Meeting Title: Brainforge <> Contextual: Bi-Weekly Catchup Date: 2026-01-28 Meeting participants: Holly Condos, Uttam Kumaran, Luke Scorziell, Hannah Wang, Mike Klaczynski, Abhishek Varma
WEBVTT
1 00:00:13.830 ⇒ 00:00:15.050 Uttam Kumaran: Hello!
2 00:00:16.120 ⇒ 00:00:16.910 Holly Condos: Hey!
3 00:00:17.350 ⇒ 00:00:18.580 Uttam Kumaran: Brave background!
4 00:00:19.940 ⇒ 00:00:20.999 Holly Condos: Does it look.
5 00:00:21.000 ⇒ 00:00:23.000 Uttam Kumaran: It looks great!
6 00:00:23.380 ⇒ 00:00:26.060 Holly Condos: Because the brain force looks backwards to me.
7 00:00:26.330 ⇒ 00:00:28.320 Uttam Kumaran: No, it looks alright on our side.
8 00:00:28.320 ⇒ 00:00:28.690 Holly Condos: Okay.
9 00:00:28.690 ⇒ 00:00:29.320 Luke Scorziell: Yeah, to…
10 00:00:30.420 ⇒ 00:00:30.960 Holly Condos: Awesome.
11 00:00:30.960 ⇒ 00:00:33.979 Luke Scorziell: It’s inspiring me to get a BrainForge background.
12 00:00:35.700 ⇒ 00:00:37.450 Holly Condos: Mmm, yeah, you should.
13 00:00:43.600 ⇒ 00:00:46.419 Holly Condos: Did you survive the snow, Wu-Tong?
14 00:00:47.030 ⇒ 00:00:56.050 Uttam Kumaran: I survived. I wish I could show you, there’s, like, not much left, but my whole, like, yard was… it wasn’t snow, it was, like, ice here. Horrible.
15 00:00:56.050 ⇒ 00:00:59.719 Holly Condos: show you my dad. Let me see if I can do this.
16 00:01:01.850 ⇒ 00:01:04.819 Holly Condos: Yeah, hold on, maybe I can do this.
17 00:01:07.120 ⇒ 00:01:08.449 Holly Condos: You see that?
18 00:01:09.680 ⇒ 00:01:12.740 Uttam Kumaran: No, you’re… it says… shows off-camera.
19 00:01:14.180 ⇒ 00:01:15.599 Holly Condos: Okay, hold on, nice.
20 00:01:16.540 ⇒ 00:01:19.130 Holly Condos: How is it so clear, there’s so much snow.
21 00:01:25.500 ⇒ 00:01:29.100 Holly Condos: It’s still… Silly me.
22 00:01:29.680 ⇒ 00:01:30.639 Holly Condos: Oh well.
23 00:01:31.050 ⇒ 00:01:33.070 Holly Condos: I’ll see if I can figure it out during the call.
24 00:01:33.070 ⇒ 00:01:34.790 Uttam Kumaran: Our picture is fine, yeah.
25 00:01:35.190 ⇒ 00:01:36.769 Luke Scorziell: Oh, yeah. Yeah, sure.
26 00:01:37.700 ⇒ 00:01:40.140 Holly Condos: Yeah, we had about, I don’t know, 15 inches?
27 00:01:42.110 ⇒ 00:01:45.469 Uttam Kumaran: Wow. 15 inches is crazy.
28 00:01:45.830 ⇒ 00:01:46.350 Holly Condos: Doing well.
29 00:01:49.000 ⇒ 00:01:49.980 Uttam Kumaran: Hey, Mike.
30 00:01:53.080 ⇒ 00:01:53.820 Holly Condos: Hi, Mike.
31 00:01:54.240 ⇒ 00:01:55.600 Mike Klaczynski: Hi folks, how’s it going?
32 00:01:55.600 ⇒ 00:01:56.560 Uttam Kumaran: Hey, good.
33 00:01:56.770 ⇒ 00:01:57.810 Luke Scorziell: Matt, how are you?
34 00:01:58.360 ⇒ 00:01:59.350 Mike Klaczynski: It’s good.
35 00:02:00.330 ⇒ 00:02:02.780 Uttam Kumaran: I don’t know, Luke, if you met Mike. Mike’s,
36 00:02:02.960 ⇒ 00:02:15.100 Uttam Kumaran: on the contextual team has been sort of, helping us steer, and I think today we’re gonna talk a little bit about… Luke’s sort of owning a lot of our campaign calendar on the marketing side. He’s also…
37 00:02:15.250 ⇒ 00:02:16.370 Uttam Kumaran: literally…
38 00:02:16.570 ⇒ 00:02:34.919 Uttam Kumaran: today, as we’re working on putting together the insurance campaign. I think that’ll kind of be the highlight of today. Maybe I can just start by, like, we wanted to plug your agent composer, I think I got… if I got it right, and, like, repost it and blast it out. I think, Luke, you’re planning on, like, next week.
39 00:02:35.020 ⇒ 00:02:35.650 Uttam Kumaran: Right?
40 00:02:35.650 ⇒ 00:02:37.319 Luke Scorziell: Yeah, we can do that next week.
41 00:02:37.470 ⇒ 00:02:45.440 Uttam Kumaran: So, just wanted to flag that to you, that I think we’ll try to get something out from our stuff next week. I did get a little bit of sneak peek last week.
42 00:02:45.630 ⇒ 00:02:49.590 Uttam Kumaran: And I, I, like, love it, I think it’s great, I think it’s, like, totally makes…
43 00:02:49.810 ⇒ 00:02:51.150 Uttam Kumaran: A ton of sense.
44 00:02:52.520 ⇒ 00:02:59.409 Uttam Kumaran: And I think we’re gonna highlight a couple of the key pieces that, in particular, I think we see some immediate, like, benefit from.
45 00:02:59.600 ⇒ 00:03:01.539 Uttam Kumaran: I’m gonna see if… if we can…
46 00:03:01.810 ⇒ 00:03:05.019 Uttam Kumaran: Start to build some of the insurance thing we did already.
47 00:03:05.330 ⇒ 00:03:07.830 Uttam Kumaran: Onto that, which would be really nice, but…
48 00:03:08.000 ⇒ 00:03:16.310 Uttam Kumaran: gave me… yell at me and say, like, we don’t have time, but we’ll see. So yeah, I just want to flag that, like, I think we’ll get to pump that next week.
49 00:03:16.620 ⇒ 00:03:31.350 Mike Klaczynski: Awesome, we appreciate it, thank you. And then, if you need any help with it, we know it’s still early days on it, so we’re happy to help rebuild it in Agent Composer, the insurance demo.
50 00:03:31.920 ⇒ 00:03:37.779 Mike Klaczynski: you know, we could sit down side by side and kind of work through it together and figure out what it would look like. So…
51 00:03:37.960 ⇒ 00:03:40.719 Mike Klaczynski: Don’t feel like you guys have to figure it out by yourselves.
52 00:03:42.480 ⇒ 00:03:43.000 Uttam Kumaran: Cool.
53 00:03:44.170 ⇒ 00:03:47.959 Uttam Kumaran: So yeah, Luke, I guess, like, we can ask a channel if we have any questions, but we’re gonna…
54 00:03:47.970 ⇒ 00:03:52.510 Luke Scorziell: I think we’re just gonna piggyback off the post, and I’ll probably add some notes to it, so…
55 00:03:54.020 ⇒ 00:03:57.330 Luke Scorziell: Yeah, yeah, and then I think, like, as we… kind of…
56 00:03:58.010 ⇒ 00:04:07.999 Luke Scorziell: I mean, we’re gonna, like, experiment a little bit with different insurance segments, so also, if you have anyone that you would be open to connecting us with that we could just, like… or if you have, yeah, just anyone in your network…
57 00:04:08.000 ⇒ 00:04:13.520 Uttam Kumaran: Do you want to even share, Luke, maybe we can just go into that, like, do we want to even talk about, like, what our flow is gonna be?
58 00:04:13.650 ⇒ 00:04:16.260 Uttam Kumaran: And yeah, fall, summer.
59 00:04:16.260 ⇒ 00:04:17.200 Luke Scorziell: a sales flow.
60 00:04:17.589 ⇒ 00:04:22.599 Uttam Kumaran: Yeah, or, yeah, or just… we could just walk through, Mike, like, how we’re kind of gonna do it, and maybe…
61 00:04:22.829 ⇒ 00:04:25.369 Uttam Kumaran: We can think about if there’s any ways we want to, like.
62 00:04:26.319 ⇒ 00:04:29.479 Uttam Kumaran: piggyback off each other, or yeah, I mean, especially if you guys are…
63 00:04:30.289 ⇒ 00:04:36.939 Uttam Kumaran: have already talked to any folks in insurance, like, we’re basically gonna put together, like, a 3- or 4-part campaign and, like, start to blast it out.
64 00:04:37.930 ⇒ 00:04:40.370 Mike Klaczynski: Yeah, Farzad was working with…
65 00:04:40.540 ⇒ 00:04:46.179 Mike Klaczynski: an insurance company down in San Diego, and I’d have to dig through the notes to see who that was.
66 00:04:46.740 ⇒ 00:04:52.640 Mike Klaczynski: But that was, yeah, like, 3-4 months ago, so lots changed since then, but maybe they’d be open to revisiting.
67 00:04:53.010 ⇒ 00:04:57.429 Uttam Kumaran: And we can even send them, like… we usually put together, like, a campaign brief, right? Like…
68 00:04:57.540 ⇒ 00:05:05.670 Uttam Kumaran: we’re gonna… we’re gonna put together these sort of video demos around the contextual demos that Gabe on our team put together, and then…
69 00:05:05.780 ⇒ 00:05:09.420 Uttam Kumaran: Part of it is, like, just normal, like, sales hook stuff, and then…
70 00:05:09.580 ⇒ 00:05:15.459 Uttam Kumaran: try to drive them towards, like, a lot of the KPIs that we talked about, which is, like,
71 00:05:15.580 ⇒ 00:05:19.979 Uttam Kumaran: Lead processing, lead volume processing, you know, things like that.
72 00:05:21.850 ⇒ 00:05:22.170 Mike Klaczynski: Yeah.
73 00:05:22.770 ⇒ 00:05:27.150 Mike Klaczynski: Totally makes sense. When you have those available to share, let us, let us know.
74 00:05:27.400 ⇒ 00:05:27.930 Uttam Kumaran: Cool.
75 00:05:28.280 ⇒ 00:05:32.109 Luke Scorziell: Yeah, I mean, like, right now we’ve got… basically, I mean, if we’re just…
76 00:05:32.560 ⇒ 00:05:43.100 Luke Scorziell: like, it’s, like, cold on LinkedIn, so, you know, you never know how that’s gonna go. But, like, a connection request, and then we were thinking of doing…
77 00:05:43.500 ⇒ 00:05:54.120 Luke Scorziell: I mean, I can show you this, it’s just all on Slack right now, though, so it’s not, like, super, formalized yet, but yeah, like, a connection request to try to get, like.
78 00:05:54.550 ⇒ 00:05:58.950 Luke Scorziell: I think it’s in the… is it any insurance that we can do, or is it mainly commercial?
79 00:05:59.580 ⇒ 00:06:04.029 Uttam Kumaran: I would probably… yeah, I guess it… it doesn’t really matter. What you’re gonna find is that…
80 00:06:04.390 ⇒ 00:06:08.740 Uttam Kumaran: The average commercial brokerage is just gonna be much bigger, and, like.
81 00:06:08.740 ⇒ 00:06:09.110 Luke Scorziell: Damn.
82 00:06:09.110 ⇒ 00:06:11.380 Uttam Kumaran: The lead volumes on personal insurance.
83 00:06:11.910 ⇒ 00:06:19.730 Uttam Kumaran: it’s just… it’s just a different type of business. So, it may be something that we’ll call Ian and sort of maybe get his thoughts on…
84 00:06:20.600 ⇒ 00:06:38.909 Mike Klaczynski: Yeah, the commercial stuff seems to be way more complicated, right? He said it could take him, like, a day or two to fill out all the forms, just because there’s so much stuff to ensure. Well, personal policy, the end client usually fills it out themselves. Like, I fill out my stuff, hand it over, and then I think they forward it on to the different underwriters.
85 00:06:39.530 ⇒ 00:06:51.489 Luke Scorziell: Yeah, okay. Well, so, I mean, we can just experiment then, and it sounds like it’s pretty customizable, so if we have to, like, build a different solution when they’re paying us to do that, then we’ll figure it out.
86 00:06:51.490 ⇒ 00:06:52.990 Mike Klaczynski: There’s a ton of overlap, right?
87 00:06:52.990 ⇒ 00:06:53.740 Uttam Kumaran: Yeah.
88 00:06:53.960 ⇒ 00:06:54.910 Luke Scorziell: Yeah, yeah.
89 00:06:54.910 ⇒ 00:06:55.580 Mike Klaczynski: We’re both…
90 00:06:56.430 ⇒ 00:07:13.769 Luke Scorziell: So, yeah, I guess it’s that, and then we, gabe did a loom, just kind of running through it, so we’ll either use that on, or maybe re-record, and then just figure, like, following up with that, either on LinkedIn, or if we can identify, like, emails of people,
91 00:07:14.500 ⇒ 00:07:17.260 Luke Scorziell: So, yeah, I mean, that’s kind of the… the…
92 00:07:17.410 ⇒ 00:07:22.540 Luke Scorziell: play that we’re thinking specific to this one. I’d be curious to know, too, like, If you…
93 00:07:23.140 ⇒ 00:07:28.779 Luke Scorziell: Yeah, I guess other, like, ideas or priorities that, like, you guys have at contextual? Because I think, like, we’re pretty…
94 00:07:29.050 ⇒ 00:07:36.980 Luke Scorziell: Or just momentum that you’re seeing, because, like, my thought with this is we’ll launch, probably…
95 00:07:37.310 ⇒ 00:07:39.959 Luke Scorziell: Or hopefully get a few messages, even today.
96 00:07:40.080 ⇒ 00:07:45.950 Luke Scorziell: And then kind of iterate through a few different, like, verticals, and see…
97 00:07:46.110 ⇒ 00:07:51.110 Luke Scorziell: Maybe who… so obviously we only have so many demos, but I’d be curious just what’s been on your mind lately.
98 00:07:51.760 ⇒ 00:07:56.550 Mike Klaczynski: Yeah, so we are doubling down specifically on the engineering segment.
99 00:07:57.530 ⇒ 00:08:09.400 Mike Klaczynski: Top-line message, semiconductor manufacturing, I mean, that’s… the campaign we launched yesterday talks about, like, rocket science, so…
100 00:08:09.400 ⇒ 00:08:11.760 Luke Scorziell: Yeah, I saw it.
101 00:08:11.760 ⇒ 00:08:18.470 Mike Klaczynski: The physical, mechanical, electrical engineering is really the area we’re gonna focus on, but everything else is fair game.
102 00:08:18.470 ⇒ 00:08:31.860 Mike Klaczynski: It’s just, we’ve traditionally been very horizontal, and it’s been tough to break through that, because we go meet with an insurance company, and they’re like, well, what do you do? And we say, well, we can do a lot of things. And they’re like, can you be specific? And we’re like, well, tell us what you guys want.
103 00:08:32.390 ⇒ 00:08:40.880 Mike Klaczynski: And it makes the conversation difficult. So that’s why we’re looking at partners such as yourself to help build out those solutions and applications. We’ll include that into our…
104 00:08:41.250 ⇒ 00:08:52.590 Mike Klaczynski: quiver of available options, and then if a customer says, hey, I’m really interested in this, then we will bring them to you and say, we have a partner that built this, and this is the partner through which you can access this solution. And so the conversations.
105 00:08:52.590 ⇒ 00:08:53.130 Luke Scorziell: Oh, cool.
106 00:08:53.130 ⇒ 00:08:54.819 Mike Klaczynski: With you guys, what’s legal.
107 00:08:54.820 ⇒ 00:08:57.050 Luke Scorziell: Real estate, insurance…
108 00:08:57.160 ⇒ 00:09:02.159 Mike Klaczynski: I think those were the three, and there might have been a fourth one. I’d have to go look through the original document, but I think those were the three.
109 00:09:02.160 ⇒ 00:09:04.490 Holly Condos: Those are the top 3 that we started with.
110 00:09:04.490 ⇒ 00:09:04.850 Uttam Kumaran: Yeah.
111 00:09:04.850 ⇒ 00:09:11.659 Holly Condos: And we do have a legal notion, also, that, you know, we can pick up
112 00:09:12.130 ⇒ 00:09:15.109 Holly Condos: Oh, cool. To brainstorm on once we’re done, or at least
113 00:09:15.400 ⇒ 00:09:20.229 Holly Condos: get the insurance going. So, we do have that, in the background.
114 00:09:21.610 ⇒ 00:09:22.020 Luke Scorziell: Okay.
115 00:09:22.060 ⇒ 00:09:24.560 Uttam Kumaran: Yeah, and I actually think overall, like,
116 00:09:25.500 ⇒ 00:09:44.349 Uttam Kumaran: what we’re seeing in the market is a lot of people have, like, tried to buy a lot of off-the-shelf tools that are just not bespoke to, like, what they’re doing. Yep. And, especially in these industries that are, like, email-based, they’ve all tried, like, some email automation tool, but then again, it only has access to what’s in your email, it doesn’t have, like, your CRM, it doesn’t… so…
117 00:09:44.600 ⇒ 00:09:50.780 Uttam Kumaran: We’re seeing some companies with that, like, in particular, Luke, we’re talking to this company called Plan Medicare, they’re a big Medicare brokerage.
118 00:09:50.780 ⇒ 00:09:52.010 Luke Scorziell: Yeah.
119 00:09:52.010 ⇒ 00:09:56.560 Uttam Kumaran: you can listen to that call, it’s literally this, he’s like, look, I,
120 00:09:57.140 ⇒ 00:10:10.250 Uttam Kumaran: I’m… we’re getting… he’s getting, like, 400 emails a day, they need a bunch of email automation, they need to hook into some API. Unfortunately, like, they’re not as, like, doc… it’s actually more just, like, simple API calls, but…
121 00:10:10.250 ⇒ 00:10:10.920 Luke Scorziell: Huh.
122 00:10:10.920 ⇒ 00:10:20.249 Uttam Kumaran: And, like, again, it’s, like, pure Medicare sales, so it’s just, like, getting people into a funnel, and then getting them to go do things, and you’re the broker. But, like, there are some industries where, like.
123 00:10:20.250 ⇒ 00:10:33.969 Uttam Kumaran: they’re trying to start at the automation of the email level, hook up other sources, and then, naturally, the industries where there’s a lot of PDF and document transfer is exactly, like, where Contextual comes in, where it’s, like, they’re basically the source of truth.
124 00:10:33.970 ⇒ 00:10:44.490 Uttam Kumaran: for AI on top of any document… any document intelligence. And ideally, if that exists, then we can orchestrate the whole thing just in contextual. Versus for folks that are just doing, like.
125 00:10:45.260 ⇒ 00:10:47.260 Uttam Kumaran: we’re just hooking into their Gmail.
126 00:10:47.550 ⇒ 00:10:53.170 Uttam Kumaran: and, like, calling a CRM API, it’s like… It’s a lot lighter. Yeah.
127 00:10:53.170 ⇒ 00:10:53.550 Luke Scorziell: It doesn’t real.
128 00:10:53.550 ⇒ 00:11:00.830 Uttam Kumaran: We’re gonna see, like, for them, we’re gonna go as deep as they can, but, like, the document use case, like, I think maybe, Luke, even in our messaging.
129 00:11:01.250 ⇒ 00:11:04.400 Uttam Kumaran: Really hammering the fact that, like, they’re trying to connect
130 00:11:04.540 ⇒ 00:11:13.099 Uttam Kumaran: email, to their CRM, to, like, their Google Drive, or, like, I don’t know, we think we’d need to hammer the document piece really hard.
131 00:11:14.650 ⇒ 00:11:22.840 Uttam Kumaran: like, even say, like, PDFs, Excel files, I think is really, like, the key, because that’s what the demo is gonna show.
132 00:11:24.010 ⇒ 00:11:32.189 Uttam Kumaran: And that’s what, like… and the other way to actually, like, create an enemy in this, that’s what I’m seeing, is, like, people are saying, like, this is way better than
133 00:11:32.370 ⇒ 00:11:37.180 Uttam Kumaran: taking a bunch of documents, throwing it at ChatGPT, basically, which is what a lot of these folks are doing.
134 00:11:37.930 ⇒ 00:11:43.649 Uttam Kumaran: Depending on who we go to, though, like, if you’re… if you find that one of these insurance companies has, like, a head of IT,
135 00:11:43.760 ⇒ 00:11:55.149 Uttam Kumaran: or, like, head of security on, like, IT side, they’re more… their concern is going to be, like, people are throwing commercial documents into personal chat GPTs, which is what everybody is doing, you know, at all their companies.
136 00:11:55.180 ⇒ 00:11:56.990 Mike Klaczynski: leakage. Like, at our company, the way I.
137 00:11:57.710 ⇒ 00:12:12.940 Uttam Kumaran: I tried to challenge that, I just bought everyone ChatGPT, so then it makes it much better for you to just use ours, at least. But, like, most people, their companies are not buying it for them, or they’re slow, and so they just are, like, ripping it on their own, which is a huge, like, risk.
138 00:12:12.940 ⇒ 00:12:20.420 Mike Klaczynski: Yep. Yeah, on that email side, even though it’s text up front, I think there’s still good value, because
139 00:12:20.840 ⇒ 00:12:29.040 Mike Klaczynski: every single one of those email conversations relates to something, either SOPs, policy documents, procedures, and so…
140 00:12:29.040 ⇒ 00:12:44.130 Mike Klaczynski: you know, you might get an email asking about certain questions, but now it has to go to the docs to look up how to handle those policies or procedures, or even say, hey, previously remedied situations followed this particular flow. So that whole decision tree of where the answer comes from.
141 00:12:44.130 ⇒ 00:12:49.409 Mike Klaczynski: Is usually where it gets more complex, right? Because yes, you can have ChatGPT or something just.
142 00:12:49.410 ⇒ 00:12:49.800 Holly Condos: for the email.
143 00:12:49.800 ⇒ 00:12:54.470 Mike Klaczynski: But if it actually looks up policies and requirements that the company.
144 00:12:54.470 ⇒ 00:12:54.930 Luke Scorziell: Yeah.
145 00:12:54.930 ⇒ 00:13:02.680 Mike Klaczynski: insurance company your healthcare provider has themselves, that’s not publicly available, so it will not be able to get answered unless it has access to those documents.
146 00:13:03.410 ⇒ 00:13:13.489 Luke Scorziell: Yeah, and I’m still learning, but this is the contextual layer, right? Or just, like, this is the body proposition, I assume, is contextual. But yeah, I mean, we’ve been talking about that a lot lately on…
147 00:13:13.590 ⇒ 00:13:22.539 Luke Scorziell: like, just… we’re doing, like, commenting on different, like, AI thought leaders and whatnot, and… and obviously within… within,
148 00:13:22.540 ⇒ 00:13:33.610 Luke Scorziell: Brainforge, too, we’re using Cursor, and we have, like, the Brainforge vault, and so if you want to ask it to do a LinkedIn campaign, it’ll just go and reference all of that stuff, which we’ve…
149 00:13:33.720 ⇒ 00:13:37.669 Luke Scorziell: But, you know, super helpful, versus, like, going in every time to chat GPT.
150 00:13:37.770 ⇒ 00:13:40.790 Luke Scorziell: I’m like, hey, can you make, you know, this whatever?
151 00:13:41.130 ⇒ 00:13:41.740 Mike Klaczynski: Exciting.
152 00:13:41.740 ⇒ 00:13:43.119 Luke Scorziell: and you gotta feed it.
153 00:13:43.120 ⇒ 00:13:47.319 Mike Klaczynski: two-paragraph prompt that explains everything. Like, it should already have that context for you.
154 00:13:47.320 ⇒ 00:13:48.100 Uttam Kumaran: Yeah.
155 00:13:48.360 ⇒ 00:13:51.420 Uttam Kumaran: Yeah, and the thing with contextual is, like, again, they’re…
156 00:13:51.630 ⇒ 00:13:56.000 Uttam Kumaran: their document intelligence is way better than what is out of the box in Cursor.
157 00:13:56.350 ⇒ 00:13:57.150 Uttam Kumaran: Right, like, cause.
158 00:13:57.750 ⇒ 00:14:03.780 Uttam Kumaran: like, their stuff is, like, like, way beyond. So, for… for us, in fact, we’re…
159 00:14:03.990 ⇒ 00:14:15.030 Uttam Kumaran: our use case for cursor is actually more textiles and code, which is great, because we’re not a big, like, we don’t, like, we’re not creating a ton of docs, we’re not creating a ton of PDFs.
160 00:14:15.150 ⇒ 00:14:21.499 Uttam Kumaran: Like, and doing a lot of document transfer. So, like, finding things within those.
161 00:14:21.720 ⇒ 00:14:31.329 Uttam Kumaran: running that through Cursor can be… it’s just definitely not going to be as accurate. And building the workflow itself to, like, go do things, create structures is a lot harder.
162 00:14:32.700 ⇒ 00:14:35.569 Uttam Kumaran: So that’s where, like, Contextual is, like, really shines.
163 00:14:36.830 ⇒ 00:14:37.480 Luke Scorziell: Yeah.
164 00:14:43.770 ⇒ 00:14:51.059 Mike Klaczynski: Yeah, so share what you have, and then we’re happy to provide feedback and additional guidance and whatnot, and we can… we can work through that, but…
165 00:14:51.490 ⇒ 00:14:56.169 Uttam Kumaran: Yeah, I think, Luke, we figure out, like, what… we figure out if we see any signal in insurance.
166 00:14:56.400 ⇒ 00:15:03.869 Uttam Kumaran: And we do think we’ve sort of tried to push also on legal as well. And so, like, I think my ask on your side is, like.
167 00:15:04.330 ⇒ 00:15:18.190 Uttam Kumaran: working with Gabe and being like, okay, like, how do we… for this… I… part of this… this exercise was like, okay, how long is it gonna take us to build the demos and then build a campaign? But the next one, I wanted to just see, like, okay, how do we cut that in half, basically? Like, now that Gabe is really
168 00:15:18.580 ⇒ 00:15:26.479 Uttam Kumaran: the platform, there’s now, like, these new UI-based builders, like, maybe it’s actually much quicker to test the next sector.
169 00:15:26.640 ⇒ 00:15:27.030 Uttam Kumaran: Yep.
170 00:15:27.350 ⇒ 00:15:29.300 Uttam Kumaran: What’s the messaging, and they iterate on the messaging.
171 00:15:29.780 ⇒ 00:15:38.679 Luke Scorziell: And when we’re thinking about real estate, is that… are we… is that, like, all-encompassing, like, mortgage brokers, loan officers, or is that kind of.
172 00:15:38.680 ⇒ 00:15:44.330 Uttam Kumaran: I think similarly, where we’ve had experience is working with commercial real estate.
173 00:15:44.470 ⇒ 00:15:50.970 Uttam Kumaran: Like, brokers. And brokers and, like, and property owners. Yeah.
174 00:15:50.970 ⇒ 00:15:52.350 Mike Klaczynski: Due diligence on properties.
175 00:15:52.350 ⇒ 00:15:53.090 Uttam Kumaran: Yeah.
176 00:15:53.090 ⇒ 00:15:56.360 Mike Klaczynski: There’s a ton of lease contract negotiation.
177 00:15:56.360 ⇒ 00:16:05.669 Uttam Kumaran: diligence, Risk exposure across a portfolio of leases. Structured entry of lease information into a lease management platform.
178 00:16:09.290 ⇒ 00:16:12.880 Mike Klaczynski: Routine tasks that are fairly complex.
179 00:16:13.450 ⇒ 00:16:16.669 Uttam Kumaran: Yeah, and it’s easy to… and people are fat-fringering it all day.
180 00:16:18.020 ⇒ 00:16:18.680 Luke Scorziell: Yeah.
181 00:16:19.300 ⇒ 00:16:30.780 Mike Klaczynski: So, on that demo and template side of things, Abhishek’s on, but I think it would be good for us to run you guys through all the new templates that our team’s produced.
182 00:16:31.270 ⇒ 00:16:34.760 Mike Klaczynski: I think there’s… Abhishek, how many are there? Are there 10 or a dozen?
183 00:16:35.280 ⇒ 00:16:41.290 Abhishek Varma: the ones that are live on the platform, or the canonical examples?
184 00:16:41.890 ⇒ 00:16:43.200 Mike Klaczynski: Whatever ones that you’re gonna…
185 00:16:43.200 ⇒ 00:16:44.760 Abhishek Varma: Like, we have, like, 6.
186 00:16:44.760 ⇒ 00:17:03.640 Mike Klaczynski: We have 6, okay. Yeah, so at some point soon, maybe on the next call, Abhishek, you could run Gabe and Utam and Luke through all of those and say, this is what we’ve built, this is how it looks like, and then you guys can show us what you’ve built, and then we can brainstorm some additional ones you guys could create, either as derivatives or as net new ones.
187 00:17:03.640 ⇒ 00:17:04.210 Uttam Kumaran: Yeah.
188 00:17:05.849 ⇒ 00:17:12.789 Abhishek Varma: Yeah, and if you can communicate, problems that, you know, you’ve noticed to your customers,
189 00:17:13.119 ⇒ 00:17:31.039 Abhishek Varma: wanting solved. So beyond document intelligence, like, you know, features such as they need some calculations done, they need PowerPoints generated, they need certain, external data sources queried, whatever you can imagine that if it can run in a computer, we can do it now.
190 00:17:31.520 ⇒ 00:17:32.060 Uttam Kumaran: Yeah.
191 00:17:35.920 ⇒ 00:17:39.880 Abhishek Varma: So yeah, just let us know, and I can prepare, like, a… Demo.
192 00:17:41.190 ⇒ 00:17:41.740 Luke Scorziell: Yeah.
193 00:17:41.960 ⇒ 00:17:46.239 Luke Scorziell: So then, I mean, maybe then with us, it’s just, like, lining up
194 00:17:47.280 ⇒ 00:17:53.489 Luke Scorziell: Like, a vertical all the time, and then just kind of sifting through, and then we can, like, repeat and just go back through that.
195 00:17:53.490 ⇒ 00:17:59.020 Uttam Kumaran: Yeah, and like, again, the faster you can throw me into calls, the faster I’ll learn, sort of, like, okay, what do they need to see?
196 00:17:59.570 ⇒ 00:18:01.169 Uttam Kumaran: In terms of a demo?
197 00:18:01.680 ⇒ 00:18:07.740 Uttam Kumaran: And like, how do we… if that’s what they want to see at all, you know, so… yeah.
198 00:18:08.750 ⇒ 00:18:09.320 Luke Scorziell: Yeah.
199 00:18:10.700 ⇒ 00:18:13.389 Luke Scorziell: Okay, that sounds great. And then…
200 00:18:14.210 ⇒ 00:18:17.750 Luke Scorziell: Well, did we want to flesh out anything else on… on this kind of subject?
201 00:18:19.330 ⇒ 00:18:25.510 Uttam Kumaran: I think that’s it, and then, yeah, I think, Mike, as we develop more demos, I know you mentioned that you’re thinking about coming up with some type of agent marketplace.
202 00:18:25.510 ⇒ 00:18:25.859 Mike Klaczynski: Something like that.
203 00:18:25.860 ⇒ 00:18:28.120 Uttam Kumaran: We can… we can push that back to you.
204 00:18:28.380 ⇒ 00:18:29.740 Uttam Kumaran: You know.
205 00:18:30.160 ⇒ 00:18:46.079 Mike Klaczynski: So, in the meantime, what we’ll do is we’ll take the demos that you guys have put together, in one-pagers, some of the stuff you guys have shared in the last three months, but just get updated ones of those, and then I’ll put that in front of my sellers, and we’ll get that into our primary sales deck.
206 00:18:46.380 ⇒ 00:18:54.969 Mike Klaczynski: So that as we’re presenting use cases, you know, your slides will pop up and say, hey, here’s some additional use cases we’re able to solve for through Brainforge.
207 00:18:55.460 ⇒ 00:18:56.050 Uttam Kumaran: Okay.
208 00:18:56.400 ⇒ 00:18:58.540 Mike Klaczynski: So you guys will become just part of our standard deck.
209 00:18:58.540 ⇒ 00:18:59.100 Uttam Kumaran: Cool.
210 00:18:59.260 ⇒ 00:18:59.790 Holly Condos: That’s great.
211 00:19:01.480 ⇒ 00:19:08.080 Luke Scorziell: And do you have a, like, design… do you have a deck already that we can kind of look at just to see where that would fit in, or…
212 00:19:08.210 ⇒ 00:19:09.699 Luke Scorziell: Maybe you could send that over to us.
213 00:19:09.700 ⇒ 00:19:15.800 Mike Klaczynski: Yeah, we do have a brand new updated one, because all the messaging changed for Agent Composer. Let me…
214 00:19:17.030 ⇒ 00:19:20.030 Mike Klaczynski: Let me find what we have, and I’ll send that to you guys.
215 00:19:20.190 ⇒ 00:19:23.569 Luke Scorziell: So we can design it to kind of reflect your brand.
216 00:19:24.110 ⇒ 00:19:24.580 Mike Klaczynski: Yep.
217 00:19:24.580 ⇒ 00:19:25.560 Luke Scorziell: Also, too.
218 00:19:25.660 ⇒ 00:19:27.470 Luke Scorziell: I’m a little bit co-branded.
219 00:19:32.560 ⇒ 00:19:40.970 Luke Scorziell: And then, one of the other things, just, I know Hannah and I were kind of talking about is events. I’d just be curious, like, if there’s anything that you guys are…
220 00:19:41.710 ⇒ 00:19:47.059 Luke Scorziell: Like, have on your radar that you’re thinking for the next, like, quarter, next quarter after that?
221 00:19:48.120 ⇒ 00:19:52.730 Mike Klaczynski: Yeah, great question. John on our end would know better, but…
222 00:19:53.060 ⇒ 00:19:54.839 Mike Klaczynski: Off the top of my mind.
223 00:19:55.350 ⇒ 00:20:03.739 Mike Klaczynski: We are going to be at NVIDIA GTC in San Jose in March. We’re going to be at Google Next in April.
224 00:20:04.130 ⇒ 00:20:07.670 Mike Klaczynski: For GTC, we have a booth, but for Google Next, we do not.
225 00:20:07.790 ⇒ 00:20:12.959 Mike Klaczynski: And then I think we’re gonna pick up a few Microsoft events, I think Ignite, and
226 00:20:13.980 ⇒ 00:20:19.739 Mike Klaczynski: I forget what the other one’s called, but we’re gonna be launching a bigger partnership with Microsoft, they’re really leaning in with us.
227 00:20:20.610 ⇒ 00:20:21.790 Luke Scorziell: Cool, that’s awesome.
228 00:20:23.370 ⇒ 00:20:29.439 Uttam Kumaran: In one particular area, or just, like, procuring through, like, their marketplace, or, like, what’s the…
229 00:20:29.720 ⇒ 00:20:38.920 Mike Klaczynski: Yeah, they invited us to this AI startup incubator, and there’s only 8 startups in it. They gave us a ton of money, like a million bucks.
230 00:20:39.150 ⇒ 00:20:55.810 Mike Klaczynski: And access to really good technical resources, so we’re gonna be… we’re gonna become available on Azure rather soon, and then deepen our integrations into AI Foundry and, like, SharePoint, and they’ve got a lot of different,
231 00:20:55.990 ⇒ 00:21:06.309 Mike Klaczynski: I guess they have two. They have M365, which is, like, their co-pilot thing, so there’s apps we can build for that, and then the other one is AI Foundry, and then they have apps for that as well. So we’ll be…
232 00:21:06.520 ⇒ 00:21:09.350 Mike Klaczynski: Getting, launched on those at some point, too.
233 00:21:09.690 ⇒ 00:21:11.779 Holly Condos: That’s great. Congrats on that.
234 00:21:11.930 ⇒ 00:21:19.209 Mike Klaczynski: I think the SharePoint integration we have built is pretty solid, but we just haven’t done much go-to-market around that, so that’s what I want to showcase more.
235 00:21:19.780 ⇒ 00:21:20.150 Holly Condos: Gotcha.
236 00:21:20.150 ⇒ 00:21:27.730 Uttam Kumaran: Yeah, I mean, even that, like, Luke, we have a… this is our CS client, it’s just they’re… they’re just dealing with so much that I’m… I just, like…
237 00:21:28.230 ⇒ 00:21:34.010 Uttam Kumaran: figuring out where we can slot in, sort of, like, a SharePoint Style document intelligence, but…
238 00:21:34.330 ⇒ 00:21:48.060 Uttam Kumaran: they’re, like, a perfect customer for that. So there may be something, Luke, where, because they have a tight coupling with Microsoft, what you can actually now do is sort of… it’s actually a little bit easier to just go find the big Microsoft shops.
239 00:21:48.760 ⇒ 00:21:53.220 Uttam Kumaran: And position, sort of, around…
240 00:21:53.450 ⇒ 00:21:56.569 Uttam Kumaran: Like, what is your document intelligence, like,
241 00:21:56.780 ⇒ 00:22:07.679 Uttam Kumaran: like, strategy, or, like, how are you managing, sort of, people QAing? Because a lot of them would have either gone with 365 Copilot, which, like, which sucks, or, like, Glean, which also sucks.
242 00:22:08.220 ⇒ 00:22:14.159 Uttam Kumaran: Or they’re, like, thinking about building their own, which is gonna be impossible at those big companies.
243 00:22:14.160 ⇒ 00:22:15.120 Mike Klaczynski: Yeah. You know?
244 00:22:15.970 ⇒ 00:22:24.349 Mike Klaczynski: Yeah, and it’s… it’s… they already have those documents in SharePoint, so the policies and Active Directories and all that stuff set up.
245 00:22:24.590 ⇒ 00:22:41.220 Mike Klaczynski: And then people are figuring out how to use those documents in a workflow. Well, this is where we can come in and design those workflows for them, and say, hey, for these types of emails or documents that come in, this is how we can process and optimize them into whatever process you need, using our platform.
246 00:22:41.540 ⇒ 00:22:47.040 Mike Klaczynski: We don’t necessarily want to do that, we want you guys to do that, but that’s where we can collaborate together.
247 00:22:47.040 ⇒ 00:22:49.889 Uttam Kumaran: We wanna do that, too. I wanna do that.
248 00:22:49.890 ⇒ 00:22:51.230 Mike Klaczynski: That’s why it’s perfect alignment.
249 00:22:54.330 ⇒ 00:22:54.700 Mike Klaczynski: Yep.
250 00:22:54.700 ⇒ 00:23:00.550 Luke Scorziell: Wait, so… yeah, that makes sense. Oh, I can kind of start thinking about… about that.
251 00:23:00.550 ⇒ 00:23:01.859 Uttam Kumaran: Yeah, I also think there’s also an.
252 00:23:01.860 ⇒ 00:23:06.860 Luke Scorziell: tech side of… or can you maybe dive into that, like, Who that client would be.
253 00:23:07.090 ⇒ 00:23:07.900 Luke Scorziell: Hey, Tom?
254 00:23:08.540 ⇒ 00:23:10.700 Uttam Kumaran: For the SharePoint thing?
255 00:23:10.970 ⇒ 00:23:11.900 Luke Scorziell: Yeah, yeah.
256 00:23:12.210 ⇒ 00:23:14.709 Uttam Kumaran: Oh, yeah, so we have… so CES is a client of ours.
257 00:23:14.710 ⇒ 00:23:15.659 Luke Scorziell: Oh, see, yes, okay.
258 00:23:15.660 ⇒ 00:23:18.859 Uttam Kumaran: Yeah, so they, they just have, like.
259 00:23:18.980 ⇒ 00:23:29.240 Uttam Kumaran: 10, 15, 20 years of just random stuff in a SharePoint, that they’re like… someone… their head of IT tried to use Glean, because people are like, where is this file, blah blah blah. It sucks.
260 00:23:29.870 ⇒ 00:23:41.890 Uttam Kumaran: And they’re like, okay, what do we do? Even additionally, like, just to help them identify how to organize that and things like that. So, like, a lot of the legacy companies with Microsoft Stacks are gonna be in this dilemma.
261 00:23:42.070 ⇒ 00:23:48.699 Uttam Kumaran: Because they wouldn’t have been able to move to Google Drive, or, like, they’re dealing with migration after migration of files.
262 00:23:48.790 ⇒ 00:23:49.989 Mike Klaczynski: And so…
263 00:23:50.260 ⇒ 00:23:55.140 Uttam Kumaran: Yeah, I feel like there’s a… There is an opportunity there.
264 00:23:56.280 ⇒ 00:23:56.650 Luke Scorziell: Yeah.
265 00:23:56.650 ⇒ 00:23:57.240 Uttam Kumaran: You know?
266 00:23:57.420 ⇒ 00:24:01.019 Uttam Kumaran: So CES might be good, a good way to angle into, like, trade shows.
267 00:24:01.190 ⇒ 00:24:04.670 Mike Klaczynski: And that’s, like, CES as in consumer electronics.
268 00:24:04.670 ⇒ 00:24:06.590 Uttam Kumaran: Yeah. Yes, yes, yeah.
269 00:24:06.900 ⇒ 00:24:07.770 Mike Klaczynski: Cool, man.
270 00:24:10.050 ⇒ 00:24:13.489 Mike Klaczynski: And then the other one you mentioned was Plan Healthcare, on the…
271 00:24:13.490 ⇒ 00:24:14.760 Uttam Kumaran: Plan Medicare?
272 00:24:14.760 ⇒ 00:24:15.690 Mike Klaczynski: and Medicare.
273 00:24:15.690 ⇒ 00:24:17.350 Uttam Kumaran: Yeah, this one…
274 00:24:18.140 ⇒ 00:24:22.730 Uttam Kumaran: I don’t know, it’s sort of like a small shop, these guys are kind of all over the place, I’m not sure, like, what’s…
275 00:24:23.690 ⇒ 00:24:27.409 Uttam Kumaran: They’re gonna… we’re gonna build the email thing they need, and then…
276 00:24:28.010 ⇒ 00:24:33.809 Uttam Kumaran: again, it’s like, I’m not sure their appetite for, like, A larger platform, but it’s…
277 00:24:34.860 ⇒ 00:24:40.960 Uttam Kumaran: It’s… I’m gonna… we’re gonna see, like, anywhere that, like, sort of rhymes with lots of documents, we’re sort of…
278 00:24:41.110 ⇒ 00:24:49.569 Uttam Kumaran: pitching contextual. And so, I think another thing, Luke, that I was just thinking about is, like, the agent composer piece, you can basically kind of, like.
279 00:24:50.150 ⇒ 00:24:56.860 Uttam Kumaran: just talk about this versus, like, N8N, maybe, and sort of, like, talk about that, or this versus other sort of, like.
280 00:24:56.970 ⇒ 00:24:58.750 Uttam Kumaran: UI-based workflows.
281 00:25:01.170 ⇒ 00:25:08.060 Uttam Kumaran: Yeah, and I talked to Abhishek a little bit about, kind of, like, why this is, like, of course, a little bit more impressive, you know?
282 00:25:08.230 ⇒ 00:25:13.360 Mike Klaczynski: For sure. Yeah, the challenge with N8N and, like, LaneViews and all those other systems is…
283 00:25:14.810 ⇒ 00:25:21.300 Mike Klaczynski: You’re using this orchestrator and then plugging it into all these different models, so you have all these loose pieces together that gets complicated.
284 00:25:21.640 ⇒ 00:25:31.369 Mike Klaczynski: scale very easily, and isn’t really enterprise-grade or secure. Like, if one thing breaks, and all of a sudden your platform’s down, you might have a thousand people that are sitting there twiddling their thumb, right?
285 00:25:31.530 ⇒ 00:25:32.050 Mike Klaczynski: Yeah.
286 00:25:32.350 ⇒ 00:25:50.880 Mike Klaczynski: we come in is we’re very focused on the enterprise. Like, we haven’t had an outage in 4 months, is my understanding, at least for our big production clients like Qualcomm, right? So, that’s what you’re getting, is not only a powerful platform that has extensibility capabilities and, you know, super accurate, but
287 00:25:51.110 ⇒ 00:25:55.359 Mike Klaczynski: It’s very secure, and it scales very well, and the uptime’s really good.
288 00:25:56.730 ⇒ 00:26:03.440 Luke Scorziell: Cool, okay. So is that… I haven’t used N8N too much. Is it similar to… is that, like, similar to Zapier?
289 00:26:03.820 ⇒ 00:26:08.599 Uttam Kumaran: It’s like, yeah, it’s like the next evolution of, like, that sort of, like, drag-and-drop
290 00:26:09.220 ⇒ 00:26:14.749 Uttam Kumaran: They just reverse the market in this, like, world of, like, drag-and-drop AI work… AI-powered workflows.
291 00:26:15.270 ⇒ 00:26:24.169 Luke Scorziell: So it’s collecting information from… like, it’s having to kind of go external versus contextual… Yeah, you can build, basically, like, connectors, you can say, do this.
292 00:26:24.170 ⇒ 00:26:26.779 Uttam Kumaran: Then do that, you sort of string along tasks.
293 00:26:27.650 ⇒ 00:26:35.220 Uttam Kumaran: So for non-technical folks, it was great, because you can kind of get into it, but, like, as soon as you build a system worth anything, and it actually matters.
294 00:26:35.330 ⇒ 00:26:44.260 Uttam Kumaran: Like, you can’t use… so we’ve moved all of our… originally, we had a lot of our internal stuff on NDN, we moved most of it. We’ve moved almost everything off now.
295 00:26:45.140 ⇒ 00:26:46.840 Uttam Kumaran: Into code, but yeah.
296 00:26:49.390 ⇒ 00:26:52.680 Luke Scorziell: Yeah, no, okay, well, I mean, this gives me a lot to go off of for,
297 00:26:53.660 ⇒ 00:26:58.400 Luke Scorziell: Yeah, I mean, we’ve got the campaigns, and what we’re trying to do, too, is,
298 00:26:58.670 ⇒ 00:27:08.020 Luke Scorziell: Do, like, partnership posts every couple of weeks for different partners, so just kind of highlighting the people that we’re working with and showing that when we work with,
299 00:27:10.940 ⇒ 00:27:21.109 Luke Scorziell: Yeah, just that we are selective about the partners that we have, and so… yeah, so we’re, like, a couple of our partners are, like, we have, like.
300 00:27:21.400 ⇒ 00:27:24.569 Luke Scorziell: I’m putting, like, ad dollars behind that, too, and stuff, but,
301 00:27:24.710 ⇒ 00:27:28.330 Luke Scorziell: Yeah, it’s interesting. Oh, this is… this is helpful.
302 00:27:28.820 ⇒ 00:27:33.979 Mike Klaczynski: Yeah, so this is the new Agent Composer interface, and you can see here…
303 00:27:34.690 ⇒ 00:27:38.650 Mike Klaczynski: you know, this… there’s templates, so, like, this just does, I think, a basic…
304 00:27:39.540 ⇒ 00:27:44.439 Mike Klaczynski: lookup or research step, so it’s just going into different data sources and doing research.
305 00:27:44.560 ⇒ 00:27:54.140 Mike Klaczynski: But we have, I want to say, like, 45 or 50 different additional modules that you can connect and plug into different components. Now, the key here is… is…
306 00:27:54.770 ⇒ 00:27:58.310 Mike Klaczynski: the reasoning step, right? So you’ll actually have a planning
307 00:27:58.410 ⇒ 00:28:08.710 Mike Klaczynski: agent that’ll design a plan and say, okay, this is what I’m actually trying to accomplish, and then you give it a set of tools, and then it’ll pick which tools it wants to go through, goes and executes those tools.
308 00:28:09.270 ⇒ 00:28:18.709 Mike Klaczynski: analyzes the response from that and says, hey, did I actually get what I’m looking for? And then it can readdress and say, oh, you know what, I need to go use a different tool, or I’m gonna re-query this, or retry this.
309 00:28:18.710 ⇒ 00:28:19.120 Luke Scorziell: Hmm.
310 00:28:19.120 ⇒ 00:28:21.550 Mike Klaczynski: And as it’s going through that reasoning process.
311 00:28:22.140 ⇒ 00:28:26.689 Mike Klaczynski: It’s, it’s working down that plan to try to get to that end solution.
312 00:28:26.840 ⇒ 00:28:31.600 Mike Klaczynski: the… I guess the best example of this would be… Log analysis.
313 00:28:31.850 ⇒ 00:28:39.780 Mike Klaczynski: So, if you have software or firmware, you get massive log files. It could be, like, 2GB, right? It’s just literally a log file of everything that’s happened over the last 8 hours.
314 00:28:40.390 ⇒ 00:28:51.430 Mike Klaczynski: doing a root cause analysis of that by a person can take a really long time. Like, they have to first get trained up on, you know, understanding the logs, and then know all the documentation, and know the patterns.
315 00:28:51.660 ⇒ 00:28:59.380 Mike Klaczynski: With a system like ours, we’re able to get some of that root cause analysis down to, like, 20 minutes. It’ll come up with 5 hypotheses of what it might be.
316 00:29:00.880 ⇒ 00:29:16.089 Mike Klaczynski: And that’s back to my earlier example when we were talking about the email, right? It’s not just that it looks at the documents, it might actually look at previously resolved issues and say, this is how I solved a similar problem based on this pattern, and then it knows that answer. And you can orchestrate all that here.
317 00:29:16.770 ⇒ 00:29:26.869 Mike Klaczynski: Now, we’re calling this dynamic as opposed to our previous system, which was a static RAG system. It was literally, you ask a question, it would just give you the answer here.
318 00:29:27.010 ⇒ 00:29:32.060 Mike Klaczynski: it’s dynamic. And then the other thing I’ll say on this, it’s…
319 00:29:32.230 ⇒ 00:29:41.789 Mike Klaczynski: with AI systems, they’re typically non-deterministic, right? You can ask the same prompt and get different answers, just because you don’t know how all the neurons are gonna fire in there.
320 00:29:42.230 ⇒ 00:29:59.130 Mike Klaczynski: But what we’re trying to do is make it more deterministic and say, here is some programming and some protocol and some context based on the workflow that we want this model to follow. So it becomes way more predictable and way more consistent in its responses.
321 00:29:59.810 ⇒ 00:30:00.370 Mike Klaczynski: So…
322 00:30:00.370 ⇒ 00:30:01.370 Luke Scorziell: Yeah, okay.
323 00:30:01.930 ⇒ 00:30:03.749 Mike Klaczynski: I hope that helps a little bit. But…
324 00:30:03.750 ⇒ 00:30:05.010 Luke Scorziell: No, no, it’s helpful.
325 00:30:05.010 ⇒ 00:30:13.449 Mike Klaczynski: I’ll share the slides that we have, and as I said, you know, over the next week or two, we’ll give you guys some demos, and it’ll start to click more.
326 00:30:13.650 ⇒ 00:30:14.180 Luke Scorziell: Okay.
327 00:30:14.180 ⇒ 00:30:14.630 Mike Klaczynski: Unreally.
328 00:30:14.630 ⇒ 00:30:31.120 Luke Scorziell: Yeah, that’d be great, and we’re happy to, like, post, like, if Utam can do, like, a walkthrough, or if we do, like, a demo, and we want to just screen record it, and then put it up, like, somewhere, we’re happy to do that, too. So I think, it’s like, the more that people can just see these products instead of
329 00:30:31.290 ⇒ 00:30:35.879 Luke Scorziell: hear about them. I think the, like, more it makes sense as to how they fit into…
330 00:30:36.020 ⇒ 00:30:37.380 Luke Scorziell: It’s everyday life, so…
331 00:30:38.090 ⇒ 00:30:46.039 Mike Klaczynski: Yeah, my old COO… would say, show it, sell it, hide it, keep it.
332 00:30:46.040 ⇒ 00:30:48.259 Uttam Kumaran: Wait, what is the… what does the second part mean?
333 00:30:48.550 ⇒ 00:30:52.600 Mike Klaczynski: So, you know, show it, and then you’ll sell it. If you hide it, then you’re gonna keep it.
334 00:30:52.600 ⇒ 00:30:53.490 Uttam Kumaran: Oh…
335 00:30:53.490 ⇒ 00:30:54.070 Luke Scorziell: Oh.
336 00:30:54.070 ⇒ 00:30:54.979 Uttam Kumaran: Hey, I see.
337 00:30:55.240 ⇒ 00:30:58.370 Mike Klaczynski: This is Mark Cranny from A16Z. He worked with,
338 00:30:58.680 ⇒ 00:31:01.570 Mike Klaczynski: Yeah. He wrote… he’s in a couple books, so…
339 00:31:01.570 ⇒ 00:31:02.190 Uttam Kumaran: Yeah, yeah, yeah.
340 00:31:02.190 ⇒ 00:31:07.180 Mike Klaczynski: Hardass, but, it works. Show it, sell it, you know? Hide it, keep it.
341 00:31:08.010 ⇒ 00:31:09.169 Uttam Kumaran: Yeah, no, you’re right.
342 00:31:09.170 ⇒ 00:31:09.750 Holly Condos: It’s good.
343 00:31:09.750 ⇒ 00:31:15.520 Luke Scorziell: I mean, that’s what we’re trying to do. We’re trying to launch, like, a Build in Public series, because it looks like all of our…
344 00:31:15.650 ⇒ 00:31:26.879 Luke Scorziell: everyone at Brainforge is building, like, super cool, workflows and using AI to do, like, crazy things, and it’s just… and for clients, for ourselves, so it’s just, like, the more that we get that out.
345 00:31:27.250 ⇒ 00:31:31.739 Uttam Kumaran: We’re all mostly engineers, so nobody’s, like, used to…
346 00:31:32.220 ⇒ 00:31:38.570 Uttam Kumaran: posting stuff, so we’re trying to, like, flip it a little bit, but that’s Luke’s job to, like, get it out of us, so…
347 00:31:38.860 ⇒ 00:31:46.989 Mike Klaczynski: Yeah, I mean… I’m hounding them. What we’re seeing is really, you have to paint the art of the possible. Like, I made the example earlier. Like, we go in and…
348 00:31:47.760 ⇒ 00:31:54.810 Mike Klaczynski: most of these executives really don’t know, right? There’s another partner we’re working with that wants us to build some CAD support.
349 00:31:54.810 ⇒ 00:32:12.770 Mike Klaczynski: And he’s like, yeah, you know, we want you guys to be able to do AI over CAD. And we said, well, what does that mean? He’s like, well, okay, we’ve got all these engineers in design review meetings, like, can you help them? And we said, okay, what does that mean? He’s like, can you do ballooning and annotation? I’m like, okay, what does that actually mean? So, like, it really needs to be… what AI does is it says.
350 00:32:13.220 ⇒ 00:32:20.720 Mike Klaczynski: there’s a process a person does, and now I actually want to replace that process, or augment it, or speed it up. Yeah. And I actually have to understand what that.
351 00:32:20.720 ⇒ 00:32:23.819 Uttam Kumaran: Yeah, don’t just say AI over CAD.
352 00:32:23.820 ⇒ 00:32:24.780 Mike Klaczynski: Exactly, right?
353 00:32:24.780 ⇒ 00:32:29.599 Uttam Kumaran: Well, you’d be surprised, that’s probably all that guy knows, is, like, there’s something to do with CAD files.
354 00:32:29.770 ⇒ 00:32:33.620 Uttam Kumaran: And… I gotta hire all these CAD people.
355 00:32:33.620 ⇒ 00:32:37.260 Mike Klaczynski: AI in a CAD and, like, let’s make money. I’m like, no, no, no, no.
356 00:32:37.570 ⇒ 00:32:38.520 Uttam Kumaran: Yeah, yeah.
357 00:32:39.110 ⇒ 00:32:48.590 Mike Klaczynski: So, to your point, we have to demonstrate it, show it, paint the art of the possible, and make it tangible. So, that’s what we’re working on as well. That’s why, like.
358 00:32:48.590 ⇒ 00:32:48.910 Luke Scorziell: Down.
359 00:32:48.910 ⇒ 00:32:57.820 Mike Klaczynski: Initially in the launch, we were gonna have a couple templates, but we said, no, we have to show people what this guy has to do, and get them hands-on with it.
360 00:32:58.640 ⇒ 00:33:04.169 Luke Scorziell: Yeah, yeah, I mean, I’m… I don’t come from a technical background, I have a journalism background, and then,
361 00:33:04.320 ⇒ 00:33:07.200 Luke Scorziell: Some, like, agency on the marketing side.
362 00:33:07.350 ⇒ 00:33:17.329 Luke Scorziell: But I’m, like, using… using cursor code, like, just figuring it out, how to use it, and I like playing around with all the different AI tools and stuff, too, so it’s been fun to…
363 00:33:17.500 ⇒ 00:33:22.039 Luke Scorziell: to be here somewhere. It’ll be fun to get my hands on, on, Contextual, too, so…
364 00:33:22.040 ⇒ 00:33:22.430 Mike Klaczynski: Yeah.
365 00:33:22.430 ⇒ 00:33:23.370 Luke Scorziell: We’ll see.
366 00:33:23.370 ⇒ 00:33:27.199 Mike Klaczynski: Have you guys tried MoltBot, formerly known as Clodbot?
367 00:33:27.200 ⇒ 00:33:36.439 Uttam Kumaran: No, I have… I’m, like, waiting for the weekend, you know? It’s… everybody’s texting me, I have, like, a day job. Like, I’m, like, trying to do… I’m, like, trying to do stuff, I can’t build…
368 00:33:36.670 ⇒ 00:33:41.679 Uttam Kumaran: WhatsApp bots to, like, buy my food right now. Yeah.
369 00:33:41.680 ⇒ 00:33:44.339 Mike Klaczynski: It was, like, you know, after midnight.
370 00:33:45.110 ⇒ 00:33:46.700 Mike Klaczynski: and I went down that rabbit hole, and I’m like.
371 00:33:46.700 ⇒ 00:33:51.209 Uttam Kumaran: I know, I just can’t do that, I can’t spend 4 hours on that right now.
372 00:33:51.210 ⇒ 00:33:54.519 Mike Klaczynski: Yeah, and it sounds like it’s just a bunch of fluff anyway, but…
373 00:33:54.520 ⇒ 00:33:55.280 Uttam Kumaran: Awesome.
374 00:33:55.280 ⇒ 00:33:56.300 Luke Scorziell: I’ll check it out.
375 00:33:56.300 ⇒ 00:33:59.040 Uttam Kumaran: Every week, it’s every week with something, I feel like, so…
376 00:33:59.510 ⇒ 00:34:00.040 Mike Klaczynski: Yep.
377 00:34:01.590 ⇒ 00:34:02.050 Uttam Kumaran: Okay.
378 00:34:02.050 ⇒ 00:34:03.260 Luke Scorziell: Cool. Cool.
379 00:34:03.400 ⇒ 00:34:13.980 Uttam Kumaran: Thank you, everyone. Yeah, Mike, we’ll send some stuff in Slack as soon as we, like, get everything together and the insurance, and then if we end up posting anything, you know, we’ll send it to you to kind of boost, if you don’t mind, and that’d be great.
380 00:34:13.989 ⇒ 00:34:17.209 Mike Klaczynski: Yeah, we’ll absolutely boost it and amplify it, and I’ll send you guys the deck.
381 00:34:17.599 ⇒ 00:34:18.219 Holly Condos: Sounds good.
382 00:34:18.219 ⇒ 00:34:21.369 Luke Scorziell: Yeah, that’d be great. Thank you guys for the time, really helpful.
383 00:34:21.370 ⇒ 00:34:22.650 Uttam Kumaran: Thank you. Thank you.
384 00:34:23.079 ⇒ 00:34:24.249 Uttam Kumaran: See you later. Talk to you soon.
385 00:34:24.250 ⇒ 00:34:24.690 Luke Scorziell: Bucks in.
386 00:34:24.690 ⇒ 00:34:25.239 Uttam Kumaran: I…