Meeting Title: Form-Filling Tech Demo and Discussion Date: 2025-08-21 Meeting participants: Ryan DeForest, Lev Katreczko, Henry Zhao, Uttam Kumaran, Mustafa Raja
WEBVTT
1 00:02:39.460 ⇒ 00:02:40.370 Uttam Kumaran: Hey guys.
2 00:02:52.860 ⇒ 00:02:59.090 Henry Zhao: Once again, I can just go to the background. You’re still in state.
3 00:03:05.190 ⇒ 00:03:07.010 Uttam Kumaran: Nice, good to see y’all.
4 00:03:07.990 ⇒ 00:03:12.299 Henry Zhao: We’re all in the same room, but on the same Zoom, you know, that’s how we’re… Lev’s in space, though.
5 00:03:12.300 ⇒ 00:03:16.379 Uttam Kumaran: Yeah, that’s the comical part of life now.
6 00:03:16.990 ⇒ 00:03:18.360 Uttam Kumaran: …
7 00:03:18.730 ⇒ 00:03:26.729 Uttam Kumaran: Cool, Mustafa should be here in a sec. We got, you know, a lot of the form-filling stuff to work this week, so we tested it with
8 00:03:27.350 ⇒ 00:03:32.110 Uttam Kumaran: … The five, 5 existing clients.
9 00:03:32.210 ⇒ 00:03:52.169 Uttam Kumaran: That we were looking… 5 existing leads we were looking at, and 5 new ones, and we’re actually able to do, a couple things. One, we were able to strip the HTML at multiple stages of the form fill, and we have… we were able to get structured outputs about, like, what the, tools they’re using are.
10 00:03:52.400 ⇒ 00:03:55.980 Uttam Kumaran: So maybe, Mustafa, do you want to share that.
11 00:03:56.170 ⇒ 00:04:00.250 Uttam Kumaran: We could just run through, like, the example you just showed me.
12 00:04:00.510 ⇒ 00:04:01.610 Uttam Kumaran: …
13 00:04:01.730 ⇒ 00:04:09.999 Uttam Kumaran: But we’re kind of… we’re basically kind of at 80-90% accuracy, I, you know, compared to where we were last week, so….
14 00:04:10.000 ⇒ 00:04:14.090 Henry Zhao: Is it to the point where I could, like, give you a domain and, like, test it out?
15 00:04:14.410 ⇒ 00:04:20.010 Uttam Kumaran: Yeah, we should do that. … Yeah.
16 00:04:20.370 ⇒ 00:04:24.949 Uttam Kumaran: Mustafa, feel free to share, and then, yeah, I don’t know, I know, I said…
17 00:04:25.060 ⇒ 00:04:38.680 Uttam Kumaran: don’t show too much, … I said, like, don’t walk too much into the technical details. We did a write-up on it, so I could share that. I can also go through that, if that’s interesting. Basically, we’re using, …
18 00:04:38.880 ⇒ 00:04:43.160 Uttam Kumaran: Like, headless browser technology, but also, …
19 00:04:43.580 ⇒ 00:04:50.579 Uttam Kumaran: we’re using some AI tech that allows us to actually, like, look at the screen, identify dynamically the forms.
20 00:04:50.780 ⇒ 00:05:01.599 Uttam Kumaran: and then fill it… fill it in with things. So, we have a loom on it, but maybe, Mustafa, like, if we’re… this result we’re looking at right now is for PlanHat.com.
21 00:05:01.600 ⇒ 00:05:06.959 Henry Zhao: What you’re seeing as the result set is, like, a structured output, so we both see, like.
22 00:05:06.960 ⇒ 00:05:08.360 Uttam Kumaran: It has a form.
23 00:05:08.710 ⇒ 00:05:19.699 Uttam Kumaran: the form type is simple, there is a scheduler, it looks like HubSpot meetings, and so we’re able to, you know, extract that. Mustafa, do you want to show the actual website?
24 00:05:20.460 ⇒ 00:05:22.090 Mustafa Raja: … yeah.
25 00:05:29.190 ⇒ 00:05:32.580 Uttam Kumaran: Yeah, so Plan Ad is an example, so if you click on Request Demo.
26 00:05:33.210 ⇒ 00:05:47.330 Mustafa Raja: once it, once it clicks on the request demo, this sort of sidebar pops up, and then it looks into the HTML, and it finds that, this thing is based on HubSpot.
27 00:05:47.740 ⇒ 00:05:52.240 Mustafa Raja: Yeah, let me know if you need to run… do a run or something.
28 00:05:52.430 ⇒ 00:05:55.550 Uttam Kumaran: Yeah, do you wanna… can we try one? Yeah, Ryan.
29 00:05:56.230 ⇒ 00:05:57.220 Henry Zhao: Hey, everyone.
30 00:05:57.760 ⇒ 00:05:59.239 Uttam Kumaran: Yeah, shoot us with one.
31 00:05:59.460 ⇒ 00:06:01.529 Henry Zhao: Try this one.
32 00:06:04.050 ⇒ 00:06:06.400 Mustafa Raja: Do you want to do something other than….
33 00:06:06.530 ⇒ 00:06:09.020 Uttam Kumaran: Yeah, it’s in the, it’s in the Zoom chat, but…
34 00:06:09.200 ⇒ 00:06:19.019 Uttam Kumaran: Let’s, let’s also just share this, like, how it’s working. So this is actually, for folks on the call, this is a simulated browser, and this is an AI agent filling this out right now.
35 00:06:19.370 ⇒ 00:06:23.019 Uttam Kumaran: So, it goes through, and it identifies
36 00:06:23.590 ⇒ 00:06:26.139 Uttam Kumaran: And it clicks, and then it opens.
37 00:06:26.380 ⇒ 00:06:32.170 Uttam Kumaran: And then I believe, yeah, it’s gonna go… at that point, it just identifies, like, that this is HubSpot.
38 00:06:35.720 ⇒ 00:06:39.139 Mustafa Raja: Mmm… I can….
39 00:06:39.410 ⇒ 00:06:39.950 Uttam Kumaran: It’s….
40 00:06:39.950 ⇒ 00:06:41.409 Mustafa Raja: Yeah, I’ve got it, I’ve got it.
41 00:06:41.410 ⇒ 00:06:42.160 Uttam Kumaran: Got it.
42 00:06:42.160 ⇒ 00:06:44.390 Henry Zhao: Yeah, that’s pretty sick.
43 00:06:46.770 ⇒ 00:06:48.299 Mustafa Raja: Yeah, let’s see if this works.
44 00:07:05.030 ⇒ 00:07:09.239 Uttam Kumaran: Can you open the… the site right now? Is there… is there a CTA?
45 00:07:14.500 ⇒ 00:07:17.620 Mustafa Raja: So that might be kind of the CTAs get started.
46 00:07:18.650 ⇒ 00:07:21.160 Mustafa Raja: Cj, let me add that.
47 00:07:22.390 ⇒ 00:07:24.780 Henry Zhao: It’s not like a request. Yeah, perfect.
48 00:07:57.890 ⇒ 00:07:58.970 Uttam Kumaran: Nice.
49 00:07:59.380 ⇒ 00:08:02.409 Henry Zhao: It just does it all behind the cookies and pops up as briefly.
50 00:08:02.410 ⇒ 00:08:03.050 Uttam Kumaran: Yeah.
51 00:08:03.050 ⇒ 00:08:05.860 Mustafa Raja: You can actually accept Yo, yay.
52 00:08:06.250 ⇒ 00:08:06.860 Uttam Kumaran: Yeah.
53 00:08:06.860 ⇒ 00:08:09.010 Mustafa Raja: I’ve commented this out, so….
54 00:08:09.500 ⇒ 00:08:14.069 Henry Zhao: No, you’re fine, just, I think it’s funny that it’s just, like, screw it.
55 00:08:14.320 ⇒ 00:08:14.650 Mustafa Raja: Yeah.
56 00:08:14.650 ⇒ 00:08:15.850 Uttam Kumaran: Okay, Hot spot, yeah.
57 00:08:16.270 ⇒ 00:08:17.099 Mustafa Raja: So, we have chili.
58 00:08:17.100 ⇒ 00:08:18.620 Uttam Kumaran: I’ll spell Anchili Viper.
59 00:08:19.030 ⇒ 00:08:21.319 Mustafa Raja: And then we have, for the forms, HubSpot.
60 00:08:22.690 ⇒ 00:08:27.329 Henry Zhao: Pretty sick. I gave you that one on purpose, because me and Lev were literally looking at that one yesterday.
61 00:08:27.890 ⇒ 00:08:35.039 Uttam Kumaran: Live demos! Dude, you’re making me nervous. I used to be a product manager, and, like, I… I literally would never allow this to happen.
62 00:08:35.049 ⇒ 00:08:36.129 Henry Zhao: That’s the ops guys.
63 00:08:36.140 ⇒ 00:08:46.290 Uttam Kumaran: works out. People would ask us, like, we would be presenting the company. I’m like, you have to talk to us after, we don’t have things set up, like…
64 00:08:46.290 ⇒ 00:08:57.650 Uttam Kumaran: But it’s pretty good, and so this is all built on, like, technology that’s only been in the market for, like, about a year. Browser-based, but also headless browsers and agentic browser technology, so…
65 00:08:57.760 ⇒ 00:09:14.650 Uttam Kumaran: like, very, very, like, on the edge. I don’t know… I mean, there are some people that are using… this is, like, Lev… this is, like, kind of what Claggent is built similarly off of, but what you’re seeing is that there’s no way that you could probably have this level of customization specific to form filling.
66 00:09:15.040 ⇒ 00:09:21.270 Henry Zhao: And from this, this payload in general, like, what does, like, form type, or form type mean?
67 00:09:21.370 ⇒ 00:09:22.090 Henry Zhao: Yeah.
68 00:09:22.090 ⇒ 00:09:24.430 Uttam Kumaran: Mustafa, do you want to go into the logic for….
69 00:09:24.430 ⇒ 00:09:30.660 Mustafa Raja: This form type. So, if we come over to…
70 00:09:30.810 ⇒ 00:09:33.560 Mustafa Raja: Apache, and this is something that…
71 00:09:34.280 ⇒ 00:09:41.359 Mustafa Raja: It might struggle with… so, if you book… want to book a demo… do you see that this is… this… this form is very…
72 00:09:41.510 ⇒ 00:09:45.690 Mustafa Raja: Different, or we can say non-conventional.
73 00:09:46.550 ⇒ 00:09:47.460 Mustafa Raja: ….
74 00:09:47.460 ⇒ 00:09:49.989 Uttam Kumaran: But that’s something… that’s, like, our logic, right?
75 00:09:50.190 ⇒ 00:09:51.360 Mustafa Raja: Yeah, yeah, yeah.
76 00:09:51.360 ⇒ 00:09:54.460 Uttam Kumaran: Can you walk through the logic for….
77 00:09:54.810 ⇒ 00:10:18.349 Mustafa Raja: Yeah, so if it’s going to be a simple form, meaning a form that has multiple fields in it that can be filled, it’s going to take that particular route, you know, it’s a form on the page, and it just needs to fill up the fields. If it’s a multi-step form, it only has two types.
78 00:10:18.350 ⇒ 00:10:24.970 Mustafa Raja: For now, I have only two types, let me… Go and find that somewhere.
79 00:10:25.310 ⇒ 00:10:27.590 Uttam Kumaran: But we can… we can expand, right? Like, this is.
80 00:10:27.590 ⇒ 00:10:29.580 Mustafa Raja: Yeah, yeah, we can expand.
81 00:10:30.110 ⇒ 00:10:43.590 Uttam Kumaran: form world, so I… we were just, like… it seems like there’s roughly two types. There’s, like, one single pane, or there’s, like, something that’s more walkthrough dynamic, but we can… we can expand this payload, long story short.
82 00:10:43.590 ⇒ 00:10:56.069 Henry Zhao: There’s more of a question just for my knowledge, less of, like, what is it? It’s more like, is there anything else that we can… that cool that we can do here? For sake of… of live demo again, can we try one more?
83 00:10:58.560 ⇒ 00:10:59.110 Mustafa Raja: Yeah, yeah.
84 00:10:59.110 ⇒ 00:11:00.160 Uttam Kumaran: Zoom chat.
85 00:11:04.850 ⇒ 00:11:12.960 Uttam Kumaran: Yeah, I mean, I don’t know what’s… I guess, like, I’m open to hearing, like, what’s helpful. I mean, we could do… we could… we could have a screenshot, we could scrape…
86 00:11:13.180 ⇒ 00:11:19.160 Uttam Kumaran: Other info, but, … Yeah.
87 00:11:21.420 ⇒ 00:11:22.699 Henry Zhao: Yeah, that’s a good question.
88 00:11:35.210 ⇒ 00:11:36.120 Uttam Kumaran: Nice.
89 00:11:46.270 ⇒ 00:11:50.389 Henry Zhao: Their… their GP2’s 4.5 out of 5.
90 00:11:51.940 ⇒ 00:11:53.579 Henry Zhao: That’s barely the A wins.
91 00:11:57.940 ⇒ 00:11:59.720 Henry Zhao: The guy that saw it, and I don’t used it.
92 00:12:03.810 ⇒ 00:12:07.990 Henry Zhao: Also, this could be… getting stuck.
93 00:12:10.090 ⇒ 00:12:11.310 Uttam Kumaran: Okay, it’s clicking.
94 00:12:12.070 ⇒ 00:12:12.810 Henry Zhao: Alright.
95 00:12:19.270 ⇒ 00:12:23.339 Uttam Kumaran: Yeah, so the only thing that’s, like, probably most…
96 00:12:24.380 ⇒ 00:12:30.229 Uttam Kumaran: problematic about this system is it’s a little bit slow these days. Like, it has to go through step by step by step.
97 00:12:30.690 ⇒ 00:12:31.360 Uttam Kumaran: ….
98 00:12:31.360 ⇒ 00:12:32.050 Mustafa Raja: Yeah, so…
99 00:12:33.900 ⇒ 00:12:41.240 Mustafa Raja: So it did find out that the form is multi-step, and then the provider of the form, it said, is Marketo.
100 00:12:41.620 ⇒ 00:12:44.669 Mustafa Raja: It wasn’t able to find any scheduler, though.
101 00:12:45.640 ⇒ 00:12:52.050 Henry Zhao: What’s funny is… Let me share my screen, I’ll show you real quick.
102 00:12:52.700 ⇒ 00:12:53.440 Mustafa Raja: Yep.
103 00:12:55.290 ⇒ 00:12:57.609 Henry Zhao: So when I go to Sendozo.
104 00:12:58.090 ⇒ 00:13:00.800 Uttam Kumaran: Yeah, I didn’t get a… I didn’t get a gift option.
105 00:13:00.800 ⇒ 00:13:03.060 Henry Zhao: No, I didn’t get a workflow.
106 00:13:04.480 ⇒ 00:13:05.540 Uttam Kumaran: It’s weird.
107 00:13:05.540 ⇒ 00:13:07.170 Henry Zhao: I went straight to this.
108 00:13:07.460 ⇒ 00:13:10.169 Uttam Kumaran: I have a different one, actually.
109 00:13:10.370 ⇒ 00:13:12.940 Uttam Kumaran: Oh, wait, you’re on their home… you’re on their…
110 00:13:13.620 ⇒ 00:13:17.530 Uttam Kumaran: Oh, it’s like a… I ha- I get a different thing.
111 00:13:19.310 ⇒ 00:13:20.429 Henry Zhao: Pull out is, yeah.
112 00:13:20.430 ⇒ 00:13:29.059 Mustafa Raja: It might be because related to the, what’s it called? Dimensions of the browser.
113 00:13:29.840 ⇒ 00:13:33.460 Uttam Kumaran: Oh, really? Like, this is what I… I just sent in the Zoom, like, this is what I get.
114 00:13:34.590 ⇒ 00:13:47.429 Henry Zhao: Yeah, one thing to call out here is that regardless of the demo flow, there’s Revenue Hero and lean data code and HTML, like, very easily sensible.
115 00:13:47.430 ⇒ 00:14:04.169 Henry Zhao: So, aside from, like, the walkthrough, you would be able to pull their tech stack, like, pretty quickly with, like, a much simpler method, so I’m curious if there’s a way to work in, like, either an initial check through that method, or a fallback if it doesn’t find anything.
116 00:14:04.300 ⇒ 00:14:13.439 Henry Zhao: Because this would be a scenario where, like, it didn’t return any output, but actually, you could probably have scraped output much easier than
117 00:14:13.630 ⇒ 00:14:14.420 Henry Zhao: And we’re….
118 00:14:14.420 ⇒ 00:14:22.119 Uttam Kumaran: Yeah, so, Mustafa, the existing Clay workflow that we had is ripping through the HTML, right? Like, the one we demoed last week?
119 00:14:23.510 ⇒ 00:14:24.330 Mustafa Raja: Yes.
120 00:14:24.330 ⇒ 00:14:26.029 Uttam Kumaran: The first step of the waterfall.
121 00:14:26.190 ⇒ 00:14:26.850 Mustafa Raja: Yeah.
122 00:14:27.040 ⇒ 00:14:29.610 Henry Zhao: So you would be doing that first, and then the Bob.
123 00:14:29.610 ⇒ 00:14:38.260 Uttam Kumaran: So this is almost like another signal just for the forms. What we demoed last week still stands, where we are going through the HTML first.
124 00:14:38.740 ⇒ 00:14:42.580 Uttam Kumaran: And then, if we don’t have it, we then kind of waterfall.
125 00:14:42.580 ⇒ 00:14:43.090 Henry Zhao: Boop.
126 00:14:43.090 ⇒ 00:14:50.250 Uttam Kumaran: I think this is also another source of, like, maybe we can strip the HTML from these steps and add that, but….
127 00:14:50.250 ⇒ 00:14:54.529 Henry Zhao: Is that still… is that still alive to test? Like, if we put Sendozo in that first step.
128 00:14:54.530 ⇒ 00:14:56.509 Uttam Kumaran: Yeah, let’s do that. I mean, …
129 00:14:56.740 ⇒ 00:15:01.039 Uttam Kumaran: Maybe, Mustafa, we’re gonna pull up Clay, and we can just run through Sendoso for….
130 00:15:01.040 ⇒ 00:15:04.089 Henry Zhao: Just to see if it’ll show revenue hero and lean data.
131 00:15:04.090 ⇒ 00:15:04.720 Uttam Kumaran: Yeah.
132 00:15:35.630 ⇒ 00:15:36.510 Mustafa Raja: Hmm….
133 00:15:49.710 ⇒ 00:15:53.389 Uttam Kumaran: Yeah, can we test it for Sendoso.com?
134 00:15:54.060 ⇒ 00:15:58.300 Mustafa Raja: … Let me add a row for that.
135 00:15:59.310 ⇒ 00:16:02.019 Henry Zhao: We could just… he could just add the website, right?
136 00:16:02.020 ⇒ 00:16:02.730 Mustafa Raja: Yeah.
137 00:16:03.550 ⇒ 00:16:04.600 Mustafa Raja: ….
138 00:16:04.600 ⇒ 00:16:07.480 Uttam Kumaran: It’s just send, and then OSO.com.
139 00:16:13.870 ⇒ 00:16:14.570 Uttam Kumaran: Yeah.
140 00:16:40.440 ⇒ 00:16:42.510 Uttam Kumaran: Can you click on the enrichment present?
141 00:16:43.120 ⇒ 00:16:43.720 Uttam Kumaran: Blessed?
142 00:16:43.720 ⇒ 00:16:45.130 Mustafa Raja: … this one?
143 00:16:45.840 ⇒ 00:16:50.849 Uttam Kumaran: Yeah, the other column, to the, the one that says 8, yeah, yeah, yeah. Let’s see that.
144 00:16:53.840 ⇒ 00:16:54.580 Uttam Kumaran: Okay.
145 00:16:57.090 ⇒ 00:16:59.599 Henry Zhao: It might not be looking for revenue here, though, right?
146 00:16:59.600 ⇒ 00:17:01.479 Uttam Kumaran: Can you keep that open, Stoffa?
147 00:17:01.480 ⇒ 00:17:03.110 Mustafa Raja: Yeah, the, ….
148 00:17:03.550 ⇒ 00:17:04.160 Uttam Kumaran: Yeah.
149 00:17:08.780 ⇒ 00:17:09.800 Uttam Kumaran: Yeah, hello.
150 00:17:13.200 ⇒ 00:17:14.940 Henry Zhao: You said it’s on the front page?
151 00:17:14.940 ⇒ 00:17:18.709 Uttam Kumaran: I don’t know if we… I mean, there’s kind of two questions, so one….
152 00:17:18.710 ⇒ 00:17:19.150 Henry Zhao: Excellent.
153 00:17:19.150 ⇒ 00:17:23.539 Uttam Kumaran: I think we should take a look. I don’t know if we are looking for Revenue Hero, maybe we can….
154 00:17:23.740 ⇒ 00:17:25.349 Henry Zhao: Yeah, there’s a good….
155 00:17:30.240 ⇒ 00:17:33.400 Mustafa Raja: So are we looking for a certain tool over here?
156 00:17:33.400 ⇒ 00:17:35.849 Uttam Kumaran: Yeah, there’s a tool called Revenue Hero.
157 00:17:35.850 ⇒ 00:17:37.709 Henry Zhao: Let me… I can share my screen, I’ll show you.
158 00:17:37.710 ⇒ 00:17:41.539 Mustafa Raja: We, we can search, search through here if, if it is….
159 00:17:41.540 ⇒ 00:17:43.460 Uttam Kumaran: Just type it in here at the top.
160 00:17:48.040 ⇒ 00:17:49.560 Uttam Kumaran: Yeah, not really.
161 00:17:49.560 ⇒ 00:17:56.049 Henry Zhao: Nope. Or what? Look up, book it, like, B-O-O-K-I-T.
162 00:17:59.670 ⇒ 00:18:02.009 Henry Zhao: How about… how about lean data?
163 00:18:03.080 ⇒ 00:18:04.810 Uttam Kumaran: So, Lean Data was there.
164 00:18:05.280 ⇒ 00:18:07.870 Henry Zhao: Yeah, I was there, like, oh, sorry, I missed that, yeah.
165 00:18:07.870 ⇒ 00:18:16.880 Uttam Kumaran: Yeah, lean data is here, book it, it didn’t get. But maybe, Mustafa, we also… because we’re going through the other process for the form anyways.
166 00:18:16.880 ⇒ 00:18:17.330 Henry Zhao: Okay.
167 00:18:17.330 ⇒ 00:18:21.350 Uttam Kumaran: Might as well run this same process in that step.
168 00:18:21.750 ⇒ 00:18:22.820 Uttam Kumaran: Also.
169 00:18:23.540 ⇒ 00:18:30.370 Mustafa Raja: Yeah, so, by the… we mean that we want to get these technologies from their website, right?
170 00:18:30.740 ⇒ 00:18:35.089 Uttam Kumaran: Yeah, but for example, like, if you look at the screenshot Ryan sent.
171 00:18:35.400 ⇒ 00:18:38.120 Uttam Kumaran: There is a Revenue Hero script.
172 00:18:39.550 ⇒ 00:18:40.870 Uttam Kumaran: in the JS on the page.
173 00:18:40.890 ⇒ 00:18:41.860 Mustafa Raja: Okay, yeah.
174 00:18:42.320 ⇒ 00:18:46.090 Henry Zhao: Yeah, if it’s in the… if it’s in there, then we should be able to look into it.
175 00:18:54.010 ⇒ 00:18:57.409 Ryan DeForest: Obviously, I’m obviously giving you, like, a super….
176 00:18:57.890 ⇒ 00:19:00.610 Henry Zhao: No, no, this is good. This is, like, we need to….
177 00:19:01.000 ⇒ 00:19:04.800 Uttam Kumaran: We’re gonna… it’s… it’s gonna be an iteration process.
178 00:19:06.010 ⇒ 00:19:11.389 Uttam Kumaran: But I am, like, I think we’re hovering closer to, like, 70-80%.
179 00:19:11.570 ⇒ 00:19:16.400 Uttam Kumaran: I think also as we start to identify things we’re missing.
180 00:19:16.660 ⇒ 00:19:21.700 Uttam Kumaran: we’ll just layer it on. So, yeah, I think, Mustafa, one… probably my feedback is, like.
181 00:19:21.930 ⇒ 00:19:34.929 Uttam Kumaran: we should have two enrichment steps, like, one that happens here in Clay, but I know we’re gonna… we’re gonna end up hooking Clay into the computer use thing, so it would be great to also pull
182 00:19:35.380 ⇒ 00:19:37.710 Uttam Kumaran: use the HTML we get from there.
183 00:19:38.070 ⇒ 00:19:48.079 Uttam Kumaran: I’m just like, yeah, one thing we should test is why this Revenue Hero thing didn’t get caught. Like, I guess my question is, what does it think Revenue Hero is?
184 00:19:48.960 ⇒ 00:19:50.410 Mustafa Raja: Hmm. And so….
185 00:19:50.410 ⇒ 00:19:51.630 Uttam Kumaran: It should not work.
186 00:19:51.830 ⇒ 00:20:05.060 Mustafa Raja: Yeah, so we, we are getting these sorts of tools from their stack. Their stack gets these tools, from the job descriptions that Sendiso has been posting.
187 00:20:05.060 ⇒ 00:20:10.839 Uttam Kumaran: But I guess my other point was, wasn’t there… didn’t we add a step that is scraping the HTML?
188 00:20:11.280 ⇒ 00:20:15.929 Mustafa Raja: Yeah, that, that would be, for…
189 00:20:17.330 ⇒ 00:20:22.550 Mustafa Raja: For dual path, dual conversion path, and… Full form friction.
190 00:20:23.030 ⇒ 00:20:26.999 Uttam Kumaran: Like, remember we added the initial waterfall before their stack?
191 00:20:29.440 ⇒ 00:20:34.079 Mustafa Raja: That waterfall was in, I guess, this one.
192 00:20:36.940 ⇒ 00:20:40.239 Mustafa Raja: Yeah, this was, the waterfall was for the jobs.
193 00:20:40.430 ⇒ 00:20:41.940 Uttam Kumaran: Oh, okay, okay, okay.
194 00:20:42.670 ⇒ 00:20:43.220 Mustafa Raja: Yeah.
195 00:20:43.590 ⇒ 00:20:51.799 Uttam Kumaran: Yeah, so I think… I think you’re actually gonna be able to get this from the computer use thing also, and then we should add that as a first step for Waterfall.
196 00:20:52.730 ⇒ 00:20:53.670 Mustafa Raja: Yeah.
197 00:20:54.200 ⇒ 00:21:01.420 Uttam Kumaran: So, not only should the computer use, pull the form details.
198 00:21:01.540 ⇒ 00:21:10.570 Uttam Kumaran: But as it steps through the site, very easy for us to just rip the HTML and start to build that list as we go. And that should be the first step of the waterfall.
199 00:21:16.540 ⇒ 00:21:17.450 Uttam Kumaran: Okay.
200 00:21:17.870 ⇒ 00:21:18.690 Uttam Kumaran: Cool.
201 00:21:19.580 ⇒ 00:21:20.340 Uttam Kumaran: …
202 00:21:20.720 ⇒ 00:21:28.530 Uttam Kumaran: So I guess my ask is going to be, can we test with a larger group, or Ryan, how do you feel about trying to get.
203 00:21:29.210 ⇒ 00:21:32.680 Ryan DeForest: Oh, I’m ready. I got… I got a whole list for you.
204 00:21:32.680 ⇒ 00:21:33.980 Uttam Kumaran: Okay, okay.
205 00:21:34.640 ⇒ 00:21:35.900 Uttam Kumaran: …
206 00:21:36.330 ⇒ 00:21:48.890 Uttam Kumaran: So maybe we can run through… again, like, I think we kind of crawl, walk, run, so if we feel good about this crawl step, then we can start to walk towards, you know, the….
207 00:21:48.890 ⇒ 00:21:49.580 Ryan DeForest: brace.
208 00:21:49.580 ⇒ 00:21:50.609 Uttam Kumaran: 40 leads.
209 00:21:50.610 ⇒ 00:21:51.370 Ryan DeForest: I agree.
210 00:21:51.370 ⇒ 00:21:51.990 Uttam Kumaran: ….
211 00:21:51.990 ⇒ 00:21:52.810 Ryan DeForest: 80s.
212 00:21:52.810 ⇒ 00:21:56.109 Uttam Kumaran: We can basically start to poke at the things that.
213 00:21:56.110 ⇒ 00:21:56.550 Ryan DeForest: I didn’t see.
214 00:21:56.550 ⇒ 00:21:59.759 Uttam Kumaran: I mean, they were a little bit suspect on, but I think the accuracy will be pretty high.
215 00:21:59.760 ⇒ 00:22:05.120 Ryan DeForest: the combo, and it scatters pretty high. Yep.
216 00:22:05.320 ⇒ 00:22:09.279 Ryan DeForest: So, how many do you want me to send you? I could send it over Slack.
217 00:22:09.490 ⇒ 00:22:10.430 Ryan DeForest: That’s it.
218 00:22:10.430 ⇒ 00:22:12.780 Uttam Kumaran: I think 25 to 50 would be great.
219 00:22:20.650 ⇒ 00:22:23.500 Uttam Kumaran: We can’t hear… we can’t hear anyone.
220 00:22:24.170 ⇒ 00:22:30.480 Henry Zhao: I mean, because you were on mute, so… Oh, I put that one up. You said….
221 00:22:31.040 ⇒ 00:22:32.670 Uttam Kumaran: Like, 25 to 50.
222 00:22:35.210 ⇒ 00:22:37.659 Henry Zhao: Oh, 25, 50? Okay, gotcha, okay.
223 00:22:38.550 ⇒ 00:22:40.729 Uttam Kumaran: If that’s too much, we could do less, but…
224 00:22:41.400 ⇒ 00:22:50.390 Uttam Kumaran: I want to, one… for me, the big tests are, like, I want to see it go end-to-end on a volume that high. I want to also start to measure
225 00:22:50.510 ⇒ 00:22:53.600 Uttam Kumaran: The cost and the duration for each run.
226 00:22:54.590 ⇒ 00:23:11.749 Henry Zhao: I have… I have a list that I’ve been working on on the side. I’ll send you. It’s only the names of the company, so you’ll need to do some magic with the domains. Okay, okay. But I’ll just send you a list of, like, 50… I’ll send you a list of 50 company names, and I’ll let you go from there.
227 00:23:11.750 ⇒ 00:23:20.280 Uttam Kumaran: small enough for us to also poke at it manually and just double check, but also, I want to start to get, like, some sample size, so….
228 00:23:20.930 ⇒ 00:23:32.699 Henry Zhao: Yeah, and all of these I’ll send you, I just went through this list yesterday, they all have, like, a demo request form. I don’t know what it is, or what happens after you submit, or anything like that, I just know that there’s a button that says, like.
229 00:23:32.950 ⇒ 00:23:36.039 Henry Zhao: Request demo, or contact sales, and stuff like that.
230 00:23:36.370 ⇒ 00:23:37.020 Uttam Kumaran: Okay.
231 00:23:37.180 ⇒ 00:23:43.099 Uttam Kumaran: Yeah, and then if you guys think about anything that’s helpful, I mean, we are having this system go onto their website.
232 00:23:43.260 ⇒ 00:23:44.759 Uttam Kumaran: And go through the form.
233 00:23:45.200 ⇒ 00:23:49.440 Uttam Kumaran: I have a feeling there’s probably, like, one or two other things that we could get.
234 00:23:49.670 ⇒ 00:23:56.049 Uttam Kumaran: like, I don’t know, I don’t know if a screenshot is helpful, or, like, what else is helpful in the sales process.
235 00:23:56.360 ⇒ 00:24:04.169 Uttam Kumaran: Or, like, maybe we… we scrape what the demo CTA is, or… I don’t know, but, like.
236 00:24:06.590 ⇒ 00:24:09.660 Uttam Kumaran: Yeah, I don’t know, I haven’t thought about it enough.
237 00:24:12.430 ⇒ 00:24:13.020 Henry Zhao: Okay.
238 00:24:13.230 ⇒ 00:24:20.110 Henry Zhao: Screenshots is a good one, like, because, for example, if you go to a website and you put in your…
239 00:24:20.260 ⇒ 00:24:28.100 Henry Zhao: data, and it says, we’ll contact you shortly, right? And no calendar pops up? Like, that’s a huge, giant red flag. Okay.
240 00:24:28.380 ⇒ 00:24:29.400 Henry Zhao: Target.
241 00:24:30.750 ⇒ 00:24:36.359 Uttam Kumaran: So maybe what could be helpful is screenshots as it fills each position.
242 00:24:36.700 ⇒ 00:24:43.350 Uttam Kumaran: That way, you have that. While you talk to the… before you hop on with the prospect, or if you’re emailing them, you have
243 00:24:43.480 ⇒ 00:24:49.020 Uttam Kumaran: You basically have, like, what filling out their form looks like, you know.
244 00:24:49.020 ⇒ 00:24:58.649 Henry Zhao: Well, because, like, I’m even, like, thinking about crazy ideas, like, it could be, like, a super cold outbound email that says, like, this page is dropping your conversion rates by 50%.
245 00:24:58.770 ⇒ 00:25:09.829 Henry Zhao: Yeah. Something like that. And that’s super automated based off that screenshot, so, like, something very simple like that would be pretty helpful, so let me know, kind of game plan when we do the better list of stuff.
246 00:25:10.360 ⇒ 00:25:13.499 Uttam Kumaran: You know, what you’re seeing a lot in, like, automated,
247 00:25:13.620 ⇒ 00:25:21.300 Uttam Kumaran: like, a lot of people for outbound right now are doing these, like, they’re using Bitplayer stuff to, like, do custom video outbounds, where they use AI.
248 00:25:21.470 ⇒ 00:25:29.579 Uttam Kumaran: And they’re like, hi, Sendoso, and the Sendoso part is, like, AI-generated. In the background, also, they have an AI that goes to the site and, like.
249 00:25:29.750 ⇒ 00:25:35.249 Uttam Kumaran: scrolls, so it looks, like, very, very personal. I get hit by this stuff all the time.
250 00:25:35.430 ⇒ 00:25:38.890 Uttam Kumaran: But this is pro- probably don’t want to Yeah.
251 00:25:40.380 ⇒ 00:25:53.460 Henry Zhao: Well, because what I… what… like, a crazy idea we were thinking about literally yesterday was being able to take their form, vibecode a copy of it, and then in the demo that’s coming up, we literally, like, show their form and, like, go from there.
252 00:25:53.460 ⇒ 00:25:54.510 Uttam Kumaran: No, I saw the….
253 00:25:54.510 ⇒ 00:25:55.350 Henry Zhao: We’re gonna be like.
254 00:25:55.350 ⇒ 00:26:04.510 Uttam Kumaran: I saw it on LinkedIn from… yeah, I forgot who… I forgot his name on our team. I saw that he posted that. This is, like, a shoe-in to that, as well.
255 00:26:04.870 ⇒ 00:26:08.540 Uttam Kumaran: Where maybe it is, like, hey, we generated you a version of it.
256 00:26:08.720 ⇒ 00:26:25.649 Uttam Kumaran: And that’s, like, an asset that gets added. Again, like, whatever the angle is, for me, I think it’s important just for the salesperson to have multiple angles, whether it’s, like, more of a… they want to just reach out through LinkedIn, and they just have the details, or it’s, like, super cold and they want to send something.
257 00:26:25.780 ⇒ 00:26:30.460 Uttam Kumaran: … Yeah.
258 00:26:31.770 ⇒ 00:26:42.519 Henry Zhao: Yeah, I’m on the same. So, yeah, let me and Lev kind of noodle on the idea of what else we can have, but, like, screenshots, for example, is, like, a top one that I think would be good. There might be other stuff that we can, like, look for.
259 00:26:42.830 ⇒ 00:26:43.690 Uttam Kumaran: Okay, okay.
260 00:26:44.080 ⇒ 00:26:44.989 Uttam Kumaran: Okay, cool.
261 00:26:45.490 ⇒ 00:26:52.100 Uttam Kumaran: So we’ll get that going for that list. We’ll also run that list that you sent through the other signal combo.
262 00:26:52.310 ⇒ 00:26:53.890 Uttam Kumaran: The one with the jobs?
263 00:26:54.140 ⇒ 00:26:56.390 Uttam Kumaran: And then we can verify that, and then…
264 00:26:56.930 ⇒ 00:27:01.119 Uttam Kumaran: I think that’ll be… I mean, we could… if you send that over, we can do that today.
265 00:27:01.650 ⇒ 00:27:04.349 Henry Zhao: Okay, sweet. I’ll send it over probably in the next hour or so.
266 00:27:04.350 ⇒ 00:27:05.000 Uttam Kumaran: Okay.
267 00:27:06.590 ⇒ 00:27:17.480 Uttam Kumaran: Cool, and then, yeah, Henry’s… Henry’s… Henry’s there. I don’t know if he mentioned a lot of the work we’re doing on the product analytics side with Justin. I don’t know if Justin’s in, but Henry, you should poke him, but help us with that, but …
268 00:27:17.520 ⇒ 00:27:27.350 Uttam Kumaran: Yeah, we’re… we’re starting… we’re gonna be measuring that. I think, Henry, what could be good is also to understand a little bit about what the PLG flow for the next product is going to look like.
269 00:27:28.110 ⇒ 00:27:29.120 Uttam Kumaran: ….
270 00:27:32.420 ⇒ 00:27:36.620 Henry Zhao: Sorry, Wi-Fi’s a little shitty over here, but yeah.
271 00:27:36.620 ⇒ 00:27:38.050 Uttam Kumaran: Oh, okay, okay.
272 00:27:39.230 ⇒ 00:27:46.660 Uttam Kumaran: I said my question was going to be what the… for you to start to talk to people about what the PLG blow is going to look like in the new product.
273 00:27:47.870 ⇒ 00:27:52.320 Henry Zhao: Yeah, that’s a good point. Yeah, yeah, yeah. …
274 00:27:52.960 ⇒ 00:27:56.329 Henry Zhao: No one’s here yet, so… we’ll see.
275 00:27:57.920 ⇒ 00:27:59.479 Uttam Kumaran: Yeah, yeah, it’s so early there, yeah.
276 00:27:59.480 ⇒ 00:28:01.239 Henry Zhao: And I’ll see the demons right now.
277 00:28:01.840 ⇒ 00:28:03.110 Henry Zhao: that you….
278 00:28:04.500 ⇒ 00:28:05.640 Uttam Kumaran: Yeah.
279 00:28:05.640 ⇒ 00:28:09.650 Henry Zhao: Sure, for sure. Great work. No, it looks fantastic. I love it.
280 00:28:09.890 ⇒ 00:28:12.850 Uttam Kumaran: Okay, alright. Thank you guys. I’ll chat over Slack.
281 00:28:13.360 ⇒ 00:28:15.610 Henry Zhao: Perfect. Thanks a lot. Thank you, guys.