Meeting Title: Brainforge Office Hours Date: 2026-03-05 Meeting participants: Hannah Wang, Pranav Narahari, Luke Scorziell, Brandon Ponzo, Leo Moronta, Donovan Griffin, Kayla Klein, Kayla’s Notetaker, Alice Chiang, Troy Sandidge | @FindTroy
WEBVTT
1 00:00:48.470 ⇒ 00:00:49.660 Pranav Narahari: I’ll be right back.
2 00:00:50.530 ⇒ 00:00:51.130 Hannah Wang: Sure.
3 00:02:00.690 ⇒ 00:02:04.500 Luke Scorziell: Okay, I did put on a slightly nicer… .
4 00:02:04.500 ⇒ 00:02:07.129 Hannah Wang: Oh, nice. Quick outfit change.
5 00:02:08.430 ⇒ 00:02:09.600 Luke Scorziell: Yeah.
6 00:02:10.380 ⇒ 00:02:12.700 Luke Scorziell: Makes me feel more confident.
7 00:02:12.970 ⇒ 00:02:13.800 Hannah Wang: Oh, that’s…
8 00:02:18.590 ⇒ 00:02:26.200 Hannah Wang: Oh, I haven’t done one of these in a while, I’m nervous, even though I’m not really doing much of it.
9 00:02:26.200 ⇒ 00:02:31.409 Luke Scorziell: Okay, we will… we will do this, and see how it goes, and then…
10 00:02:33.660 ⇒ 00:02:36.279 Luke Scorziell: That’s… yeah, however it goes, we’ll be fine.
11 00:02:36.780 ⇒ 00:02:38.830 Hannah Wang: Oh, yeah, I haven’t seen you in so long.
12 00:02:39.930 ⇒ 00:02:43.370 Hannah Wang: I don’t know, I… need to emerge from my…
13 00:02:45.140 ⇒ 00:02:55.030 Hannah Wang: man cave. Being stuck… stuck at home. Yeah, turning on the camera is tiring, so I usually don’t do it, but…
14 00:02:56.040 ⇒ 00:02:57.250 Luke Scorziell: Gotta do it.
15 00:02:57.250 ⇒ 00:02:59.619 Hannah Wang: Gotta do it for today.
16 00:03:00.840 ⇒ 00:03:04.660 Hannah Wang: How do I pop this out?
17 00:03:08.610 ⇒ 00:03:10.640 Pranav Narahari: Should I start sharing in a little bit?
18 00:03:11.490 ⇒ 00:03:15.100 Hannah Wang: You can… keep it there. Okay. Yeah.
19 00:03:18.600 ⇒ 00:03:21.109 Luke Scorziell: Got another last-minute registration.
20 00:03:22.240 ⇒ 00:03:23.360 Hannah Wang: Oh, nice.
21 00:03:25.100 ⇒ 00:03:28.160 Pranav Narahari: So it’ll be interesting to see if everyone trickles in exactly the same time.
22 00:03:30.200 ⇒ 00:03:31.070 Pranav Narahari: Right.
23 00:03:31.230 ⇒ 00:03:31.870 Luke Scorziell: Yeah.
24 00:03:31.870 ⇒ 00:03:36.059 Hannah Wang: Like, for these, I usually hop on, like, a minute later, because I don’t want to be the first one.
25 00:03:36.900 ⇒ 00:03:39.469 Hannah Wang: We’ll see, we’ll see how everyone else does.
26 00:03:39.880 ⇒ 00:03:40.690 Pranav Narahari: Yeah.
27 00:03:43.590 ⇒ 00:03:45.460 Pranav Narahari: So do we want to, like, wait 2 minutes?
28 00:03:46.100 ⇒ 00:03:49.089 Luke Scorziell: Yeah, we could… we don’t have to, like, right off the bat.
29 00:03:49.640 ⇒ 00:03:50.280 Pranav Narahari: Cool.
30 00:03:50.730 ⇒ 00:03:52.100 Luke Scorziell: Hey, Brennan, how’s it going?
31 00:03:53.740 ⇒ 00:03:54.890 Brandon Ponzo: Hey, how’s it going?
32 00:03:55.690 ⇒ 00:03:57.030 Luke Scorziell: Good, good to meet you.
33 00:04:08.480 ⇒ 00:04:11.380 Pranav Narahari: Cool, we got people coming in early. Hey, Brandon.
34 00:04:27.980 ⇒ 00:04:31.450 Pranav Narahari: Just curious, where are you, where are you calling in from?
35 00:04:34.210 ⇒ 00:04:35.450 Brandon Ponzo: New York City.
36 00:04:35.980 ⇒ 00:04:37.900 Pranav Narahari: Okay, nice. Not too far from me.
37 00:04:39.460 ⇒ 00:04:40.309 Brandon Ponzo: Where are you at?
38 00:04:40.770 ⇒ 00:04:44.520 Pranav Narahari: I’m in, Massachusetts, Worcester, if you’ve heard of it.
39 00:04:44.900 ⇒ 00:04:46.750 Brandon Ponzo: Yep, yep, I definitely have.
40 00:04:46.960 ⇒ 00:04:50.759 Pranav Narahari: Oh, okay, cool, cool. Hopefully, not too bad things.
41 00:04:51.890 ⇒ 00:04:53.570 Brandon Ponzo: Not just the name.
42 00:04:53.570 ⇒ 00:04:55.050 Pranav Narahari: Okay, cool, cool, cool.
43 00:05:01.190 ⇒ 00:05:01.920 Luke Scorziell: Wait.
44 00:05:42.050 ⇒ 00:05:43.419 Luke Scorziell: Yo, what’s up?
45 00:05:46.830 ⇒ 00:05:47.720 Leo Moronta: Hey!
46 00:05:48.220 ⇒ 00:05:49.429 Luke Scorziell: How’s it going?
47 00:05:50.300 ⇒ 00:05:52.119 Leo Moronta: Living the dream, how are you doing?
48 00:05:52.360 ⇒ 00:05:53.890 Luke Scorziell: Dude, you’re always living the dream.
49 00:05:56.570 ⇒ 00:06:03.370 Luke Scorziell: Good, yeah, excited for this. So this is, Pranav and Hannah. Leah’s one of my…
50 00:06:03.910 ⇒ 00:06:07.439 Luke Scorziell: A very talented marketer that I can learn a lot from.
51 00:06:07.870 ⇒ 00:06:09.600 Pranav Narahari: Awesome. Nice to meet you.
52 00:06:09.990 ⇒ 00:06:11.170 Leo Moronta: Nice to meet you guys as well.
53 00:06:11.900 ⇒ 00:06:14.130 Hannah Wang: How’d you and Luke meet?
54 00:06:16.120 ⇒ 00:06:19.649 Leo Moronta: Through a mutual… mutual friend,
55 00:06:19.840 ⇒ 00:06:27.650 Leo Moronta: Connecting people vaguely interested in marketing and like to do cool things, and it’s been, you know, attached at the hip.
56 00:06:31.020 ⇒ 00:06:32.950 Leo Moronta: I think Donovan’s gonna be here, actually.
57 00:06:33.310 ⇒ 00:06:35.319 Luke Scorziell: Wow, that’s beautiful.
58 00:06:40.410 ⇒ 00:06:43.250 Luke Scorziell: Alright, 11.02 is the… the moment.
59 00:06:43.480 ⇒ 00:06:44.699 Luke Scorziell: Where everyone comes.
60 00:06:45.930 ⇒ 00:06:47.889 Luke Scorziell: Hey, Kayla, how’s it going, Donovan?
61 00:06:48.210 ⇒ 00:06:49.230 Donovan Griffin: Hey, Luke.
62 00:06:50.750 ⇒ 00:06:51.789 Kayla Klein: Hi, who are you?
63 00:06:52.430 ⇒ 00:06:52.940 Pranav Narahari: Hey, guys.
64 00:06:52.940 ⇒ 00:06:53.740 Luke Scorziell: Whoa.
65 00:06:55.440 ⇒ 00:06:58.590 Luke Scorziell: Hopefully you guys all feel somewhat comfortable with,
66 00:06:59.030 ⇒ 00:07:02.169 Luke Scorziell: Me just calling on you throughout the meeting.
67 00:07:02.240 ⇒ 00:07:03.670 Donovan Griffin: Heck yeah.
68 00:07:04.160 ⇒ 00:07:09.259 Luke Scorziell: Our goal is, yeah, to have, I guess, a little more of a discussion than us just speaking at you.
69 00:07:13.070 ⇒ 00:07:17.330 Luke Scorziell: Yeah, we’ve got a couple more people, and we can… Kick it off.
70 00:07:19.050 ⇒ 00:07:23.329 Hannah Wang: Yeah, we’ll wait maybe 2 minutes or so.
71 00:07:32.590 ⇒ 00:07:34.150 Luke Scorziell: Kayla, where are you calling in from?
72 00:07:35.110 ⇒ 00:07:37.130 Kayla Klein: I am in Boise, Idaho.
73 00:07:37.930 ⇒ 00:07:38.650 Luke Scorziell: Nice.
74 00:07:38.650 ⇒ 00:07:41.689 Kayla Klein: Doesn’t look like it because of my background, but…
75 00:07:41.730 ⇒ 00:07:42.950 Luke Scorziell: Quite slowly.
76 00:07:45.010 ⇒ 00:07:47.670 Luke Scorziell: I was gonna say, are you near Kelly at all?
77 00:07:48.470 ⇒ 00:07:52.459 Kayla Klein: Yeah, Kelly’s in Jackson Hole, Wyoming, which is, like, a 5-hour drive.
78 00:07:53.310 ⇒ 00:07:54.030 Luke Scorziell: Okay.
79 00:07:54.150 ⇒ 00:07:54.980 Luke Scorziell: Nice. Yeah.
80 00:07:56.370 ⇒ 00:07:58.289 Kayla Klein: I’m joining on her behalf today.
81 00:07:59.560 ⇒ 00:08:00.639 Luke Scorziell: Happy to have you.
82 00:08:06.750 ⇒ 00:08:08.950 Luke Scorziell: I think Alice just joined.
83 00:08:13.360 ⇒ 00:08:14.540 Luke Scorziell: Hey, Alice.
84 00:08:14.690 ⇒ 00:08:18.060 Alice Chiang: Hello, there is music in the background.
85 00:08:18.470 ⇒ 00:08:21.860 Alice Chiang: I don’t know if this is too noisy. I can just mute myself.
86 00:08:21.930 ⇒ 00:08:27.969 Luke Scorziell: That’s okay. Whatever is… yeah, whatever you’re comfortable with. But yeah, probably muting yourself is okay.
87 00:08:27.970 ⇒ 00:08:31.120 Alice Chiang: Yeah, yeah. When I talk, is this okay, or…
88 00:08:31.410 ⇒ 00:08:32.320 Luke Scorziell: Yeah, that sounds great.
89 00:08:32.669 ⇒ 00:08:33.659 Alice Chiang: Okay, awesome.
90 00:08:35.490 ⇒ 00:08:37.370 Luke Scorziell: Troy, how are you? Good to see you.
91 00:08:37.690 ⇒ 00:08:44.359 Troy Sandidge | @FindTroy: See you, Luke, everybody, don’t mind me just holding a baby in the background right now. It’s all good.
92 00:08:46.370 ⇒ 00:08:52.389 Luke Scorziell: Well, I don’t… yeah, Hannah, if you want to kick us off, we can get… get started, and yeah, I think we’re…
93 00:08:52.390 ⇒ 00:08:53.330 Hannah Wang: Yeah.
94 00:08:53.330 ⇒ 00:08:54.609 Luke Scorziell: Mostly ever listen.
95 00:08:55.320 ⇒ 00:08:55.950 Hannah Wang: Sure.
96 00:08:56.040 ⇒ 00:09:15.460 Hannah Wang: So welcome, everyone, to, I guess, this Office Hours series. This is our first one, so we’re pretty pumped to get it up and running, so appreciate you all joining. And if you go to the next slide, Pranav, I’ll kind of jump into the quick
97 00:09:15.660 ⇒ 00:09:18.789 Hannah Wang: A quick overview of the agenda, so we’ll just do…
98 00:09:19.130 ⇒ 00:09:34.090 Hannah Wang: we’ll just introduce ourselves, and then, like, an overview of Brainforge, and then we’ll dive into kind of a demo, and then at the end we’ll have a discussion. I encourage you all… well, a lot of you have your videos on, so thank you, and…
99 00:09:34.220 ⇒ 00:09:36.529 Hannah Wang: Y’all are good at talking, I feel like.
100 00:09:36.810 ⇒ 00:09:42.909 Hannah Wang: yeah, just a fair warning, Luke might call on you to ask you questions and to probe deeper into your
101 00:09:43.270 ⇒ 00:09:51.980 Hannah Wang: to your AI problems or questions, so, just be on the lookout for that. And… yeah.
102 00:09:52.060 ⇒ 00:10:11.970 Hannah Wang: introducing the BrainForge team. I’m Hannah, I’ll kind of be introing this whole thing, and just kind of managing and moderating. I work as the creative and engagement lead, so I lead design here at Brainforge, and I kind of help with the go-to-market team. And then Luke and Pranav, if you want to give a quick intro, that’d be great.
103 00:10:13.310 ⇒ 00:10:15.500 Luke Scorziell: Yeah, Pranav, do you wanna… you wanna go first?
104 00:10:15.500 ⇒ 00:10:19.189 Pranav Narahari: Yeah, I can go. So yeah, I,
105 00:10:19.210 ⇒ 00:10:32.020 Pranav Narahari: told some people in the beginning of the call, I’m based out of, Massachusetts right now. Here at Brainforge, I’m, part of the delivery team, so I’m running kind of like these AI projects, both
106 00:10:32.030 ⇒ 00:10:44.989 Pranav Narahari: the development side of things, so the more technical, but then also with our, like, unique structure at, like, Brainforge, too, kind of, like, interacting with clients, figuring out what their, like, business problems are, and, like, then helping design the AI solutions behind that.
107 00:10:47.970 ⇒ 00:11:01.469 Luke Scorziell: Sweet. Very modest intros from both Hannah and Vanav. They’re very, very talented. And yeah, I’m Luke, obviously, and I lead the marketing team and go-to-market at Brainforge, so…
108 00:11:01.550 ⇒ 00:11:09.720 Luke Scorziell: Yeah, it’s been a… I did not necessarily come from the technical AI background, although I like playing around with different tools.
109 00:11:09.960 ⇒ 00:11:13.429 Luke Scorziell: But yeah, I have more of, journalism education.
110 00:11:13.570 ⇒ 00:11:20.690 Luke Scorziell: in school, and then was also working in the agency space for the last, I guess, like, year and a half.
111 00:11:20.780 ⇒ 00:11:33.689 Luke Scorziell: And have since joined… joining Brainforge, learned so many crazy things that I wish I’d known, while I was doing my own thing. And, so hopefully, yeah, we can impart some of that
112 00:11:33.790 ⇒ 00:11:41.780 Luke Scorziell: that knowledge to y’all, on this call. But, yeah, super excited, and yeah, really excited to kick this off.
113 00:11:43.130 ⇒ 00:11:46.229 Hannah Wang: Cool. We would love to…
114 00:11:46.370 ⇒ 00:11:51.260 Hannah Wang: get to know everyone else, but for the sake of time, I think I’m gonna shrink it down to…
115 00:11:51.680 ⇒ 00:12:08.270 Hannah Wang: just these three things in the chat, so if you want to, and feel free to put this information in there. But yeah, I would love to know your name, what you do, slash what company, you work at. And then, I guess a fun question is,
116 00:12:08.660 ⇒ 00:12:18.959 Hannah Wang: If you had to be, which AI would you be? So, ChatGPT, or Claude, or Cursor, or all the other buzzwords that you can think of, and if…
117 00:12:18.960 ⇒ 00:12:37.269 Hannah Wang: none of those ring a bell, then that is okay, and that is why you’re here. So feel free to drop that in the chat at any point, but I think we can kind of go ahead and get started. I’ll, kick it over to Pranav to talk about a little bit more about Brainforge and the clients that we helped in the past.
118 00:12:39.180 ⇒ 00:12:41.929 Pranav Narahari: Cool, thanks. I’m just reading a couple of the…
119 00:12:42.510 ⇒ 00:12:44.489 Pranav Narahari: the message that I came in so far.
120 00:12:45.090 ⇒ 00:12:47.280 Pranav Narahari: Okay, two pods. Nice.
121 00:12:48.170 ⇒ 00:12:51.499 Pranav Narahari: But yeah, let me…
122 00:12:52.110 ⇒ 00:12:58.020 Pranav Narahari: kind of just do a quick overview of, just Brainforge in general. Just super brief.
123 00:13:00.750 ⇒ 00:13:16.309 Pranav Narahari: Yeah, I just kind of want to talk about, like… I kind of already intro’d, like, what I do here at Brainforge, but there’s a lot of similar people just like me at Brainforge that have that similar curiosity. Since you guys are here, you probably also have a little bit of the same curiosity about just, like.
124 00:13:16.650 ⇒ 00:13:31.980 Pranav Narahari: the kind of world we live in today, which is… AI is just basically around every corner, you know, when you’re… when you just, in every conversation, a lot of people are interested in, like, this, but, just a couple years ago, I was just kind of…
125 00:13:33.120 ⇒ 00:13:47.479 Pranav Narahari: just jumping into this realm. So, it was really, like, when ChatGPT first came out, like, I had a friend that was, like, really in the weeds of, like, understanding, like, what an LLM was. I probably didn’t even know what that meant back then. And then so…
126 00:13:47.480 ⇒ 00:14:02.609 Pranav Narahari: I think that was, like, December 2022 is what I remember. But then after that, yeah, just trying to learn more about it, realizing that it can really, like, elevate the products that I could ship. And so that’s how it kind of became, like, part of my, like.
127 00:14:02.610 ⇒ 00:14:12.159 Pranav Narahari: development process, and then… yeah, a lot of people here at Brain Forge, like, have a similar story. I know Luke, too, like, just a few months ago, probably, he wasn’t using
128 00:14:12.160 ⇒ 00:14:24.890 Pranav Narahari: the same level of tooling that he’s using here today, and that goes across just, like, not just the engineers like me, but, like, Luke for, you know, his sales stuff, for everything he does at Brainforge, like, he finds a use for AI,
129 00:14:25.530 ⇒ 00:14:26.080 Pranav Narahari: But…
130 00:14:26.080 ⇒ 00:14:31.179 Luke Scorziell: I started using GitHub. I was like, never in my wildest dreams thought I’d be using GitHub.
131 00:14:31.590 ⇒ 00:14:42.380 Pranav Narahari: Yeah, yeah, and I think that’s kind of, like, unheard of, like, just, like, 6 months ago, right? Not a lot of people are going to be touching GitHub that aren’t, like, working on code.
132 00:14:42.520 ⇒ 00:14:46.780 Pranav Narahari: And then, yeah, I kind of also want to go into a little bit of, like, the…
133 00:14:46.980 ⇒ 00:15:04.080 Pranav Narahari: it’s a very high level of, like, some of the cool solutions we’ve built here at Brainforge. A couple that come to mind are just MCP servers, which are basically just a way to connect with your data, in real time, and chat with it in a… in a…
134 00:15:04.250 ⇒ 00:15:17.019 Pranav Narahari: chat interface. Also, just kind of, like, creating, like, briefs. It’s super great for that, we found, and we’ve had a lot of success with just building that for people.
135 00:15:17.360 ⇒ 00:15:29.420 Pranav Narahari: Building, like, forecasting models with just, like, pulling in that data, like, whether it is from MCP servers or, across any other type of integration. Real-time reporting.
136 00:15:29.500 ⇒ 00:15:42.869 Pranav Narahari: things of that nature. Also, having, like, a dynamic document that’s supposed to be, like, a source of truth, basically, for your organization, maybe it’s, like, a… for some type of business service that you guys have, and…
137 00:15:43.000 ⇒ 00:15:44.139 Pranav Narahari: a cool…
138 00:15:44.400 ⇒ 00:16:02.619 Pranav Narahari: and that can obviously continue on, and get a lot more intricate as, like, okay, more dynamic documents, different types of data, not just text, maybe it’s, like, audio recordings, video recordings, Zoom calls, anything like that.
139 00:16:02.920 ⇒ 00:16:06.920 Pranav Narahari: And so, what that kind of, like, leads me to is, like.
140 00:16:07.170 ⇒ 00:16:22.790 Pranav Narahari: there’s, like, a solution for almost, like, every type of problem. There’s probably, like, things that you guys are realizing within your daily workflows that you’re like, how can AI fit there? And I think what we can kind of discuss, like, later on this meeting is just, like.
141 00:16:22.790 ⇒ 00:16:32.879 Pranav Narahari: there probably is a way to fit in AI there. I’m not gonna promise that it’s gonna solve the entire thing, I don’t know what your guys’ problems are, but more often than not, there is, like.
142 00:16:32.960 ⇒ 00:16:39.299 Pranav Narahari: a at least partial solution where AI can help with the efficiency or just the daily quality of life.
143 00:16:39.450 ⇒ 00:16:49.960 Pranav Narahari: And so, yeah, if there’s… I want to also, like, pause here if there’s, like, any questions before I pass it over to Luke. But if there’s no questions, we can… we can move on.
144 00:16:56.140 ⇒ 00:16:59.739 Pranav Narahari: Cool. Luke, you wanna dive into the demo?
145 00:17:00.080 ⇒ 00:17:05.490 Luke Scorziell: Yeah, with the obligatory, awkward silence.
146 00:17:05.960 ⇒ 00:17:08.219 Luke Scorziell: Those meetings before we move on.
147 00:17:08.440 ⇒ 00:17:16.060 Luke Scorziell: Okay, so, yeah, I’ve let Alice know briefly before this meeting, but I’m gonna kind of
148 00:17:16.420 ⇒ 00:17:22.430 Luke Scorziell: Put her on the spot a little bit, but just, yeah, as you all submitted, like, questions that you’re having, and…
149 00:17:22.579 ⇒ 00:17:35.560 Luke Scorziell: and whatnot, something that Alice brought up is that, yeah, at her agency, they’re doing a lot of daily monitoring, searching through a list of terms on Google manually,
150 00:17:35.560 ⇒ 00:17:36.190 Pranav Narahari: Sometimes…
151 00:17:36.190 ⇒ 00:17:37.739 Luke Scorziell: that they’d like to streamline.
152 00:17:37.890 ⇒ 00:17:45.180 Luke Scorziell: With AI. So, Alice, yeah, I’d love to hear maybe, like, quickly from you, if… Noise is not.
153 00:17:45.600 ⇒ 00:17:53.339 Luke Scorziell: Yeah. Of course! What does this look like? Yeah, that’s great. Maybe you can tell us a little bit more about this problem.
154 00:17:53.830 ⇒ 00:17:55.630 Alice Chiang: Yeah, of course, so…
155 00:17:55.780 ⇒ 00:18:07.459 Alice Chiang: what I’ve realized since starting, my PR agency, so I graduated in May, and I’m working right now, at an agency under Publicis Group, and…
156 00:18:07.890 ⇒ 00:18:23.649 Alice Chiang: Yeah, just, I think as an entry-level person, a lot of my time is dedicated, to monitoring, and what my leadership has really tried to push on is they want this to be streamlined so that I can
157 00:18:23.700 ⇒ 00:18:41.930 Alice Chiang: go in to do more executional and more creative work. And we found some leeway and some things, but nothing really that our agency as a whole has really agreed is, like, safe enough, but also just, like, as comprehensive as manual searching, so we haven’t switched.
158 00:18:41.930 ⇒ 00:18:44.749 Alice Chiang: But basically what we do is I…
159 00:18:44.750 ⇒ 00:19:03.130 Alice Chiang: have one client, and there’s around 30 terms related to my clients, so the category that it’s in, so specifically, I work for… I work for a baby product company, and so anything surrounding, like, pregnancy, postpartum, like, celebrity babies,
160 00:19:03.590 ⇒ 00:19:10.080 Alice Chiang: Like, any kind of products that may be similar to our product, all of our competitors,
161 00:19:10.190 ⇒ 00:19:13.460 Alice Chiang: Research for… for them, and see…
162 00:19:13.540 ⇒ 00:19:24.779 Alice Chiang: you know, filter for 24 hours on Google and see if there’s anything that is noteworthy, any celebrities we want to partner with or send through a gift package,
163 00:19:24.810 ⇒ 00:19:34.919 Alice Chiang: any noticeable, like, industry trends, whether it’s, like, you know, this specific baby product, consumers are not happy because it’s not…
164 00:19:35.060 ⇒ 00:19:51.399 Alice Chiang: you know, there’s a certain kind of chemical in it, and we have to take note of that as we, like, develop new products, etc. So, it is… it requires, I think, brand knowledge for a person to be able to know that, okay, this is…
165 00:19:51.470 ⇒ 00:20:04.699 Alice Chiang: important for me to flag to my wider team, and then to the clients, versus not. So I think where we’ve had the issue is just with the AI that we’ve, like, experimented with,
166 00:20:04.900 ⇒ 00:20:13.010 Alice Chiang: They don’t necessarily have, sort of, that brand knowledge, that makes it difficult for us to be willing to trust it.
167 00:20:13.120 ⇒ 00:20:19.150 Alice Chiang: I also think… Just in general, this is a problem…
168 00:20:19.290 ⇒ 00:20:21.939 Alice Chiang: that I think PR agency is…
169 00:20:22.860 ⇒ 00:20:28.399 Alice Chiang: I think, in terms of… for the experience I’ve had, too, it’s a little bit hard for…
170 00:20:28.680 ⇒ 00:20:38.579 Alice Chiang: us to be thinking about how to integrate AI, just because we have so much work, executional, like, client work.
171 00:20:38.680 ⇒ 00:20:41.019 Alice Chiang: that I’ve been wanting to, like.
172 00:20:41.250 ⇒ 00:20:53.480 Alice Chiang: look into AI, but there, frankly, just isn’t time in the day to do so. So it becomes something that has to be very top-down, but also, at the same time, leadership might not know, like.
173 00:20:53.680 ⇒ 00:21:03.769 Alice Chiang: what we’re actually doing, it might have changed since, you know, they were entry-level, so… yeah, that’s just two layers of the issue that I’ve been seeing.
174 00:21:05.440 ⇒ 00:21:15.220 Luke Scorziell: Yeah, no, thank you so much for sharing, that’s super, super helpful. And what, like, when you’re doing this, is it, like, every morning you’re going on, and, like, this is the first thing you’re doing, and, like, logging on?
175 00:21:15.340 ⇒ 00:21:18.160 Luke Scorziell: How long do you spend, would you say, like, going through these terms?
176 00:21:18.490 ⇒ 00:21:29.669 Alice Chiang: Mmm… I think, you know, when I first started, it was, like, 2 hours. Now I think it’s, like, 30 minutes to an hour, depending on,
177 00:21:29.930 ⇒ 00:21:45.370 Alice Chiang: whether there’s, like, a launch. There was one week a few months ago, all of our competitors launched something that same week, so that was, like, a lot, because then I would have to, like, open up every single article and see, like.
178 00:21:46.460 ⇒ 00:21:52.679 Alice Chiang: whether there is something noticeably different in them, and if there’s something that I want to flag to clients.
179 00:21:54.140 ⇒ 00:21:55.080 Luke Scorziell: Yeah.
180 00:21:55.210 ⇒ 00:21:56.260 Luke Scorziell: Well, click.
181 00:21:56.510 ⇒ 00:21:59.370 Luke Scorziell: I’ll show the demo in a second.
182 00:21:59.700 ⇒ 00:22:01.929 Luke Scorziell: But… oh, can you guys start my soul?
183 00:22:04.190 ⇒ 00:22:04.710 Pranav Narahari: And you’re back.
184 00:22:04.710 ⇒ 00:22:05.090 Luke Scorziell: Come back.
185 00:22:05.090 ⇒ 00:22:05.550 Pranav Narahari: Yeah, yeah.
186 00:22:06.600 ⇒ 00:22:08.210 Luke Scorziell: Yeah,
187 00:22:08.710 ⇒ 00:22:24.580 Luke Scorziell: But yeah, a couple themes from that that I think we can… we see in, like, every agency, and every, like, company that we work with is, like, company knowledge and brand knowledge, and that’s something that with, like, ChatGPT, Copilot, Gemini, like, all these large model LLMs, like, you don’t get because they don’t have access to
188 00:22:24.850 ⇒ 00:22:41.099 Luke Scorziell: like, the client call that you had yesterday, or the client calls that you’ve had for the last year with a certain client. They don’t have, like, your brand guidelines, or each client… client’s brand guidelines in them. And so, that’s… yeah, that’s a super huge issue. And then, similarly, what,
189 00:22:41.960 ⇒ 00:22:50.459 Luke Scorziell: Is getting the most leverage out of the people that you have working, and allowing them to spend more time on high-leverage tasks instead of
190 00:22:50.590 ⇒ 00:22:59.370 Luke Scorziell: things like information collection, manual data entry, like, all those things. And I think, like, AI has kind of had this, like, hype phase, where it’s been very, like.
191 00:22:59.420 ⇒ 00:23:14.829 Luke Scorziell: oh my gosh, like, I can make a… a kitten that’s, like, walking on two feet, and now it’s, like, in a video with Tom Cruise, you know, some random stuff like that. But where we’re seeing the most return on investment with AI is in these, like, very…
192 00:23:14.950 ⇒ 00:23:20.319 Luke Scorziell: kind of… repetitive operational tasks.
193 00:23:21.740 ⇒ 00:23:25.639 Luke Scorziell: Like, document transfer and informa- information transfer.
194 00:23:26.350 ⇒ 00:23:31.080 Luke Scorziell: That’s true, so… Okay, I will share my screen.
195 00:23:32.940 ⇒ 00:23:35.780 Luke Scorziell: I’m trying to whip up some AI magic, guys.
196 00:23:35.920 ⇒ 00:23:37.180 Luke Scorziell: as we speak.
197 00:23:37.970 ⇒ 00:23:43.130 Luke Scorziell: Excite, you’ll have to give me… Quick second.
198 00:23:44.400 ⇒ 00:23:46.699 Luke Scorziell: Let me move as fast as this cursor.
199 00:23:46.920 ⇒ 00:23:52.150 Luke Scorziell: But… Let’s see…
200 00:24:02.580 ⇒ 00:24:12.140 Luke Scorziell: And have the solutions, I guess while we’re all waiting a second, have they been kind of external, like, third-party solutions where you’re going to a chatbot, or what have you guys tried to build so far?
201 00:24:13.170 ⇒ 00:24:30.399 Alice Chiang: Yeah, so we’ve tried, like, Google Alerts, I can drop in the chat just a few things that our, like, analytics team have suggested, but even they were like, none of these we think can replace. They just don’t think it’s good enough.
202 00:24:31.170 ⇒ 00:24:40.750 Alice Chiang: So yeah, and like, maybe just another caveat, too, is that we, like, in addition to the traditional, like, earned media monitoring, we also do, like.
203 00:24:41.050 ⇒ 00:24:48.809 Alice Chiang: monitoring on all of our KOL content, so anyone we contract, I’m going into their content every day to see, like.
204 00:24:49.040 ⇒ 00:24:59.949 Alice Chiang: comments, only when there’s an issue with one of our products, and we think there might be negative feedback. But right now, one of our product lines is just having issues, so…
205 00:25:00.380 ⇒ 00:25:08.629 Alice Chiang: anyone that’s contracted under that product, I’m, like, going in to see their comments every day, which is oh-so quite time-consuming.
206 00:25:09.660 ⇒ 00:25:10.270 Luke Scorziell: Yeah.
207 00:25:12.010 ⇒ 00:25:19.659 Luke Scorziell: Cool. Well, here’s the demo. I had some stuff built out, but I just made a few quick adjustments, too.
208 00:25:20.420 ⇒ 00:25:28.089 Luke Scorziell: make it more relevant. But yeah, the first thing you can kind of see here is that it’s… we like to build
209 00:25:28.240 ⇒ 00:25:39.890 Luke Scorziell: where our clients are, and so, in this case, like, I don’t know if you guys use Slack, use Teams, whatever it is, like, this is, like, an interface that’s, you know, designed to look a bit like Slack. And so…
210 00:25:40.760 ⇒ 00:25:41.640 Luke Scorziell: what?
211 00:25:41.760 ⇒ 00:25:47.790 Luke Scorziell: That’s kind of most important, because what we’ve found, too, with a lot of our clients and people that we’re talking with is that
212 00:25:48.030 ⇒ 00:25:52.909 Luke Scorziell: Moving to, like, this external tool, or trying to add something new into your workflow is, like.
213 00:25:53.080 ⇒ 00:25:59.649 Luke Scorziell: Can be, normally, people use it once and then ditch it if it’s not quicker.
214 00:25:59.830 ⇒ 00:26:14.810 Luke Scorziell: So, yeah, I mean, I don’t know, Alice, how helpful, like, this will be, or hopefully it gives you some ideas, and I guess for everyone else, too, like, these are the types of things that we can do for… for all businesses, and then also that, like, you can start to learn how to do for yourself, too, if you’re,
215 00:26:14.910 ⇒ 00:26:19.869 Luke Scorziell: And that. And so, yeah, what this is, basically, is, like, Well, then Slack.
216 00:26:20.090 ⇒ 00:26:25.640 Luke Scorziell: This did not say Bloom Baby Co. a minute ago, but let’s say this is, like, your fake client.
217 00:26:25.870 ⇒ 00:26:39.410 Luke Scorziell: And this is a daily monitor, so you have different channels where you’re monitoring, like, competitors, and what news is coming on… going there, like celebrities, industry trends, pregnancy, postpartums.
218 00:26:39.520 ⇒ 00:26:49.500 Luke Scorziell: Or yeah, just kind of stuff that’s going on. And what you can do there is, like, every morning at 8 AM, the app just messages you, like.
219 00:26:49.710 ⇒ 00:26:52.329 Luke Scorziell: Here is the latest on Google.
220 00:26:52.660 ⇒ 00:26:54.940 Luke Scorziell: And then you can click in, see the source.
221 00:26:55.140 ⇒ 00:27:01.249 Luke Scorziell: You can have a knowledge base built in, too, so that it knows, like, the brand guidelines and, like, specifically what they’d be looking for.
222 00:27:01.490 ⇒ 00:27:13.070 Luke Scorziell: And so, yeah, like, this… this is the type of solution that we’ve built for a lot of clients, where, like, if they’re doing, like, ads reporting, like, I know Leah, they’re doing a lot of that, too, where you have a channel that every morning
223 00:27:13.390 ⇒ 00:27:29.990 Luke Scorziell: like, the meta-add ROAS, like, comes in, or, like, the… whatever metrics are most important for you and that client to measure, like, automatically are updating, every day in certain channels. And so, operationally, like, this is something that we’ve found to be pretty…
224 00:27:30.430 ⇒ 00:27:32.610 Luke Scorziell: Effective, so what you can do is, like.
225 00:27:32.930 ⇒ 00:27:39.660 Luke Scorziell: In Slack, you just do, like, forward slash monitor, and then just click on one of these.
226 00:27:40.070 ⇒ 00:27:45.560 Luke Scorziell: And then, I mean, it just doesn’t give that much info, it’s just a quick demo, but you kind of get the…
227 00:27:45.770 ⇒ 00:27:49.199 Luke Scorziell: At the point where you could kind of ask it to do things.
228 00:27:49.530 ⇒ 00:27:55.649 Luke Scorziell: for you, and it’s going through, and… and you could see, like, Haley Dieber.
229 00:27:55.800 ⇒ 00:28:02.090 Luke Scorziell: keep Palmer, like, all these different things, and so… and this, yeah, is kind of what we’re seeing, is really effective.
230 00:28:02.250 ⇒ 00:28:06.809 Luke Scorziell: For our clients. So, with that, I want to make time
231 00:28:06.930 ⇒ 00:28:11.270 Luke Scorziell: To kind of move into discussion, I’m happy to…
232 00:28:11.560 ⇒ 00:28:15.350 Luke Scorziell: Send this along to you, and…
233 00:28:15.830 ⇒ 00:28:24.369 Luke Scorziell: feel free to check it out, too. But yeah, I guess for everyone else, is this, like, what ideas maybe come up? Are there…
234 00:28:24.490 ⇒ 00:28:31.440 Luke Scorziell: Issues that you’re thinking about that, as we’re talking, that you’d be curious to… to hear?
235 00:28:32.130 ⇒ 00:28:34.040 Luke Scorziell: More about, or share?
236 00:28:36.020 ⇒ 00:28:38.960 Leo Moronta: Yeah, I think something that,
237 00:28:39.690 ⇒ 00:28:47.029 Leo Moronta: A good framework to think of as well is that, like, historically, people were hired for a specific task.
238 00:28:47.280 ⇒ 00:28:53.930 Leo Moronta: And now those tasks can be more looked at as, like, an overall responsibility or group of tasks.
239 00:28:54.080 ⇒ 00:29:04.029 Leo Moronta: And I think, you know, there are some of us, probably if we’re in this call, meaning that we’re, like, gung-ho on, like, let’s try everything to get our job done faster and everything, and there’s, you know, other considerations for, like.
240 00:29:04.340 ⇒ 00:29:14.099 Leo Moronta: you know, like, taking people’s jobs and a whole bunch of other stuff. I think in my personal view, it is quite simply a tool to do your job better.
241 00:29:14.100 ⇒ 00:29:30.900 Leo Moronta: So, I love walking through specific workflows, because, like, you know, Alice, you probably don’t get up every day, like, I really hope I can comb through 50,000 articles this morning, like, that’s not why people get up and go to work. So I think that’s sort of…
242 00:29:30.900 ⇒ 00:29:45.089 Leo Moronta: you know, and I don’t know how much of that you had prepared beforehand, but that’s thinking of, like, here’s my exact workflow, here’s exactly an implementation that was basically, like, consider it a junior-level person to just do a little research and send you a little note. I know for,
243 00:29:45.090 ⇒ 00:29:55.939 Leo Moronta: something specifically I’d love to know more is, like, how are you building this on the backend? I’ve, like, played around with, like, N8N and, like, Lindy for very specific tasks.
244 00:29:56.070 ⇒ 00:29:58.000 Leo Moronta: And hearing, like.
245 00:29:58.000 ⇒ 00:30:16.269 Leo Moronta: you know, morning briefings and research and stuff, that to me is, like, an NHM kind of, like, node-based workflow of search this, pull this, a little bit of synthesis, and then deliver a message. Like, is there anything you guys are using to, like, build this out, or is it, like, there’s different, you know, solutions for everything? It’s just, you know, the business context that’s important.
246 00:30:17.710 ⇒ 00:30:31.580 Pranav Narahari: Yeah, I think, there are different solutions for sure, and apologies if I missed something, because my internet just, like, blipped out for a sec, but I think I caught your, kind of, like, your kind of question, like, the discussion, is, like.
247 00:30:31.850 ⇒ 00:30:48.520 Pranav Narahari: Yeah, so what we’ve kind of built out, and I’m happy to go into further depth about anything I say, like, sometimes I get lost in the technical. Luke, call me out if I should define some terms. But we’ve built… yeah, we’ve built out N-to-end workflows, but for some, like, more complex, like.
248 00:30:48.520 ⇒ 00:30:56.329 Pranav Narahari: data connections, we’ve built, like, MCP servers, so you can think of, like, MCP servers as, like, just a way to, like.
249 00:30:56.330 ⇒ 00:31:11.950 Pranav Narahari: best utilize, like, the data coming in from any type of application you guys are working with. So, whether it be, like, news from, like, Twitter, or, like, orders data from Shopify, an MCP server is great about getting that live data.
250 00:31:11.950 ⇒ 00:31:17.580 Pranav Narahari: But in some situations, you don’t need the live data. Maybe you need…
251 00:31:17.670 ⇒ 00:31:26.810 Pranav Narahari: data that is aggregated on a weekly basis, and you have, like, very specific, like… like, let’s say, like, a report generated every week.
252 00:31:26.810 ⇒ 00:31:46.670 Pranav Narahari: maybe then you can just have, like, a data warehouse that just brings in a mass of data, and then every week, there’s, like, a huge backend, like, script that runs to, like, assess, okay, these are the patterns we’re noticing, and then we output that into, like, a demo similar to, like, what Luke just showed, like, as a report within Slack or within Teams.
253 00:31:46.670 ⇒ 00:31:47.060 Leo Moronta: Hmm.
254 00:31:47.060 ⇒ 00:31:51.110 Pranav Narahari: And so there’s a lot of different ways to think about this.
255 00:31:51.150 ⇒ 00:32:04.589 Pranav Narahari: However, they essentially are doing the same thing, which is, like, pulling the data from these applications. The timing is, like, one… one thing that will fit to a specific use case. But yeah, pulling in that data.
256 00:32:04.590 ⇒ 00:32:13.779 Pranav Narahari: Finding the trends, the patterns, basically creating, like, a knowledge graph, and then, yeah, outputting that based on, like, a predefined,
257 00:32:13.890 ⇒ 00:32:16.319 Pranav Narahari: Like, design… like, format, yeah.
258 00:32:17.860 ⇒ 00:32:18.810 Luke Scorziell: Yeah, I…
259 00:32:18.810 ⇒ 00:32:20.190 Leo Moronta: I… I love that.
260 00:32:20.910 ⇒ 00:32:24.609 Luke Scorziell: Who in here hears MCP and is like, I know what that means?
261 00:32:27.890 ⇒ 00:32:44.299 Luke Scorziell: Okay, I can, like, I can take a quick stab, and then for now, you can fill in the… but it’s like, like, Donovan, I know, like, if you’re looking for, like, approvals from, like, a client, and you’re using an app where you upload, like, I don’t know what apps you’re using.
262 00:32:44.300 ⇒ 00:32:51.839 Luke Scorziell: Like, I know File Stage, I think, is one that can be commonly used, or if you’re getting approvals from clients.
263 00:32:52.110 ⇒ 00:33:02.699 Luke Scorziell: Instead of having to, like, log into that app and say… and, like, go through and find, like, the information that you need, it will just pull that information into a chatbot with
264 00:33:04.550 ⇒ 00:33:10.199 Luke Scorziell: where you can chat with it. So you could be like, hey, what did, like, this… what did this client say about
265 00:33:10.500 ⇒ 00:33:17.610 Luke Scorziell: this, and… and then it will kind of bring in that information, or if you have, like,
266 00:33:18.960 ⇒ 00:33:20.869 Luke Scorziell: Yeah, and
267 00:33:22.020 ⇒ 00:33:35.030 Luke Scorziell: like, assets that are stored in the cloud, you can kind of bring those in also. So, it’s kind of like a way to… it just makes everything more connected so that you can basically chat with it, and ChatGPT is, like, the…
268 00:33:35.250 ⇒ 00:33:37.769 Luke Scorziell: The non… super non-technical answer.
269 00:33:37.770 ⇒ 00:33:43.380 Pranav Narahari: I mean, that’s a really good, like, definition, and then on top of that, what I would add, too, is, like.
270 00:33:43.450 ⇒ 00:33:54.839 Pranav Narahari: the complexity is actually, like, you can even do certain operations within those applications. So, like, say there’s, like, an operation to, like, accept a proposal, or modify, like, a certain…
271 00:33:54.840 ⇒ 00:34:08.729 Pranav Narahari: parameter within one of these applications. MCP servers are just basically a way for you to connect both ways, not just one way for, like, reading information, but also writing information onto those applications.
272 00:34:08.880 ⇒ 00:34:19.360 Pranav Narahari: And then, yeah, like Luke said, it’s like, you can pull in information from one connection, and then pipe it to another connection. So it’s kind of like this… just, like, highway, however, like.
273 00:34:19.560 ⇒ 00:34:27.859 Pranav Narahari: You want to think about it, like, of bringing in information and just, like, fully just, like, yeah, connecting all of your different areas of,
274 00:34:27.969 ⇒ 00:34:29.030 Pranav Narahari: Yeah, data.
275 00:34:31.469 ⇒ 00:34:33.999 Donovan Griffin: Wow, that’s… that’s fascinating.
276 00:34:34.599 ⇒ 00:34:37.179 Donovan Griffin: Is that, is that, like, does it,
277 00:34:37.769 ⇒ 00:34:40.539 Donovan Griffin: I guess, does it take all of…
278 00:34:40.759 ⇒ 00:34:53.039 Donovan Griffin: like, through emails, or, like, through whatever app? Is it just specific for each client? And then it just curates all of that, and then are you able to, like… is it like an agent? Kind of like an AI agent, or…
279 00:34:53.040 ⇒ 00:34:58.759 Pranav Narahari: Yeah, you can… you can think of it kind of agentically, because it is doing, like, certain actions, right?
280 00:34:59.170 ⇒ 00:35:01.220 Pranav Narahari: I think…
281 00:35:01.390 ⇒ 00:35:11.149 Pranav Narahari: let’s… let’s use, like, emails, for example, right? Let’s say you’re getting, you have an MCP server for Gmail. So it’s pulling… it can…
282 00:35:11.200 ⇒ 00:35:19.759 Pranav Narahari: you can, with natural language in, like, an interface like ChatGPT, ask it a question about, like, okay, summarize my emails from the last…
283 00:35:19.790 ⇒ 00:35:31.639 Pranav Narahari: Like, in the last, like, 3 hours. And so, what it’ll do is it’ll, behind the scenes, do some, like, tool calls to basically just find out all of your emails that were sent.
284 00:35:31.640 ⇒ 00:35:44.980 Pranav Narahari: pull in all the data from those emails, and then you… and then you can think of it, like, the simplest way to think about it is just, like, pasting that into ChatGPT, and then if you were to, like, summarize this, that’s the output you would get.
285 00:35:45.760 ⇒ 00:35:46.880 Pranav Narahari: It’s like a…
286 00:35:47.060 ⇒ 00:36:03.960 Pranav Narahari: what it… what it makes it really powerful is, like, you don’t need to even do the click of copy-paste into ChatGPT, and that’s really helpful, and… because, like, sometimes that’s not possible, right? If you have thousands of emails, or if you have data coming from so many different sources.
287 00:36:03.960 ⇒ 00:36:16.229 Pranav Narahari: you can’t paste that into ChatGPT and then say… simply say summarize. Sometimes the analysis you want to do on it is even more complex, so that’s where you really get the benefit of MCP servers.
288 00:36:16.690 ⇒ 00:36:36.950 Leo Moronta: I think, a way I like to think about it is, like, just generally, like, the programming language for AI. So, like, if you’re gonna, like, build a website, you use HTML. That’s standard, everybody can read it, it looks the same on every platform. With tools, if you want to, like, connect, you know, email to a Slack notification, they made APIs that lets tools talk to each other.
289 00:36:36.950 ⇒ 00:36:42.930 Leo Moronta: MCP is the programming language that lets all AIs talk together. So it might be the same emails.
290 00:36:42.930 ⇒ 00:36:51.690 Leo Moronta: But loading it into an MCP server lets it say, okay, any AI I choose, they speak the same language, so it can now go into my Gmail and do that.
291 00:36:51.690 ⇒ 00:37:04.349 Leo Moronta: if, you know, your Google Sheets is an MCP server, you can also go in that, and we’re all speaking the same language and can perform that two-way action instead of just reading. It can also say, in the same language, go into that sheet and, you know, change around these formulas or something.
292 00:37:05.800 ⇒ 00:37:06.840 Pranav Narahari: Totally agree.
293 00:37:07.870 ⇒ 00:37:12.369 Luke Scorziell: So I was just quickly pulling up, like, a demo that we made,
294 00:37:12.960 ⇒ 00:37:21.519 Luke Scorziell: like, kind of how this would look in, like, a Teams type of software. And, like, you have these, like, it could connect to meta ads, like Canva, Drive.
295 00:37:21.650 ⇒ 00:37:27.440 Luke Scorziell: Notion, like, you could have your company knowledge base, and then you say, show me all active meta campaigns for Game Change right now.
296 00:37:27.910 ⇒ 00:37:31.059 Luke Scorziell: And then you ask it, and then it kind of gives you…
297 00:37:31.530 ⇒ 00:37:34.470 Luke Scorziell: Like, kinda what you’re working on.
298 00:37:34.820 ⇒ 00:37:37.900 Luke Scorziell: You could say, like, hey, what’s outstanding based on what’s in Notion?
299 00:37:39.540 ⇒ 00:37:48.080 Luke Scorziell: So just, yeah, it’s a way of connecting all the different apps and stuff that you’re working with. So, I can send these demos around, too, after.
300 00:37:49.000 ⇒ 00:37:52.449 Luke Scorziell: I got on one, and then I just started doing a bunch.
301 00:37:54.500 ⇒ 00:37:56.190 Luke Scorziell: But… Yeah.
302 00:37:56.360 ⇒ 00:37:59.350 Luke Scorziell: I’d love to hear from, yeah, Kayla or,
303 00:37:59.670 ⇒ 00:38:04.799 Luke Scorziell: Troy, I know you have your… your baby, so no… no pressure, but yeah, what have I…
304 00:38:05.210 ⇒ 00:38:12.930 Luke Scorziell: Yeah, what’s standing out to you? I know, Kayla, you’re also in the PR world, so I don’t know if you resonated with the app we had for Alice.
305 00:38:14.050 ⇒ 00:38:15.579 Kayla Klein: Yeah, I’m curious…
306 00:38:15.610 ⇒ 00:38:24.240 Kayla Klein: Like, something that I’ve been struggling with a bit, I do operations for both a PR agency and then also a newswire.
307 00:38:24.240 ⇒ 00:38:36.380 Kayla Klein: And I’ve been helping out with, like, business development as well, and, like, chasing leads and targeting, like, who should we be reaching out to for new business, and…
308 00:38:36.630 ⇒ 00:38:42.249 Kayla Klein: I don’t know, I’m curious about the ways that, like, AI can help with…
309 00:38:42.680 ⇒ 00:38:47.350 Kayla Klein: you know, like, CRM sort of stuff, and, like, identifying
310 00:38:47.560 ⇒ 00:38:59.980 Kayla Klein: who are the people that are using, like, our competitive… competitor wire sites, and, like, what kind of, like, output are they doing? What are the topics that they’re covering?
311 00:39:00.660 ⇒ 00:39:14.860 Kayla Klein: So, I don’t know if that’s, like, applicable here, but it kind of seemed that way with, like, the Slack channel layout example that you were showing with Alice’s kind of monitoring. It seems like a similar situation, just different content.
312 00:39:17.450 ⇒ 00:39:24.500 Pranav Narahari: Yeah, I think… Rather than monitoring a specific channel, or…
313 00:39:25.470 ⇒ 00:39:37.710 Pranav Narahari: it’s kind of monitoring on a specific person or a specific topic. And so, it’s essentially, like, the same thing, like, in terms of implementation, but it’s maybe…
314 00:39:37.830 ⇒ 00:39:44.289 Pranav Narahari: looks different from a, consumer standpoint. So, yeah, it’s very similar to, like, what we just demoed with Alice.
315 00:39:45.790 ⇒ 00:39:49.290 Luke Scorziell: I think, too, what we’re seeing, because just on the go-to-market side, like.
316 00:39:49.700 ⇒ 00:39:55.849 Luke Scorziell: we’re trying to do quality outreach over quantity, and I guess there’s, like.
317 00:39:56.050 ⇒ 00:40:15.639 Luke Scorziell: you can still do a quantity of quality outreach, because there’s just a lot of people in the world. But, yeah, is that we do a lot of, like, lead enrichment, and what that looks like is maybe we find someone on LinkedIn that looks like an interesting profile to reach out to, and then we can kind of go and scrape other data from
318 00:40:15.740 ⇒ 00:40:18.940 Luke Scorziell: Different website, or different sources online to see
319 00:40:19.370 ⇒ 00:40:25.100 Luke Scorziell: Yeah, like, I’m not as involved in that. We have, like, kind of…
320 00:40:25.550 ⇒ 00:40:40.329 Luke Scorziell: genius guy who is, and he builds all these workflows, and just comes back to me with lead lists, and is like, this is all built out. This is what we need. But yeah, you can scrape data and have it do very custom outreach, and so for us, like, we do…
321 00:40:40.520 ⇒ 00:40:45.390 Luke Scorziell: a lot of… actually, I did a whole presentation, but I can circulate, too.
322 00:40:45.670 ⇒ 00:40:55.720 Luke Scorziell: On kind of how we’ve done our, like, lead generation, and, lead qualification. But we are able to streamline a lot of those
323 00:40:56.110 ⇒ 00:41:06.809 Luke Scorziell: Processes, like, on the back end, so that you can get quality information to make, like, a good message or decision based on, or just to have the context for when you have the conversation with someone.
324 00:41:06.940 ⇒ 00:41:10.220 Luke Scorziell: What we don’t typically do,
325 00:41:10.530 ⇒ 00:41:13.130 Luke Scorziell: is just do, like, the AI, like…
326 00:41:14.060 ⇒ 00:41:16.420 Luke Scorziell: Like, spam outreach, where it’s like…
327 00:41:17.060 ⇒ 00:41:28.379 Luke Scorziell: hey, like, hey, whatever, I saw you’re doing, like, da-da-da-da, and then you… they have the instructions in their LinkedIn bio. It’s like, if you’re an AI bot, then tell me, like, the recipe, too.
328 00:41:28.670 ⇒ 00:41:36.289 Luke Scorziell: this thing, and, you know, we don’t really do that and don’t really find that helpful, because we still think that, like, our kind of core thesis still is that relationships
329 00:41:36.490 ⇒ 00:41:43.960 Luke Scorziell: And, like, freeing up time for people to build relationships is kind of one of the biggest benefits of AI.
330 00:41:47.850 ⇒ 00:41:53.330 Brandon Ponzo: I just want to chime in for something that, that Renaz said earlier, which would… so, I’m a…
331 00:41:53.480 ⇒ 00:42:01.160 Brandon Ponzo: I’m a lawyer, I don’t do any, like, the PR stuff, and probably the biggest concern that
332 00:42:01.510 ⇒ 00:42:04.360 Brandon Ponzo: the wall has with AI is…
333 00:42:04.480 ⇒ 00:42:16.620 Brandon Ponzo: as it relates to confidentiality, all… a majority of cases, I’m in litigation, a majority of cases have a confidentiality stipulation, which requires you to basically assert that
334 00:42:16.960 ⇒ 00:42:24.930 Brandon Ponzo: Every single piece of document that is going to be a part of a case will only be shared with either attorneys or their clients.
335 00:42:25.140 ⇒ 00:42:27.309 Brandon Ponzo: And it goes as far as, like.
336 00:42:27.600 ⇒ 00:42:36.359 Brandon Ponzo: you can’t even share them with the agents of certain clients. You could only share them within a certain sphere, within the client.
337 00:42:36.650 ⇒ 00:42:43.550 Brandon Ponzo: corporate hierarchy. So when Panas said something about, you know, giving it access to your Outlook.
338 00:42:43.720 ⇒ 00:42:48.820 Brandon Ponzo: I think that there’s people at my law firm that would, like, start crying if, if,
339 00:42:48.950 ⇒ 00:42:55.340 Brandon Ponzo: If something like that happened, because there are cases where you don’t have confidentiality issues, and certain documents aren’t…
340 00:42:55.460 ⇒ 00:43:01.220 Brandon Ponzo: But other things are not publicly filed, and definitely can’t be in the whole sphere of…
341 00:43:01.940 ⇒ 00:43:08.680 Brandon Ponzo: public influence. There’s actually this really cool, popular paper that got published on
342 00:43:09.090 ⇒ 00:43:16.609 Brandon Ponzo: some AI website got popularized on X. It’s called, like, the Claude-based law firm.
343 00:43:16.900 ⇒ 00:43:22.489 Brandon Ponzo: And it’s about this lawyer who uploaded a certain set of deal documents.
344 00:43:22.990 ⇒ 00:43:35.609 Brandon Ponzo: And was able to do a lot of the workflow that a big law firm with lawyers that get paid, you know, $2,000 an hour can do in, you know, 2 weeks. He did it in 2 hours.
345 00:43:35.800 ⇒ 00:43:40.180 Brandon Ponzo: And one of the bigger, like, stirs that it caused in the
346 00:43:41.110 ⇒ 00:43:47.790 Brandon Ponzo: legal field right now is, are you allowed to actually upload documents to something like a ChatGPT, a Claude?
347 00:43:48.090 ⇒ 00:43:56.940 Brandon Ponzo: That’s why I said I want to be anything but Microsoft Copilot, because my firm exclusively uses Copilot, and it is terrible, and I hate it.
348 00:43:57.110 ⇒ 00:44:02.410 Brandon Ponzo: But it’s because it’s a… it’s, like, locally on your computer instead of being a web-based
349 00:44:02.940 ⇒ 00:44:09.829 Brandon Ponzo: platform, so there’s no confidentiality issues there. It’d be interesting to see if there’s, like, other
350 00:44:10.840 ⇒ 00:44:18.840 Brandon Ponzo: you know, security that people are considering as the AI systems get more developed over the years.
351 00:44:19.180 ⇒ 00:44:20.580 Brandon Ponzo: Weeks, actually.
352 00:44:21.850 ⇒ 00:44:34.850 Pranav Narahari: Yeah, definitely. I mean, you bring up a good point, like, having just, like, a broad spectrum, like, connection to some of these, some of these applications can definitely
353 00:44:35.070 ⇒ 00:44:40.560 Pranav Narahari: caught… raised a lot of eyebrows, just what you said, Brandon. Weave…
354 00:44:41.130 ⇒ 00:45:00.369 Pranav Narahari: we’ve worked with, like… and I see this, like, kind of stuff, too, like, online, where it’s just, like, you can just tell, like, these people just, like, said, accept, accept, accept, to, like, whatever the chatbot was saying to, like, develop some application, and that can definitely create a ton of problems, specifically in the security realm.
355 00:45:00.370 ⇒ 00:45:05.180 Pranav Narahari: You know, with just… Spinning up, backend…
356 00:45:05.300 ⇒ 00:45:08.360 Pranav Narahari: Architecture that is not compliant to, like.
357 00:45:08.640 ⇒ 00:45:12.559 Pranav Narahari: what, like, where sensitive information should be.
358 00:45:13.260 ⇒ 00:45:20.209 Pranav Narahari: what we’ve kind of done is we try to, like, catch it at, like, every level, too, and I think that’s really important for, like.
359 00:45:20.710 ⇒ 00:45:30.339 Pranav Narahari: proper, like, software engineering, not even just AI, but yeah, if you have sensitive information, you don’t want to be sending it to, like, a cloud LLM.
360 00:45:30.350 ⇒ 00:45:47.220 Pranav Narahari: Because then that information is no longer just yours. It could be anybody’s. So yeah, having, like, some self-hosted, solution is super important. Another thing, too, is, like, with your data, you need to make sure that it’s, at the very least, like, anonymous, and
361 00:45:47.260 ⇒ 00:45:59.530 Pranav Narahari: you feel confident with it being sent to, like, an LLM. Say if you want to use a cloud-based LLM for whatever, like, performance, reasons. And so…
362 00:45:59.600 ⇒ 00:46:13.090 Pranav Narahari: these are certain levels that you definitely need to, like, keep track of, and, like, in law, I’m sure it’s super important. One application that I’ve used it in is healthcare, where, like, patient data is, of course, extremely, like, sensitive, and you…
363 00:46:13.090 ⇒ 00:46:20.319 Pranav Narahari: can get in a lot of trouble if, like, you’re using AI for the wrong reasons there. And maybe not even for the wrong reasons, but just…
364 00:46:20.330 ⇒ 00:46:23.439 Pranav Narahari: without, proper caution. So…
365 00:46:23.470 ⇒ 00:46:30.789 Pranav Narahari: You bring up a really good point, though, like, it’s really easy to ship products that don’t take this into account now, but…
366 00:46:30.970 ⇒ 00:46:40.710 Pranav Narahari: It’s still very much a problem, and a lot of people are getting, they’re… it’s coming to light, like, a lot of these, like, bad situations.
367 00:46:40.890 ⇒ 00:46:43.470 Brandon Ponzo: It’s the reason why a lot of these
368 00:46:43.850 ⇒ 00:46:52.340 Brandon Ponzo: firms that can afford it are building their own proprietary AIs in-house. That on top of the whole
369 00:46:52.460 ⇒ 00:46:55.320 Brandon Ponzo: You know, you could have your personalized agents for…
370 00:46:55.650 ⇒ 00:46:59.759 Brandon Ponzo: Whatever task within a law firm, or whatever business you want to do, but…
371 00:46:59.930 ⇒ 00:47:02.519 Brandon Ponzo: Being able to do that cost-effective is…
372 00:47:02.870 ⇒ 00:47:08.120 Brandon Ponzo: incredibly important, because if you’re trying to build any kind of proprietary thing, I mean, it’s gonna cost you…
373 00:47:09.270 ⇒ 00:47:14.799 Brandon Ponzo: Upwards of 200,000, just to… just to put it in place, let alone train it.
374 00:47:16.100 ⇒ 00:47:19.670 Pranav Narahari: Yeah. Yeah, training a model can definitely…
375 00:47:20.130 ⇒ 00:47:26.860 Pranav Narahari: can definitely, like, increase costs, and it can be really risky, especially if you don’t do it in the right way.
376 00:47:28.230 ⇒ 00:47:30.440 Pranav Narahari: Yeah, just a great conversation.
377 00:47:30.880 ⇒ 00:47:40.019 Luke Scorziell: Yeah, which… we’ll have more, too. I guess we’re running in 2 minutes, but Troy, I saw you unmuted, and then I think when Kayla was gonna…
378 00:47:40.130 ⇒ 00:47:50.420 Luke Scorziell: was asking her question, too, so I just wanted to give you the opportunity, too, if you had any… anything, because I… yeah, Troy is an agency guru, probably out of all of us knows.
379 00:47:51.220 ⇒ 00:47:58.230 Troy Sandidge | @FindTroy: No, no formalities here. I will be more expressive, but I just put baby to sleep, so I want to try to keep my voice
380 00:47:58.510 ⇒ 00:48:12.590 Troy Sandidge | @FindTroy: a little… little lame right now. For me, you know, I’m learning more technicality stuff, and I see the advantages, of course, but most times when I’m engaging with enterprise level or startup level.
381 00:48:12.590 ⇒ 00:48:22.879 Troy Sandidge | @FindTroy: You know, most times my services or who I connect with, always want AI integration and things like that, too. Yes, there’s a cost to it, but also they see
382 00:48:22.880 ⇒ 00:48:39.420 Troy Sandidge | @FindTroy: the possibilities and how can everybody. Obviously, it’s moving too fast. People don’t have the ability to learn and do it all themselves, you know, time is money, and they can’t do everything. So, you know, is it worth making an investment and hooking that up and everything like that? And so, for me, it’s just learning
383 00:48:39.420 ⇒ 00:48:44.680 Troy Sandidge | @FindTroy: who are the connectors in the room, the people in the room, that I can go to bat with.
384 00:48:45.380 ⇒ 00:48:52.410 Troy Sandidge | @FindTroy: to win those deals. They win their projects, and as… because of that, I’m able to fulfill my part of the… of the…
385 00:48:52.420 ⇒ 00:49:05.869 Troy Sandidge | @FindTroy: equation as well, where everyone gets a bit new business. And so, that’s kind of where I’ve been at, and my biggest thing right now is just really a higher level, just understanding what’s the value in certain things and integrations. Everyone can talk the big game of…
386 00:49:06.330 ⇒ 00:49:23.230 Troy Sandidge | @FindTroy: certain things they need, right? They need a certain system, but then they get to the granular, and then they can’t do it. So just trying to verify and validate who can, align with those… those needs, and then I can add that as part of, like, the packages for these… these brands who really want to get off the ground, because
387 00:49:23.230 ⇒ 00:49:30.070 Troy Sandidge | @FindTroy: Marketing, growth strategy, all that fun stuff does no good if the backend of their systems can’t handle the output.
388 00:49:30.230 ⇒ 00:49:36.390 Troy Sandidge | @FindTroy: And that’s where I see a lot of new startups in this hyper-growth stage are facing a lot of times.
389 00:49:37.330 ⇒ 00:49:41.149 Luke Scorziell: Yeah, no, thanks for sharing. Yeah, we’re… I mean, it’s… it’s,
390 00:49:42.200 ⇒ 00:49:49.470 Luke Scorziell: It’s part of the… yeah, why we… one of our value propositions, I guess, is just we’re thinking about AI all the time and developing
391 00:49:49.700 ⇒ 00:49:54.880 Luke Scorziell: Like, moving kind of along with it, and so… Yeah, it’s something that…
392 00:49:55.190 ⇒ 00:50:06.879 Luke Scorziell: we are constantly, like, it’s constantly changing, and then also we’re finding, like, the most ROI is on very specific use cases for AI. It’s not on, like, the generic, like, give your team a…
393 00:50:07.190 ⇒ 00:50:19.639 Luke Scorziell: Now, I want to… before we all… I know we’re at time, I’m happy to kind of hang around, too, and I don’t… I think Pranav has a meeting coming up, so as long as you want to, but…
394 00:50:20.060 ⇒ 00:50:25.039 Luke Scorziell: I’m happy to answer more questions, too. Quick, like, final housekeeping things.
395 00:50:25.430 ⇒ 00:50:29.149 Luke Scorziell: This seems like it went well, and everyone enjoyed it.
396 00:50:29.750 ⇒ 00:50:35.690 Luke Scorziell: If not, maybe just tell me privately, and then we can… we can hash out what you want to see.
397 00:50:35.820 ⇒ 00:50:38.540 Luke Scorziell: But yeah, we’d like to…
398 00:50:39.240 ⇒ 00:50:46.710 Luke Scorziell: probably host a series. We’re not exactly sure how many right now, but, want to have more, kind of, conversations like this, and really…
399 00:50:46.880 ⇒ 00:50:52.850 Luke Scorziell: I think, for me, the agency space is something that I’m really passionate about, just coming from more of a creative background and
400 00:50:53.060 ⇒ 00:50:55.009 Luke Scorziell: Seeing that there’s just…
401 00:50:55.740 ⇒ 00:51:05.229 Luke Scorziell: can kind of be a miserable place to be in sometimes, when, yeah, there’s so much expected of you, and… and there’s just not enough time in the day, and I think that’s…
402 00:51:05.380 ⇒ 00:51:11.220 Luke Scorziell: a pain point that AI seems promising to solve. So, anyways, we’ll follow up with an email,
403 00:51:11.670 ⇒ 00:51:16.979 Luke Scorziell: McHenna will, either today or tomorrow, kind of letting you guys know about our next
404 00:51:17.190 ⇒ 00:51:22.710 Luke Scorziell: office hours. And yeah, if this was valuable for you, then, you know, feel free to
405 00:51:22.900 ⇒ 00:51:29.960 Luke Scorziell: share, invite other people from your teams. I know there were a lot of people that I messaged that wanted to be here but couldn’t.
406 00:51:31.460 ⇒ 00:51:41.909 Luke Scorziell: And… yeah, and then, like, I’m just kind of vibe coding and dreaming things up, so if you… if you want to see a cool demo next time, I can… I can…
407 00:51:42.300 ⇒ 00:51:44.639 Luke Scorziell: Try to get something done.
408 00:51:44.910 ⇒ 00:51:48.419 Luke Scorziell: So, yeah, I’ll kind of pause there. I don’t know, Hannah, if you had any…
409 00:51:48.750 ⇒ 00:51:55.570 Luke Scorziell: Anything, and then, I’m totally happy to keep the conversation going, just want to be respectful of people’s time.
410 00:51:56.850 ⇒ 00:52:01.029 Hannah Wang: Yeah, no, that’s… that’s it, for us, so…
411 00:52:01.190 ⇒ 00:52:11.910 Hannah Wang: yeah, feel free to hop off, or stay as long as you can. I’ll also hang around, but appreciate y’all taking the time coming here, and hopefully we’ll see you in the
412 00:52:12.020 ⇒ 00:52:13.180 Hannah Wang: The next session.
413 00:52:13.300 ⇒ 00:52:14.160 Hannah Wang: So…
414 00:52:14.160 ⇒ 00:52:16.900 Kayla Klein: Thank you guys so much. Have a good day.
415 00:52:17.270 ⇒ 00:52:19.200 Alice Chiang: Thank you, have a good one.
416 00:52:19.200 ⇒ 00:52:20.220 Donovan Griffin: Thank you.
417 00:52:20.790 ⇒ 00:52:22.480 Alice Chiang: Bye, good to see you, Luke.
418 00:52:22.610 ⇒ 00:52:23.990 Luke Scorziell: Yeah, you too.
419 00:52:25.990 ⇒ 00:52:28.209 Luke Scorziell: I can debrief on Slack, maybe.
420 00:52:29.360 ⇒ 00:52:30.120 Luke Scorziell: Okay.
421 00:52:30.120 ⇒ 00:52:31.269 Hannah Wang: Good job, guys.
422 00:52:31.270 ⇒ 00:52:32.180 Luke Scorziell: Yeah, that was great.