are you, asking about that?
58 00:11:13.540 ⇒ 00:11:33.450 Greg Stoutenburg: So I’m asking from a technical standpoint, but also just, like, from a product… a project leadership standpoint. Like, so if… if what you’ve been given is this sort of big, messy tangle, what did you do to make… to improve that so that then you can go to stakeholders or the client and go, okay, things were this way, now here’s how I’ve made them better.
59 00:11:33.920 ⇒ 00:11:47.449 Advait Nandakumar Menon: Right. So, like I said, I gathered the requirements from them, like which data points or customer-facing data that they are trying to look at. I get those requirements from them, then I get into the technical work.
60 00:11:47.450 ⇒ 00:11:57.739 Advait Nandakumar Menon: Basically, I was standardizing the fields, validating the data and cleaning it. This was really, within Power BI, I was going… using Power Query.
61 00:11:57.790 ⇒ 00:12:08.950 Advait Nandakumar Menon: and a little of DAX as well. So, once I’ve structured the data and the transformation in the Power BI layer, I bring it into the semantic modeling layer, basically,
62 00:12:08.950 ⇒ 00:12:19.400 Advait Nandakumar Menon: One example would be, I created a star schema wherein the dimension table was surrounded by the fact tables, so that’s really the technical,
63 00:12:19.470 ⇒ 00:12:33.549 Advait Nandakumar Menon: work I put into this, trying to really match different data sets and see what’s going on, and the end result is that the business really, like, the CEO and the CFO doesn’t really need to know what’s going on technically, unless they really want to, which is another case, but…
64 00:12:33.550 ⇒ 00:12:45.080 Advait Nandakumar Menon: They really want to get to the right numbers, and that’s where the semantic model… modeling helps and get to the right answers and… for the questions that they have, so that’s something I did.
65 00:12:45.590 ⇒ 00:12:53.860 Greg Stoutenburg: Yeah, yeah, okay, great, yeah. But like you said, you’re less interested in that sort of work now, and want to do more of the business-facing reporting. Right.
66 00:12:53.990 ⇒ 00:13:13.520 Advait Nandakumar Menon: I would say it’s a mix of both, like, I started out technical, and I’ve matured now into a role, or into this part of my career where I want to do a mix of both. So, I would, say I want to wear a lot of hats and own the business aspect as well as the technical aspect. So, a mix of both is
67 00:13:13.520 ⇒ 00:13:15.129 Advait Nandakumar Menon: I would aptly put it as.
68 00:13:15.320 ⇒ 00:13:18.240 Greg Stoutenburg: Yeah, yeah, yeah. Okay, great.
69 00:13:20.100 ⇒ 00:13:23.280 Greg Stoutenburg: What would you say makes an analysis actionable?
70 00:13:23.750 ⇒ 00:13:25.610 Greg Stoutenburg: For those business stakeholders.
71 00:13:27.270 ⇒ 00:13:39.180 Advait Nandakumar Menon: So, that’s a good question. Like, there have been instances in my career, like, when I started out, where I built a full-blown technical product, like a dashboard or whatever.
72 00:13:39.180 ⇒ 00:13:39.860 Advait Nandakumar Menon: But…
73 00:13:39.860 ⇒ 00:14:03.570 Advait Nandakumar Menon: seeing it not getting adopted, or really helping any decision, or saving money, time, or whatever. So, I would say analysis is actionable for me when that is being really getting adopted, like a dashboard is getting adopted, it’s driving business decisions, or automating some process, or saving money for the end client. So, I would say that’s when an analysis
74 00:14:03.570 ⇒ 00:14:05.070 Advait Nandakumar Menon: It’s actionable for me.
75 00:14:05.070 ⇒ 00:14:11.900 Greg Stoutenburg: Yeah, yeah, yeah, okay, okay, yeah. Because it’s pointing to… so it’s pointing to ways to improve on…
76 00:14:12.010 ⇒ 00:14:14.559 Greg Stoutenburg: Key metrics, like that, yeah. Yeah.
77 00:14:14.560 ⇒ 00:14:14.930 Advait Nandakumar Menon: Yeah.
78 00:14:16.090 ⇒ 00:14:32.849 Greg Stoutenburg: how do you ensure that clients can operate your work without you? So, put differently, one of the goals is always at least some amount of… enabling some amount of self-service for clients using, like, BI tools and reports and things like that. How do you ensure… how do you enable that?
79 00:14:33.900 ⇒ 00:14:47.730 Advait Nandakumar Menon: So, once… so, generally, what I’ll do is, once, we hand off the dashboard or whatever system you have built to the clients or business users, I, lead a
80 00:14:47.950 ⇒ 00:15:00.629 Advait Nandakumar Menon: enablement session, so as they know, like, how to use the dashboards, like, some people, they really don’t know how… they’re coming from an Excel spreadsheet or whatever, and they’re new to Power BI or a Tableau dashboard, so…
81 00:15:00.630 ⇒ 00:15:10.589 Advait Nandakumar Menon: I lead enablement sessions, in which the main leaders or the stakeholders will join as well, and really try to help,
82 00:15:10.600 ⇒ 00:15:23.410 Advait Nandakumar Menon: the sales reps, or whoever is going… using the end dashboard, trying to walk them through the definitions of the key metrics, the logics that’s being used in coming up with the numbers for those metrics, and really
83 00:15:23.620 ⇒ 00:15:42.260 Advait Nandakumar Menon: teaching them how to use the dashboard, basically. Another instance of enablement I have done is for the same manufacturing plant. So, there were some business users who used to come to us from time to time to ask basic questions or ad hoc analysis they want to do.
84 00:15:42.260 ⇒ 00:15:48.899 Advait Nandakumar Menon: So, which the dashboard really serves the purpose, but maybe they are not that technically advanced to get to the right answer.
85 00:15:48.900 ⇒ 00:16:03.889 Advait Nandakumar Menon: So, what I did is I implemented the Power BI Copilot feature within those dashboards, so it involved using the Q&A feature in Power BI, and some semantic modeling, and the Power BI service itself, where the Copilot agent exists.
86 00:16:03.890 ⇒ 00:16:17.680 Advait Nandakumar Menon: So, implementing that on a semantic layer, like, enable those business users to ask questions in natural language. It’s like chatting with ChatGPT, basically, but in this case, it’s restricted to your own data and that semantic model.
87 00:16:17.680 ⇒ 00:16:25.600 Advait Nandakumar Menon: So, they can ask, like, what’s the top sales, this month, top orders, and things like that, and get real-time answers without…
88 00:16:25.740 ⇒ 00:16:28.899 Advait Nandakumar Menon: My requirement over there. So, yeah.
89 00:16:29.380 ⇒ 00:16:35.430 Greg Stoutenburg: Yeah, yeah, yeah, okay, great, fantastic. So, sounds like you’re familiar with thinking through what the user needs to see.
90 00:16:35.760 ⇒ 00:16:40.130 Greg Stoutenburg: And ensuring that you can put that in front of them. When you did those enablement sessions, how did you structure them?
91 00:16:40.130 ⇒ 00:16:40.680 Advait Nandakumar Menon: Huh.
92 00:16:41.300 ⇒ 00:16:56.970 Advait Nandakumar Menon: So, I would first, give an overview, from myself, like, what the dashboard or the technical solution is about, like, what process it’s trying to make easy, or what business questions it’s trying to answer.
93 00:16:56.970 ⇒ 00:17:06.780 Advait Nandakumar Menon: I will… so on all my dashboards, I’ve had this FAQ page wherein they can regularly check if they have any doubts about the KPIs or any formulas that’s going into it.
94 00:17:06.780 ⇒ 00:17:18.789 Advait Nandakumar Menon: So I will give a rundown of the FAQ page, and then the later part of the session would be basically me and the stakeholder going through each page of the dashboard and
95 00:17:18.790 ⇒ 00:17:34.409 Advait Nandakumar Menon: basically teaching them how to use the filters, or… this is for people who are really not that technical, by the way. So, yeah, it’s a combination of overview of what the dashboard does, a little business logic, and what was going behind the scenes to come up with the numbers.
96 00:17:34.410 ⇒ 00:17:40.180 Advait Nandakumar Menon: And the tutorial itself is how I have structured in my enablement sessions.
97 00:17:40.390 ⇒ 00:17:42.829 Greg Stoutenburg: Yeah, yeah, yeah. And have they tended to go well?
98 00:17:43.220 ⇒ 00:18:00.820 Advait Nandakumar Menon: Yeah, for the most part, it has gone well. Like, there are always instances wherein some of the reps or stakeholders would come back to us saying the number isn’t right, or they’re, like, not applying the right filters to get to their data point. So there have been instances like that, but
99 00:18:00.820 ⇒ 00:18:10.099 Advait Nandakumar Menon: It’s not a direct result of the enablement. I would say it’s a question apart from, like, the enablement that they had after the session, so, yeah.
100 00:18:10.100 ⇒ 00:18:20.089 Greg Stoutenburg: Yeah, yeah, yeah, okay, yeah, sure, fair. Tell me about when you choose spreadsheets over a BI tool. When would you just make a spreadsheet?
101 00:18:20.650 ⇒ 00:18:21.290 Advait Nandakumar Menon: Right.
102 00:18:21.700 ⇒ 00:18:36.509 Advait Nandakumar Menon: So this, depends on the client’s stack, first of all, because, one instance I can give you in my most recent role is, we were working with a non-profit organization.
103 00:18:36.510 ⇒ 00:18:50.489 Advait Nandakumar Menon: And as you can imagine, nonprofits don’t want to invest that much. They want to save as much as money they can. So this particular nonprofit, they were doing some manual transformations by hand in Excel.
104 00:18:50.490 ⇒ 00:18:58.229 Advait Nandakumar Menon: to import their data from an event platform to their custom donor CRM. So, this was a donation platform that they were running.
105 00:18:58.230 ⇒ 00:19:12.440 Advait Nandakumar Menon: So, there was… there were two options, really, we were giving them. Like, there is an API from the event platform that automatic… automates all of that process and just dumps the data into the CRM platform, but that was pretty expensive for them.
106 00:19:12.440 ⇒ 00:19:19.440 Advait Nandakumar Menon: Around $5,000 or something in cost, which, obviously, as a non-profit, you want to try to avoid, so…
107 00:19:19.550 ⇒ 00:19:30.869 Advait Nandakumar Menon: I went with a simpler solution of using an Excel template wherein, they just have to drop in the data, and using lookups, data validation, and…
108 00:19:30.870 ⇒ 00:19:39.760 Advait Nandakumar Menon: conditional aggregations, the data automatically gets transformed, and they can just save it as a CSV file and import it on the CRM platform, so…
109 00:19:39.860 ⇒ 00:19:56.649 Advait Nandakumar Menon: That’s… so, simplicity is the goal, like, if the client is okay with that simple of a tool, then that’s fine. So, the end result is making them satisfied, and if that’s saving them money as well in the long run, then that’s good enough.
110 00:19:57.100 ⇒ 00:20:06.820 Greg Stoutenburg: Yeah, yeah, yeah, yeah, great, excellent, excellent. Okay, what is your approach to getting some quick wins in the first 30 days of a new engagement?
111 00:20:08.240 ⇒ 00:20:11.409 Advait Nandakumar Menon: Quicklins. Okay, that’s a good question. So…
112 00:20:12.170 ⇒ 00:20:15.489 Advait Nandakumar Menon: How I would look at it, like, is,
113 00:20:15.730 ⇒ 00:20:28.600 Advait Nandakumar Menon: before getting really deep into a project, I would see, like, what can be done immediately, like, what can be delivered immediately, and what can be iterated upon later, so…
114 00:20:28.760 ⇒ 00:20:47.300 Advait Nandakumar Menon: earlier, I had this, in my career when I started out. Like I said, I was focused on delivering well-rounded technical solutions when it wasn’t answering questions, but what I noticed in most recent roles is that, clients are more satisfied if they get, like.
115 00:20:47.690 ⇒ 00:20:56.270 Advait Nandakumar Menon: Quick, delivery of, shorthand, things like, quick, spreadsheet of their,
116 00:20:56.410 ⇒ 00:21:05.470 Advait Nandakumar Menon: totals for the month, like, with respect to orders and sales and whatnot. And basically, trying to…
117 00:21:06.190 ⇒ 00:21:17.110 Advait Nandakumar Menon: give something quick and easy to whet their appetite, in that sense, and then building on top of it, and giving more well-rounded solutions, and going from there, so…
118 00:21:17.250 ⇒ 00:21:29.149 Advait Nandakumar Menon: that’s how I would approach a new project, like, really try to quickly understand the data structure quickly, what they’re trying to solve, and try to give them something tangible really quickly, and work on top of it.
119 00:21:29.550 ⇒ 00:21:32.040 Greg Stoutenburg: Yeah, yeah, okay, great.
120 00:21:32.310 ⇒ 00:21:37.949 Greg Stoutenburg: Alright, 5 minutes left. I should probably pick, like, maybe just one question.
121 00:21:37.950 ⇒ 00:21:38.550 Advait Nandakumar Menon: Yeah.
122 00:21:41.170 ⇒ 00:21:45.080 Greg Stoutenburg: Describe how you’ve, well, no, you already did one like that.
123 00:21:45.550 ⇒ 00:21:56.160 Greg Stoutenburg: Tell me about a template or a system you improved, or some kind of automation that you set up, so that something that was a manual task or project got faster and more repeatable.
124 00:21:56.160 ⇒ 00:21:56.790 Advait Nandakumar Menon: Right.
125 00:21:57.140 ⇒ 00:21:58.980 Advait Nandakumar Menon: So…
126 00:21:59.220 ⇒ 00:22:14.979 Advait Nandakumar Menon: In my most recent role, one example of that is, we used to get a lot of ad hoc analysis requests from clients, as you can imagine, like, they might be dumping CSVs on us, and CSV files on us, and asking to analyze it and give some insights about it. So.
127 00:22:15.620 ⇒ 00:22:33.600 Advait Nandakumar Menon: I observed that within our team, we were spending quite some time to basically, like, even before the analysis part, we need to clean it and do all the validation and all those pre-processing stuff on the data so that we can ready it up for analysis and analyze it. So I was…
128 00:22:33.600 ⇒ 00:22:52.720 Advait Nandakumar Menon: I prototyped an AI agent, so it was an experiment. Basically, I introduced an AI agent, which I built using Langchain and Azure OpenAI, so most of our clients were within the Microsoft ecosystem, so that’s why I stuck with Azure OpenAI and not to go out of the Microsoft ecosystem.
129 00:22:52.810 ⇒ 00:23:09.039 Advait Nandakumar Menon: So, we… I created an agent that basically reads the CSV files, which we can upload, and it basically, cleans the data, checks for any inconsistencies, like duplicates, or invalid, field types, or whatever, and really gives a…
130 00:23:09.040 ⇒ 00:23:21.279 Advait Nandakumar Menon: starter insight or visuals about the data itself. It’s not really meant to automate the analysis part because, as you might know, AI is nowhere close near to just analyze and give us the result.
131 00:23:21.320 ⇒ 00:23:37.369 Advait Nandakumar Menon: it’s slowly approaching there, but I wanted to automate the part where the pre-processing validation steps, where the amount of time we were spending is being taken into factor. So, that’s a process I automated, and it saved about
132 00:23:37.450 ⇒ 00:23:48.289 Advait Nandakumar Menon: I would say we were spending around 3 hours per analysis request. It cut down to 45 minutes just for the analysis, removing the validation and preprocessing step.
133 00:23:48.580 ⇒ 00:23:51.679 Greg Stoutenburg: Yeah, that’s fantastic. So you built the agent for them?
134 00:23:52.070 ⇒ 00:24:09.200 Advait Nandakumar Menon: Yeah, yeah. It’s not for a client, it’s an internal work we did, because we were dealing with all these clients, and like I said, most of them were within the Microsoft ecosystem, so I believed it would make sense to develop something like this. So, yeah, it’s a prototype that I came up with.
135 00:24:09.200 ⇒ 00:24:18.680 Greg Stoutenburg: Yeah, sure, no, that’s great that, you know, building custom internal tools is fantastic, yeah. Yeah, yeah. Yeah, yeah, cool. Okay, all right, well…
136 00:24:18.680 ⇒ 00:24:38.249 Greg Stoutenburg: Yeah, sort of at the cusp there. So, do you have any questions for me? Now, I’m just… I’m not… I’m not a manager or anything like that, you know, I’d be at the same, you know, I’m just another consultant at the company. But yeah, do you have any questions about, you know, if you have any questions about, like, the day-to-day or anything like that, you know, maybe I can answer something like that, if you have it.
137 00:24:39.030 ⇒ 00:24:54.630 Advait Nandakumar Menon: Yeah, so I was, just thinking, like, since you mentioned you’re also a consultant, so in your experience so far, what separates a good client engagement from a great one at Brainforge? Like, how does that, go for you?
138 00:24:54.630 ⇒ 00:25:19.619 Greg Stoutenburg: Yeah, yeah, yeah, good question. So, I have, I have, I guess, touched on, I guess, four different client engagements, and done, some significant work for, I guess, two and a half of them. And my experience has been, like, it really depends upon the relationship with the client. So, one of the things that’s important is just having that good relationship with the client. Things that can be really great are when you and the client are really aligned on what they want to see done.
139 00:25:19.970 ⇒ 00:25:39.399 Greg Stoutenburg: And, you know, you’re able to establish that trust with them, that you know what you’re doing, and you can provide value, and you’re, you know, you’re able to demonstrate that, things that can be challenging are when the client can be sort of wishy-washy on what their own needs are, and so then they’re, you know, if they’re sort of, like.
140 00:25:39.400 ⇒ 00:25:49.170 Greg Stoutenburg: If they’re living in the tyranny of the urgent, you know, every day is a new fire, that can be really challenging to demonstrate success in if you’re having to be too reactive.
141 00:25:50.260 ⇒ 00:26:02.530 Greg Stoutenburg: Yeah, so… so I… I mean, I think, yeah, that would probably be the… the main, you know, good relationship quality, and also the main bad relationship quality. It’s really important to…
142 00:26:02.550 ⇒ 00:26:12.629 Greg Stoutenburg: it’s really important to, like, have a direction that you set, that you’re taking the client in. You’re able to leverage that expertise and trust, and deliver some value for them.
143 00:26:12.930 ⇒ 00:26:18.379 Greg Stoutenburg: Because if you’re not, if you’re too reactive, even if it’s their fault that you’re reactive, because
144 00:26:20.200 ⇒ 00:26:29.289 Greg Stoutenburg: Even if it is their fault, sooner or later, they’re going to conclude, okay, you know, it’s been a quarter, or it’s been two quarters, and I’m not getting out of this what I thought I was going to get out of it.
145 00:26:29.290 ⇒ 00:26:29.720 Advait Nandakumar Menon: Yeah.
146 00:26:29.720 ⇒ 00:26:47.089 Greg Stoutenburg: Yeah, yeah, yeah. So, there’s a lot of change, it’s, you know, it’s fast-paced, there’s a lot of new challenges all the time, and, you know, but for some people, that’s really great. I mean, I can tell you, you know, I’ve only been here for about 6 weeks, but it’s… I’ve just learned a ton. I’ve learned a ton. It’s been really great. This is my first time…
147 00:26:47.090 ⇒ 00:26:47.480 Advait Nandakumar Menon: That’s awesome.
148 00:26:48.510 ⇒ 00:26:49.340 Advait Nandakumar Menon: Before that.
149 00:26:49.340 ⇒ 00:26:53.320 Greg Stoutenburg: It was always full-time in-house, yeah, as, like, as a product manager. Yeah.
150 00:26:53.810 ⇒ 00:26:54.230 Greg Stoutenburg: Yeah.
151 00:26:54.230 ⇒ 00:26:54.550 Advait Nandakumar Menon: And…
152 00:26:54.550 ⇒ 00:26:55.060 Greg Stoutenburg: Yeah, cool.
153 00:26:55.060 ⇒ 00:27:13.240 Advait Nandakumar Menon: No, you… you mentioning some of this stuff really clicks with me, because that’s something I’ve seen play out in my roles so far as well, so… Yeah. Yeah, I think the key is really to build a relationship, if it’s a new client, with quick wins, and something tangible, and something small, and take it from there, so… Yeah.
154 00:27:13.240 ⇒ 00:27:28.330 Greg Stoutenburg: Yep, yep, yep, so. So, yeah, cool. All right, well, yeah, thanks for your time. I’ll, I’ll give my feedback to the team, and then whoever’s been, whoever’s been in touch with you so far will, will reach out for whatever’s next. Yep, yep, sounds good. Great to talk. Nice to meet you.
155 00:27:28.330 ⇒ 00:27:31.260 Advait Nandakumar Menon: It was great talking to you. Have a good one. Bye-bye. Take care.