Meeting Title: AI Team Retro/Weekly Close Date: 2025-02-07 Meeting participants: Janna Wong, Uttam Kumaran, Miguel De Veyra, Casie Aviles


WEBVTT

1 00:03:05.540 00:03:07.870 Miguel de Veyra: Hey, Adam? Wow! You’re early!

2 00:03:09.240 00:03:10.789 Uttam Kumaran: I know. Dude, I beat you.

3 00:03:11.808 00:03:13.199 Miguel de Veyra: I can’t! I can’t!

4 00:03:13.200 00:03:14.670 Uttam Kumaran: Let you win every single time.

5 00:03:15.690 00:03:17.380 Miguel de Veyra: 1st time in forever.

6 00:03:17.800 00:03:20.150 Miguel de Veyra: I was expecting to see Jana here.

7 00:03:20.150 00:03:21.650 Uttam Kumaran: You underestimate me. Dude.

8 00:03:26.950 00:03:28.559 Miguel de Veyra: You tried the link or not, yet.

9 00:03:28.844 00:03:33.105 Uttam Kumaran: Not yet. I wanna see. I wanna see you guys go through the demo

10 00:03:33.510 00:03:35.599 Uttam Kumaran: because I kind of want to just like.

11 00:03:35.980 00:03:39.889 Uttam Kumaran: because I was so close to them. I can kind of give some, give some good feedback.

12 00:03:40.010 00:03:43.399 Uttam Kumaran: and I want to just run through like cause I want. I would prefer if you guys.

13 00:03:43.600 00:03:49.020 Uttam Kumaran: Demo, also because I want to introduce you guys to the team but we can talk a little bit about that

14 00:03:51.520 00:03:54.640 Uttam Kumaran: if do you think Jan is gonna join as well

15 00:03:55.671 00:03:58.480 Uttam Kumaran: for later? No, I don’t think so.

16 00:03:58.960 00:04:06.310 Miguel de Veyra: At this meeting. I I’m not sure she should, but she’s been working like since the morning, so I’m not sure.

17 00:04:06.460 00:04:07.030 Uttam Kumaran: Okay.

18 00:04:07.950 00:04:08.355 Miguel de Veyra: Yeah.

19 00:04:14.180 00:04:16.910 Miguel de Veyra: I spoke to Anton a little earlier.

20 00:04:17.540 00:04:18.670 Uttam Kumaran: You call him on the phone.

21 00:04:19.314 00:04:20.659 Miguel de Veyra: No, no! Only on the chat.

22 00:04:20.660 00:04:22.802 Uttam Kumaran: Okay, yeah. Dude you should.

23 00:04:23.890 00:04:28.050 Uttam Kumaran: you should chat with him, because yeah, I’ll send you. He’s a

24 00:04:29.520 00:04:34.500 Uttam Kumaran: I think he’s gonna be like really, really incredible. And he’s also

25 00:04:34.740 00:04:36.709 Uttam Kumaran: like not looking for a ton of money.

26 00:04:38.690 00:04:41.110 Uttam Kumaran: So I do think that there’s something that could work out.

27 00:04:43.100 00:04:45.030 Miguel de Veyra: That seems for 35, or something.

28 00:04:45.630 00:04:49.950 Uttam Kumaran: I don’t know. We haven’t talked numbers, but he’s trying to quit his job.

29 00:04:50.320 00:04:55.400 Uttam Kumaran: And yeah, I don’t know. It’s open to opportunities. So.

30 00:04:55.990 00:05:02.179 Miguel de Veyra: I see. Yeah, he did mention. You know, he wants to move from Sweden degree, and he wants like part time work.

31 00:05:02.370 00:05:03.360 Uttam Kumaran: Yeah.

32 00:05:03.540 00:05:06.170 Miguel de Veyra: To supplement. I guess whatever is gonna work.

33 00:05:06.390 00:05:08.619 Uttam Kumaran: And if he’s on your time zone it’s good, too. So.

34 00:05:08.940 00:05:11.329 Miguel de Veyra: Yeah. Japan’s like, 2 or 3 h ahead of us.

35 00:05:11.590 00:05:12.180 Uttam Kumaran: Okay.

36 00:05:15.980 00:05:18.585 Uttam Kumaran: okay, cool. Should we get started?

37 00:05:19.420 00:05:25.772 Uttam Kumaran: yeah, I guess. Let me get. Let me give you guys a brief of how the meeting went yesterday. So

38 00:05:26.190 00:05:32.250 Uttam Kumaran: Do you guys happen to get a chance to watch the the meeting? Invite.

39 00:05:34.100 00:05:36.150 Miguel de Veyra: You mean the loop that you sent.

40 00:05:38.150 00:05:38.620 Uttam Kumaran: Yes!

41 00:05:39.426 00:05:42.189 Miguel de Veyra: Yeah, only a bit. Not the entire. Almost 3 h.

42 00:05:43.300 00:05:44.129 Uttam Kumaran: Okay. Okay. Cool.

43 00:05:44.350 00:05:45.010 Miguel de Veyra: Well, if you.

44 00:05:45.010 00:05:49.590 Uttam Kumaran: Find a chance to listen to that. It’s like a podcast really worth listening to.

45 00:05:49.720 00:05:54.470 Uttam Kumaran: because we walk through every single part of the project?

46 00:05:55.820 00:05:59.639 Uttam Kumaran: Which part did you guys catch? And which part should I like? Try to go through.

47 00:06:02.500 00:06:04.690 Miguel de Veyra: Honestly, I spent more time just.

48 00:06:04.690 00:06:14.240 Uttam Kumaran: Okay. So let me let me let me walk through an overall. And also I’m gonna be on a plane for like a while today. So I’m gonna work on like a project plan.

49 00:06:15.970 00:06:17.069 Uttam Kumaran: I’m going to Argentina.

50 00:06:17.070 00:06:19.540 Miguel de Veyra: Oh! Oh! To Nico.

51 00:06:19.540 00:06:26.570 Uttam Kumaran: Yeah, I’m gonna go spend some time with a friend, and then I’ll be meeting with Nico next week. He’s currently on his bachelor party.

52 00:06:27.870 00:06:29.509 Miguel de Veyra: Oh, cause he’s gonna get married soon, right.

53 00:06:29.510 00:06:30.579 Uttam Kumaran: Yeah, yeah.

54 00:06:30.580 00:06:32.190 Miguel de Veyra: Oh, yeah, shit. Yeah, yeah.

55 00:06:33.030 00:06:35.490 Miguel de Veyra: So you won’t be here Monday, or you’ll be here Monday.

56 00:06:36.750 00:06:38.600 Uttam Kumaran: I will. Oh, I’ll be on. Yeah, I’ll be on.

57 00:06:38.760 00:06:41.459 Uttam Kumaran: I may take a day off next week.

58 00:06:42.200 00:06:49.000 Uttam Kumaran: but it’ll be last minute. It’ll be like. I mean, I feel like Wednesday. It looks like my schedule is pretty clear.

59 00:06:53.430 00:06:56.519 Uttam Kumaran: But you know, only God knows.

60 00:07:02.300 00:07:03.866 Uttam Kumaran: Okay, cool. So let me

61 00:07:05.140 00:07:08.885 Uttam Kumaran: let me give you a brief overview.

62 00:07:10.690 00:07:13.330 Uttam Kumaran: So the

63 00:07:13.600 00:07:23.210 Uttam Kumaran: we’re working with this. And I I think one of the things that you know, I’m just gonna give an overview of 10 min, and then we can do it, maybe a longer session on Monday, because I know we’re sort of rushing on this.

64 00:07:23.430 00:07:28.899 Uttam Kumaran: But ABC home and Commercial is one of the most

65 00:07:29.610 00:07:35.870 Uttam Kumaran: is is an ex extremely well known home and Commercial Services Company

66 00:07:37.030 00:07:43.120 Uttam Kumaran: based in Texas. They’re they’re almost like a 75 year old company, a little bit older.

67 00:07:43.666 00:07:56.849 Uttam Kumaran: And they’re one of the largest home services. What are home services? So home services are pest control, roofing, appliance repair. When you own a home, you may need one or many of these.

68 00:07:57.965 00:07:59.569 Uttam Kumaran: And so I

69 00:08:00.020 00:08:14.451 Uttam Kumaran: I want you guys to think about who. In fact, like we’re we’re really dealing with here and sort of what their, what their needs are. I’m gonna I I sort of scrambled a bunch of notes while I was meeting with them.

70 00:08:16.160 00:08:24.803 Uttam Kumaran: yesterday. And so I’ll just walk through a couple of things. So basically,

71 00:08:26.480 00:08:43.430 Uttam Kumaran: ABC, home has several divisions. As I mentioned, we are primarily working with the pest, division, pest, and rodent division. These are all their services. Bed bug treatment fire and control mosquito control. You guys like, what do you guys have like we don’t have like in Texas? We have scorpions

72 00:08:43.960 00:08:52.289 Uttam Kumaran: like, what are the bugs there like, do you guys have the same type of bugs? Or is it like different bugs like we have cock? We have some cockroaches, but it’s mostly mosquitoes here.

73 00:08:52.830 00:08:56.500 Miguel de Veyra: And it’s mostly cockroaches and termite.

74 00:08:59.010 00:08:59.970 Casie Aviles: Yeah. Let’s see.

75 00:09:01.200 00:09:03.080 Miguel de Veyra: We call the termite we call anai.

76 00:09:03.840 00:09:04.560 Uttam Kumaran: Hmm!

77 00:09:05.030 00:09:06.700 Uttam Kumaran: How do you spell a MAI.

78 00:09:06.890 00:09:10.100 Miguel de Veyra: Yeah, why is it? Why, I’m not? Yeah. Why.

79 00:09:10.573 00:09:17.450 Miguel de Veyra: like, if you go to a residential area, you see it everywhere like, and I something, and I something like.

80 00:09:17.450 00:09:21.579 Uttam Kumaran: Because because they try to, they try to help like clear it out.

81 00:09:22.140 00:09:25.250 Miguel de Veyra: Yeah, yeah, cause it eats up your cabinets and stuff, you know. So it’s.

82 00:09:25.250 00:09:29.540 Uttam Kumaran: Okay. So you got. So you know, you’re familiar with like home services as a category.

83 00:09:29.980 00:09:43.100 Uttam Kumaran: So basically they have this pest control division pest division has 3 has 3 core people, Janice.

84 00:09:43.980 00:09:50.390 Uttam Kumaran: She’s the head of the she’s the head of the pest. So there’s 3 people I’m gonna I’ll just write down who they are. So there’s Yvette.

85 00:09:51.370 00:09:53.060 Uttam Kumaran: She is like head.

86 00:09:53.300 00:09:56.240 Uttam Kumaran: I don’t know where she’s basically like head of operations. I think.

87 00:09:56.810 00:09:58.680 Miguel de Veyra: Yeah, yeah. Our bot knows that.

88 00:09:59.440 00:10:00.429 Uttam Kumaran: Oh, really. Okay.

89 00:10:00.430 00:10:01.180 Miguel de Veyra: Yeah.

90 00:10:01.960 00:10:04.232 Uttam Kumaran: Did you look into? Did you look into like the

91 00:10:08.220 00:10:12.710 Uttam Kumaran: Yeah, I forgot what? Let me look at. Yeah, she’s the customer service director. So.

92 00:10:13.400 00:10:15.209 Miguel de Veyra: Event, customer, care director.

93 00:10:15.210 00:10:21.560 Uttam Kumaran: Customer service director. You have the next person, which is Shannon.

94 00:10:22.170 00:10:24.699 Uttam Kumaran: She’s like the the second.

95 00:10:24.700 00:10:26.750 Miguel de Veyra: Summer Service division manager.

96 00:10:28.130 00:10:29.280 Uttam Kumaran: So who’s Janice?

97 00:10:29.840 00:10:34.530 Miguel de Veyra: Janice is customer care Center, Qa. Training team.

98 00:10:35.740 00:10:41.539 Uttam Kumaran: So so, Shannon. And then, Kenny. So this is the this is the escalation path.

99 00:10:43.830 00:10:45.820 Uttam Kumaran: So who’s Kenny? Ask it? Who’s Kenny?

100 00:10:46.090 00:10:47.729 Miguel de Veyra: Is it Kelly or Kenny.

101 00:10:47.730 00:10:48.490 Uttam Kumaran: Kenny.

102 00:10:49.020 00:10:54.540 Miguel de Veyra: No, the one. The one in the document is Kelly Kelly’s true. There’s not Kenny.

103 00:10:54.760 00:11:05.850 Uttam Kumaran: Okay, okay. So so Janice is, let’s just say, head of past Csr back end in command.

104 00:11:08.200 00:11:08.800 Uttam Kumaran: Actually.

105 00:11:08.800 00:11:11.429 Miguel de Veyra: Oh, here’s Kenny. He’s support Supervisor. Sorry.

106 00:11:11.750 00:11:14.660 Uttam Kumaran: Yeah, okay, so Csr Supervisor.

107 00:11:15.000 00:11:20.010 Uttam Kumaran: And then Shannon is like, like the Csr manager.

108 00:11:20.010 00:11:20.899 Miguel de Veyra: Measure, yeah.

109 00:11:20.900 00:11:22.929 Uttam Kumaran: Yeah. Division manager. Yes.

110 00:11:26.480 00:11:29.370 Uttam Kumaran: these are our stakeholders.

111 00:11:30.014 00:11:32.859 Uttam Kumaran: And these are the people whose lives we’re gonna be

112 00:11:33.040 00:11:47.529 Uttam Kumaran: improving with this solution. So not all Yvette is gonna be on this call. I’m gonna see if we can start adding Janice, but ideally, today, I want to make this tool available to to them, to to play around with

113 00:11:47.996 00:12:01.789 Uttam Kumaran: and these event and Janice are like they were so nice. Watch the end of the loom video, there are some video, so you can kind of see, see who they are. And then, ideally, Miguel, next time I go in person would love. Maybe I’ll try to get it scheduled earlier in the morning.

114 00:12:02.323 00:12:11.079 Uttam Kumaran: But would love for you to meet them. They’re really really great, and would love for you to sort of talk to them and ask some questions. But overall, this is this is like the the.

115 00:12:11.470 00:12:15.730 Uttam Kumaran: the order of operations, and then we also have all of our customer service reps right.

116 00:12:19.650 00:12:21.070 Uttam Kumaran: of which there are.

117 00:12:21.070 00:12:21.460 Miguel de Veyra: Question.

118 00:12:21.460 00:12:22.800 Uttam Kumaran: 97 of them.

119 00:12:23.350 00:12:24.400 Miguel de Veyra: Oh, okay.

120 00:12:25.380 00:12:27.610 Uttam Kumaran: So there’s a.

121 00:12:28.930 00:12:34.108 Uttam Kumaran: So there’s like, there’s Route 97 of these folks. And so overall

122 00:12:35.500 00:12:41.029 Uttam Kumaran: the pest control division basically has. Like.

123 00:12:41.190 00:12:46.139 Uttam Kumaran: if you ultimately what this comes down to is a customer calls with one or several questions

124 00:12:46.320 00:13:13.170 Uttam Kumaran: right? And you may have already seen some of these in the documents, but they may ask about what services are covered under what they purchased? When is their technician coming? Can they reschedule can they add another service. Ca, what like? And so all of these questions need to get routed right? And I know Miguel and maybe others on the call. You guys have worked on call center automation stuff. You guys understand? Like these are just questions that need to get routed and answered. The problem with pest control is that

125 00:13:13.290 00:13:17.400 Uttam Kumaran: it takes them about 3 months to train a new employee.

126 00:13:18.597 00:13:24.679 Uttam Kumaran: And so that’s 1 of the biggest problems is there’s a lot of knowledge that needs to

127 00:13:24.840 00:13:29.559 Uttam Kumaran: be referenced in order to do this job effectively. Right? So

128 00:13:32.920 00:13:37.409 Uttam Kumaran: so you could consider that as like retrieval, right?

129 00:13:37.530 00:13:42.800 Uttam Kumaran: So retrieval for the Csrs is hard.

130 00:13:43.150 00:13:46.139 Uttam Kumaran: What does this lead to? This leads to?

131 00:13:46.697 00:13:49.480 Miguel de Veyra: This leads to putting people on hold.

132 00:13:50.210 00:13:53.670 Uttam Kumaran: This leads to, I’ll call you back.

133 00:13:54.200 00:13:57.639 Uttam Kumaran: This leads to transfer to another department.

134 00:13:59.880 00:14:09.920 Uttam Kumaran: All of this leads to customers being unhappy or churning right.

135 00:14:10.610 00:14:13.090 Uttam Kumaran: So this is what we’re ultimately solving.

136 00:14:14.540 00:14:17.420 Uttam Kumaran: I’ll pause there. Does that make sense, I guess, like.

137 00:14:17.790 00:14:30.650 Uttam Kumaran: you know, I know, Miguel, but I’m and I I also want to include you, Casey, because I think in case we need more help on this, you’ll jump in. But for Jana and Casey like, does this make sense overall in terms of like the problem we’re solving here.

138 00:14:33.380 00:14:34.470 Janna Wong: Yeah, I think so.

139 00:14:34.470 00:14:36.870 Casie Aviles: Yeah, yeah, pretty clear.

140 00:14:36.870 00:14:37.470 Uttam Kumaran: Okay.

141 00:14:37.790 00:14:43.360 Uttam Kumaran: great sorry if I’m being like very specific. But I I think it helps to sort of frame everything

142 00:14:44.510 00:14:56.039 Uttam Kumaran: even for me to be like, cause the scope, as you guys will see, is gonna is very can get very wide. So I want to continue to focus us on what the real problem is. So

143 00:14:56.240 00:14:58.780 Uttam Kumaran: to kind of move a little bit further.

144 00:14:58.930 00:15:05.589 Uttam Kumaran: the way that the customer service reps, understand how to answer questions is through the documents that are in

145 00:15:05.740 00:15:06.839 Uttam Kumaran: that file.

146 00:15:07.296 00:15:12.050 Uttam Kumaran: And if I go to pest let’s just go to drive, and I’ll see if I can pull it up

147 00:15:16.140 00:15:17.180 Uttam Kumaran: past.

148 00:15:17.470 00:15:23.360 Uttam Kumaran: So all of the documents here are how the pest team answers questions.

149 00:15:24.740 00:15:26.519 Uttam Kumaran: Let’s just pick a random one.

150 00:15:29.870 00:15:34.490 Uttam Kumaran: Imagine you’re a customer service rep couple of questions.

151 00:15:34.780 00:15:40.059 Uttam Kumaran: You get a call from a client you have. You don’t know how to answer the question. You open the pest department folder.

152 00:15:41.670 00:15:42.860 Uttam Kumaran: Where do you go.

153 00:15:44.370 00:15:45.820 Uttam Kumaran: No clue right.

154 00:15:45.820 00:15:46.440 Miguel de Veyra: Yeah.

155 00:15:46.680 00:15:47.970 Uttam Kumaran: So what’s happening.

156 00:15:49.340 00:15:56.959 Uttam Kumaran: pop? Quiz, what do you think is, what do you think is that happening when the Csr doesn’t know? I guess Jana like, what do you think? What do you think they do when they

157 00:15:57.190 00:15:59.230 Uttam Kumaran: they don’t know exactly like where to go?

158 00:16:00.965 00:16:08.260 Janna Wong: They’ll find time to like hold the person in the call while they find the document.

159 00:16:08.930 00:16:11.659 Uttam Kumaran: Yeah, okay. And then I guess, Casey.

160 00:16:12.150 00:16:16.070 Uttam Kumaran: while they put the person on hold, let’s say they can’t find

161 00:16:16.200 00:16:19.050 Uttam Kumaran: it really quickly. What do you think their next option is.

162 00:16:23.280 00:16:25.139 Casie Aviles: Not not really sure, but.

163 00:16:25.140 00:16:26.510 Miguel de Veyra: Elevator, music.

164 00:16:26.820 00:16:28.020 Casie Aviles: Yeah, I guess.

165 00:16:29.030 00:16:30.899 Uttam Kumaran: Let’s say your cus your customers on hold.

166 00:16:31.300 00:16:34.029 Casie Aviles: You’re like scrambling. You can’t find it here.

167 00:16:34.200 00:16:35.640 Uttam Kumaran: You’re in a call center.

168 00:16:36.350 00:16:37.910 Uttam Kumaran: There’s people around you.

169 00:16:38.660 00:16:39.670 Uttam Kumaran: What do you do?

170 00:16:40.460 00:16:41.700 Casie Aviles: Ask someone.

171 00:16:41.700 00:16:45.059 Uttam Kumaran: Yes, you ask someone next to you. Right?

172 00:16:45.710 00:16:49.100 Uttam Kumaran: So one is it gets us. It gets you ask someone next to you

173 00:16:49.560 00:16:54.350 Uttam Kumaran: what happens then, Miguel, right, and let’s talk about the escalation path.

174 00:16:54.810 00:16:55.370 Miguel de Veyra: Yeah.

175 00:16:55.510 00:16:56.590 Uttam Kumaran: What happens next?

176 00:16:57.970 00:16:59.129 Miguel de Veyra: Yeah. Well.

177 00:16:59.630 00:17:07.650 Miguel de Veyra: couple of things, either. The person you ask doesn’t doesn’t know to. So you guys skim through together, or if he knows, then you’re in luck.

178 00:17:08.599 00:17:15.669 Uttam Kumaran: Exactly. And so the the ABC home is on Google Chats. So either they they and I was just in their call center yesterday.

179 00:17:16.059 00:17:23.149 Uttam Kumaran: You guys know it’s cubicles call center. They stand up. They said, Hey, customer has this question, what should I do? Maybe they get an answer. Maybe they don’t

180 00:17:23.279 00:17:26.759 Uttam Kumaran: they have Google chat? They chat their friend. Hey? What should I do?

181 00:17:27.089 00:17:31.579 Uttam Kumaran: Maybe they get answered. Maybe they don’t. Then the escalation happens. Okay, let me call Shannon.

182 00:17:31.999 00:17:33.289 Uttam Kumaran: Shannon gets called.

183 00:17:33.989 00:17:35.199 Uttam Kumaran: Shannon doesn’t know.

184 00:17:35.379 00:17:43.159 Uttam Kumaran: Okay, Shannon. Then messages Kenny Kenny doesn’t know. Then Kenny messages Janice Janice doesn’t Janice doesn’t know and then Janice has to message, Yvette.

185 00:17:44.239 00:17:47.169 Uttam Kumaran: you guys, even in our slack channel.

186 00:17:47.419 00:17:49.909 Uttam Kumaran: think about how long it takes sometimes to get an answer.

187 00:17:50.209 00:17:52.769 Uttam Kumaran: and we’re pretty. We’re pretty good at like documenting.

188 00:17:53.149 00:17:56.449 Uttam Kumaran: Imagine the pain that these guys go through right?

189 00:17:56.749 00:18:00.269 Uttam Kumaran: So if customer may not hear back for hours, maybe a day.

190 00:18:00.959 00:18:07.069 Uttam Kumaran: they’re like fuck these guys like, you know, we’re gonna we’re gonna we’re not. Gonna I don’t. We’re gonna move to someone else.

191 00:18:07.199 00:18:09.899 Uttam Kumaran: And so here’s a really good example of what this happened.

192 00:18:10.929 00:18:15.039 Uttam Kumaran: They offer the service called rodent annuals rodent. Of course, mice

193 00:18:15.159 00:18:24.269 Uttam Kumaran: like, I don’t know beavers. I don’t know some other possums. Other Texas rodents basically annually. They’ll come and put out traps for rodents.

194 00:18:25.770 00:18:32.080 Uttam Kumaran: So a customer called and said, I have a back patio that is not attached, but is outside.

195 00:18:32.400 00:18:34.499 Uttam Kumaran: Is this covered by the rodent annual.

196 00:18:37.130 00:18:41.120 Uttam Kumaran: Can you put this question into the thing and see what it says? Miguel.

197 00:18:42.120 00:18:46.579 Miguel de Veyra: Okay, appreciate, can you? I have a backpack here that is not attached to this one.

198 00:18:48.010 00:18:54.050 Uttam Kumaran: And so what happens when they when they get this question one they like, they quickly go here.

199 00:18:54.490 00:19:00.379 Uttam Kumaran: They’re like, Okay, I don’t know. Let me just search for like rodent annual.

200 00:19:03.710 00:19:06.100 Uttam Kumaran: Okay, I don’t know. Like cheat notes.

201 00:19:07.697 00:19:13.360 Uttam Kumaran: Okay, let me just let me just do like rodent rodent rodent. Okay, rodent annual.

202 00:19:16.220 00:19:20.479 Uttam Kumaran: I don’t know. Okay, I sorry, sir. I have to let me put you on hold

203 00:19:20.880 00:19:23.059 Uttam Kumaran: until I find out. Right? So

204 00:19:23.600 00:19:26.619 Uttam Kumaran: look at this document. This is this is like brutal.

205 00:19:27.660 00:19:28.660 Uttam Kumaran: So

206 00:19:28.860 00:19:45.569 Uttam Kumaran: basically, the problem we’re trying to solve, given that example is the customer service Rep should be able to quickly on the call type in that question and get told either the answer or get told that it’s not covered like, get told that there is not an answer. Please escalate

207 00:19:45.760 00:19:46.280 Uttam Kumaran: right.

208 00:19:46.280 00:19:46.930 Miguel de Veyra: Yeah.

209 00:19:47.320 00:19:51.362 Uttam Kumaran: So that’s this is a clear example of like, what’s happening.

210 00:19:52.120 00:19:55.000 Miguel de Veyra: I thought then there’s no information about it.

211 00:19:55.000 00:19:58.720 Uttam Kumaran: Okay, so that’s perfect in that. This is now not a problem of

212 00:19:58.900 00:20:03.810 Uttam Kumaran: retrieval. This is a problem of the information doesn’t exist. So 1st thing

213 00:20:05.280 00:20:09.179 Uttam Kumaran: for every single escalation they lose 30% of people. Meaning.

214 00:20:09.710 00:20:15.679 Uttam Kumaran: let me, let me ask, let me talk to let me go transfer you to a department, 30% of people are gonna be like, hang up.

215 00:20:15.680 00:20:16.280 Miguel de Veyra: Yeah.

216 00:20:16.670 00:20:22.255 Uttam Kumaran: Right. Second is their communication is either verbally Google Chats or email, right?

217 00:20:23.596 00:20:30.620 Uttam Kumaran: And so that’s this is sort of realistically like the ultimate problem we’re trying to solve is one.

218 00:20:31.610 00:20:33.950 Uttam Kumaran: it’s it’s really twofold one.

219 00:20:34.250 00:20:39.690 Uttam Kumaran: We need to improve the organization of this knowledge base.

220 00:20:40.030 00:20:43.900 Uttam Kumaran: I want to be careful to to be clear that this doesn’t mean

221 00:20:44.100 00:20:48.989 Uttam Kumaran: creating more documents doesn’t mean cleaning better documents. I don’t really know yet.

222 00:20:49.120 00:20:52.699 Uttam Kumaran: but ultimately there’s 50 documents in here.

223 00:20:53.240 00:20:59.650 Uttam Kumaran: and nobody’s using them. Even the even Janice and Yvette said, like.

224 00:20:59.800 00:21:03.320 Uttam Kumaran: most likely, that people are using 2 documents, and

225 00:21:03.510 00:21:05.219 Uttam Kumaran: I don’t even know whether they’re using them.

226 00:21:05.470 00:21:13.539 Uttam Kumaran: So we have a, we have a, very, we have a, we have a problem. Second, they have the data on the people call back rates and everything, and it’s not

227 00:21:13.790 00:21:21.910 Uttam Kumaran: really really great. And so our number one goal is to allow for the improvement of same call resolutions.

228 00:21:22.190 00:21:29.970 Uttam Kumaran: and to decrease the amount of escalations that happen. And so the way we’re going to do that is, first, st by indexing all these documents.

229 00:21:30.390 00:21:34.249 Uttam Kumaran: Second, is going to be proposing a solution on an ultimate

230 00:21:34.370 00:21:37.520 Uttam Kumaran: form to access and store this knowledge.

231 00:21:37.930 00:21:45.919 Uttam Kumaran: We have 2 problems there, though one is, we need a, we need a method by which this doc. These documents can get updated. And so this is the second problem.

232 00:21:46.550 00:21:53.900 Uttam Kumaran: Knowledge is always getting added. Right? So there’s new services. They they realize that there’s a new way to

233 00:21:54.060 00:21:58.279 Uttam Kumaran: handle the rodent annuals, and Denise or Yvette want to make an update.

234 00:21:58.400 00:21:59.610 Uttam Kumaran: Where do they go?

235 00:21:59.840 00:22:02.470 Uttam Kumaran: It is 50 documents. Where do I go to make an update.

236 00:22:02.990 00:22:11.400 Uttam Kumaran: So this is not. This is the second agent that we want to build, or if I I honestly don’t even know whether it needs to be 2 agents. Maybe we just route it or like whatever. But

237 00:22:11.880 00:22:14.679 Uttam Kumaran: the second agent really helps them update knowledge

238 00:22:15.860 00:22:29.840 Uttam Kumaran: event may say, Hey, we wanna, we wanna add a clause for our rodent annuals. Okay, what’s the clause? Great? Okay? I need this information, this more information until I can add until I can add that. Okay, I got the information cool. I added it to the knowledge base. It’s now permeated.

239 00:22:30.880 00:22:37.509 Uttam Kumaran: What that eliminates is the 2, is it? It makes a whole cycle better? Right? So knowledge gets updated. And it’s clean right.

240 00:22:37.510 00:22:38.020 Miguel de Veyra: Yeah.

241 00:22:38.020 00:22:42.049 Uttam Kumaran: If the knowledge is clean, the retrieval gets better. If the retrieval gets better than the numbers go up

242 00:22:42.250 00:22:42.730 Uttam Kumaran: right.

243 00:22:42.730 00:22:43.410 Miguel de Veyra: Yep.

244 00:22:43.410 00:23:07.550 Uttam Kumaran: The last piece of this is as that system is going, we’re gonna spit out information. This is really the reason I wanted to use vellum for the logs right, and we haven’t looked at that yet. We’ll look at that next week. And I’ll work with you guys to see how we can get it into a, into a dashboard to visualize. But one of the key things we’re gonna help them do is understand? What are the types of questions or Csrs are having right? So what are they asking? The bots

245 00:23:07.780 00:23:13.179 Uttam Kumaran: that will help them enable what’s more, documents they need to create, like what other knowledge they need to create.

246 00:23:13.780 00:23:18.710 Uttam Kumaran: And the last piece is, we’re we’re ideally, this is gonna live in Google Chat as an agent.

247 00:23:19.410 00:23:33.630 Uttam Kumaran: So Casey, very similar to how we’re building the slack agents. My proposal with them is like, I don’t want this to live separately, of course I know, for now it lives separately. That’s just like I I we need that to happen to sort of get something out for a demo.

248 00:23:33.760 00:23:36.630 Uttam Kumaran: But ideally, this lives as a Google chat agent,

249 00:23:40.110 00:23:45.010 Uttam Kumaran: and so those are the really the biggest components. There’s some more information

250 00:23:45.200 00:23:47.749 Uttam Kumaran: mean. There’s some other tools here, like, for example.

251 00:23:48.060 00:23:58.979 Uttam Kumaran: just similar. How we have notion for everything. They have a tool called Evolve, which has a bunch of information. They have a tool called Dream, which has a bunch of information. We’re gonna get the Apis from those. And I don’t. I’m gonna get that information to you guys.

252 00:24:00.490 00:24:02.810 Uttam Kumaran: But really, I think those are the

253 00:24:03.910 00:24:06.520 Uttam Kumaran: those are the number one things right.

254 00:24:07.280 00:24:13.830 Uttam Kumaran: The last piece I’ll I’ll I’ll share is one I’m gonna get some calls, some example customer calls from Yvette.

255 00:24:13.970 00:24:23.409 Uttam Kumaran: I it would be great for all of us to listen to it individually. And then I’m gonna schedule a meeting, or maybe, Miguel. It’d be great for you to schedule a meeting where we can like listen to one of those together.

256 00:24:23.680 00:24:27.049 Uttam Kumaran: because I want us to really get into the brains of the Csr.

257 00:24:27.170 00:24:34.530 Uttam Kumaran: And sort of understand like who the real customer is. Because, let’s say, our agent works really well. But it takes 2 min to get back to them.

258 00:24:34.780 00:24:39.129 Uttam Kumaran: It’s not gonna work right? And so I want us to think about what the form factor needs to be.

259 00:24:40.050 00:24:57.269 Uttam Kumaran: And then, second, we’re gonna build evals, right? And so this is the 1st agent we will have built where we have a real, consistent evals process. I want to use the real questions that they’re that Csr would ask to build the Eval data set. So as we build the agent up, we can continue to test against that data set.

260 00:24:57.961 00:25:03.589 Uttam Kumaran: and so those are really gonna be, I think, the 3 or 4 like priorities for the next 3 weeks.

261 00:25:06.200 00:25:17.180 Uttam Kumaran: I’ll pause there. This meeting is being recorded. I have all these notes. I’m gonna get this into like a project plan. So I can. So me and you, Miguel, can sort of approve it. And I can let you work towards execution.

262 00:25:17.380 00:25:20.400 Uttam Kumaran: But any questions or concerns, or like.

263 00:25:20.720 00:25:22.770 Uttam Kumaran: what about this idea or like?

264 00:25:24.010 00:25:26.509 Uttam Kumaran: Do we want to include something else like, what do you think.

265 00:25:29.862 00:25:39.529 Miguel de Veyra: Yeah. So what’s our deliverable for the 1st month at least? Like, what do we like? What’s our what’s our, you know, goalpost for success like, how do we know we succeeded.

266 00:25:39.530 00:25:45.370 Uttam Kumaran: The goalpost for success is getting an agent in in in Google Chat, that allows

267 00:25:46.300 00:25:49.520 Uttam Kumaran: Dsrs to ask questions against the data, the data set.

268 00:25:49.720 00:25:59.290 Uttam Kumaran: and the second goalpost is allowing Yvette and and Janice to update knowledge.

269 00:26:01.890 00:26:02.760 Miguel de Veyra: Okay, yeah.

270 00:26:04.780 00:26:08.169 Miguel de Veyra: Oh, include. It also includes the update already.

271 00:26:09.740 00:26:15.139 Uttam Kumaran: We want to separate those out because the Csrs should not be able to update should not be able to update.

272 00:26:16.300 00:26:18.059 Miguel de Veyra: Okay, yeah, yeah, that makes sense.

273 00:26:18.060 00:26:20.389 Uttam Kumaran: Yeah. But yeah, but the scope includes both.

274 00:26:20.750 00:26:27.950 Uttam Kumaran: I mean, I mean, I’m I wanna I wanna set the bar high. We’ll see. We’re gonna learn like, you know, we’re sprinting on this.

275 00:26:28.300 00:26:32.359 Uttam Kumaran: If we need a little bit more time, we can buy more time. But I don’t. Wanna. I don’t. Wanna.

276 00:26:32.550 00:26:35.449 Uttam Kumaran: I want to set some like lofty goals.

277 00:26:35.680 00:26:43.339 Uttam Kumaran: And we have. We have the 4 of us. Right? So ultimately, I think we can go for this.

278 00:26:45.140 00:26:49.159 Miguel de Veyra: Because I think this project will have, like more like

279 00:26:49.290 00:26:51.999 Miguel de Veyra: couple of major ports, the bot, of course.

280 00:26:52.610 00:26:55.829 Miguel de Veyra: and then the Evals should be worked on separately, I believe.

281 00:26:56.810 00:27:00.360 Miguel de Veyra: At least, so it’s not, you know, one person working on both

282 00:27:03.570 00:27:05.300 Miguel de Veyra: retrievals. Yeah.

283 00:27:05.560 00:27:09.040 Uttam Kumaran: The retrieval and an update. We have the evals.

284 00:27:09.400 00:27:09.695 Miguel de Veyra: Here.

285 00:27:15.410 00:27:21.049 Miguel de Veyra: And then the integration to Google chat because I haven’t really using Google chat.

286 00:27:28.150 00:27:31.190 Miguel de Veyra: So I think, yeah, that’s the 4 like moving parts. I believe

287 00:27:33.800 00:27:40.749 Miguel de Veyra: agent, I think, would be easy logs and data logs and evils, I think, are the same. No.

288 00:27:41.820 00:27:44.859 Uttam Kumaran: Yeah. Well, I I think I’m gonna do logs and dashboard.

289 00:27:45.460 00:27:46.540 Miguel de Veyra: Okay, okay.

290 00:28:00.960 00:28:03.059 Miguel de Veyra: yeah, yeah. This is a good question.

291 00:28:16.130 00:28:19.999 Miguel de Veyra: I think for the update. We can do it similar to

292 00:28:21.700 00:28:24.580 Miguel de Veyra: the one we did before Casey. Now, what do you think.

293 00:28:26.376 00:28:26.909 Casie Aviles: Which one?

294 00:28:27.290 00:28:31.349 Miguel de Veyra: But it’s still it’s still on Google drive with, you know.

295 00:28:31.670 00:28:37.969 Miguel de Veyra: whatever types of folders they have or the types of files they have, and then they can.

296 00:28:38.100 00:28:39.719 Miguel de Veyra: But what do you?

297 00:28:39.830 00:28:43.180 Miguel de Veyra: Oh, yeah. Cause what we want to do is the what to update it. Now.

298 00:28:44.470 00:28:52.619 Casie Aviles: Yeah, it’s going to be. I think they’ll be interfacing it through. They’re just going to talk to it right like. And then the bot will handle the updating.

299 00:28:54.100 00:28:55.080 Miguel de Veyra: Yeah.

300 00:29:02.580 00:29:05.020 Miguel de Veyra: I don’t know. I think that’s something we need to look into

301 00:29:08.590 00:29:14.830 Miguel de Veyra: the evils. What type of data do they want to see most questions? Asked one.

302 00:29:15.320 00:29:19.500 Uttam Kumaran: Yeah, I guess. Like, let’s talk about like, maybe how we want to assign these. I

303 00:29:19.830 00:29:23.500 Uttam Kumaran: I’m pretty sure I can try and take these 2

304 00:29:23.690 00:29:27.560 Uttam Kumaran: that really want you guys to focus on the agents. And this

305 00:29:30.970 00:29:32.700 Uttam Kumaran: like, how do we wanna.

306 00:29:33.400 00:29:36.290 Miguel de Veyra: I think the least is number 4 like I.

307 00:29:36.730 00:29:39.500 Miguel de Veyra: At least, that’s the like least priority.

308 00:29:39.970 00:29:46.460 Miguel de Veyra: The I think retrieval is pretty easy. The Major Block, Major Blocker here would be the update

309 00:29:46.910 00:29:51.399 Miguel de Veyra: at least. So what? That’s what I’m you know. I’m not sure if you agree, their case.

310 00:29:53.620 00:29:56.719 Casie Aviles: Yeah, I guess it’s with the update part that I’m

311 00:29:56.920 00:30:00.040 Casie Aviles: I. Still, it’s not like the implementation.

312 00:30:00.240 00:30:01.570 Casie Aviles: It’s still unclear.

313 00:30:02.900 00:30:03.270 Miguel de Veyra: Yeah.

314 00:30:03.270 00:30:03.860 Uttam Kumaran: Okay.

315 00:30:10.210 00:30:14.290 Miguel de Veyra: I mean, if there’s a probably a way we can do it where?

316 00:30:16.060 00:30:22.500 Miguel de Veyra: Where all the content are stored in a database, and then

317 00:30:22.900 00:30:37.359 Miguel de Veyra: we just there’s there’s like a normal database. And then an vector, database. And then we when we want to update it, we just update this database. Basically the normal one. And then it updates the vector, that should that could probably be it

318 00:30:38.860 00:30:42.410 Miguel de Veyra: the solution there. But yeah, we’ll see how that goes.

319 00:30:44.040 00:30:44.600 Uttam Kumaran: Okay.

320 00:30:47.100 00:30:51.309 Miguel de Veyra: Because there’s no way we can do this. If it’s in like Google Docs or something.

321 00:30:56.810 00:30:59.059 Miguel de Veyra: The evil Sultan, how do you plan to do it?

322 00:31:00.620 00:31:06.089 Uttam Kumaran: I’m gonna use. I’m just gonna build a data set and and vellum.

323 00:31:06.960 00:31:07.779 Miguel de Veyra: Oh, good.

324 00:31:13.750 00:31:17.959 Miguel de Veyra: Think, okay, integration. Yeah, let’s do the last.

325 00:31:27.720 00:31:34.630 Miguel de Veyra: So I, the other thing, I think that’s important for us to decide. Now, it’s final that we’re going to be using vellum for this project at least, right?

326 00:31:40.530 00:31:43.190 Uttam Kumaran: Yeah, I think we should just try to go with vellum short term.

327 00:31:47.840 00:31:54.840 Casie Aviles: Yeah, I can look into the Google Chat integration. But although if there’s something more immediate that you guys need help with.

328 00:31:57.540 00:32:04.209 Miguel de Veyra: Yeah, I think, Casey, let’s work on the update. Janna, you work on.

329 00:32:04.490 00:32:07.340 Miguel de Veyra: Oh, shit, it’s gonna be fun, too. Yeah.

330 00:32:09.390 00:32:12.570 Miguel de Veyra: Jana, yeah. Yeah. I think Jana could work on the

331 00:32:13.170 00:32:18.539 Miguel de Veyra: develop conversational Chatbot with integrations and everything

332 00:32:21.280 00:32:25.870 Miguel de Veyra: cause ideally, we’re gonna have 2 bots here right with them, because Csr shouldn’t be able to Update.

333 00:32:28.280 00:32:34.530 Uttam Kumaran: Yeah. So that’s why I want to have like retrieval agent, update agent.

334 00:32:34.930 00:32:36.579 Miguel de Veyra: Yeah, I think, like.

335 00:32:36.810 00:32:39.610 Uttam Kumaran: Between you and Jana working on these.

336 00:32:39.720 00:32:44.599 Uttam Kumaran: And then maybe, Casey, working on this that could be good

337 00:32:45.790 00:32:48.490 Uttam Kumaran: like, I will take these 2.

338 00:32:50.980 00:32:51.839 Miguel de Veyra: Okay. Yeah.

339 00:33:40.660 00:33:41.905 Uttam Kumaran: Okay, what else?

340 00:33:44.230 00:33:49.229 Uttam Kumaran: I think we’re good on on ABC. For now I think, Miguel, maybe me and you can.

341 00:33:49.970 00:33:54.600 Uttam Kumaran: I think ideally, Miguel, I think we should break. You should work on breaking these down into tickets.

342 00:33:55.250 00:33:56.780 Miguel de Veyra: Yeah, yeah, I’ll finish it up.

343 00:33:57.420 00:34:03.939 Uttam Kumaran: And then ideally, the evals. This will be really like ultimately, how you guys measure success.

344 00:34:09.780 00:34:15.670 Uttam Kumaran: The one I have to get access to Apis and service accounts.

345 00:34:16.550 00:34:19.060 Uttam Kumaran: What other things do we have like immediate actions.

346 00:34:22.288 00:34:24.239 Miguel de Veyra: Figure out how to do the updates.

347 00:34:27.239 00:34:29.749 Uttam Kumaran: Okay? So that’s that. That’s tickets for the agents.

348 00:34:29.750 00:34:30.870 Miguel de Veyra: That’s yeah.

349 00:34:35.010 00:34:37.029 Miguel de Veyra: Yeah. In the demo session there.

350 00:34:39.770 00:34:43.309 Uttam Kumaran: Ideally. I can go try to meet them again in person, Miguel, next week.

351 00:34:43.670 00:34:47.460 Uttam Kumaran: but I want to have, like an hour or 2 worth of like questions.

352 00:34:49.900 00:34:53.979 Uttam Kumaran: But honestly, we may not, because I feel like I got all the information we sort of needed.

353 00:34:54.409 00:34:56.790 Uttam Kumaran: I want to get some of the calls also.

354 00:34:57.720 00:34:59.870 Miguel de Veyra: Aren’t you in Argentina next week, though?

355 00:35:00.570 00:35:02.027 Miguel de Veyra: Oh, shit yeah, so never mind.

356 00:35:04.143 00:35:05.730 Uttam Kumaran: Following week.

357 00:35:08.145 00:35:13.850 Miguel de Veyra: But yeah, another thing I’m not sure how to handle is, how do we do? The meet? The client calls.

358 00:35:15.680 00:35:16.489 Uttam Kumaran: What do you mean?

359 00:35:16.940 00:35:21.150 Miguel de Veyra: Cause. We have a call with them on 1230, right like, how does that usually go?

360 00:35:23.580 00:35:26.119 Miguel de Veyra: Thing I had before was with Eddie. Right? And it’s.

361 00:35:26.120 00:35:31.220 Uttam Kumaran: So yeah, how? So one is, I think we’re gonna I want to start to use a deck.

362 00:35:31.410 00:35:35.610 Uttam Kumaran: I’m gonna try to prepare that, and then within like.

363 00:35:35.900 00:35:37.049 Miguel de Veyra: An hour. Yeah.

364 00:35:37.050 00:35:43.910 Uttam Kumaran: Yeah, ideally, like, right at like 1015, I’ll prepare. I’ll try to prepare that

365 00:35:44.377 00:35:47.229 Uttam Kumaran: but basically, I just want to have like 3 slides, which is

366 00:35:47.850 00:35:52.180 Uttam Kumaran: our team. I want to talk through the key priorities.

367 00:35:52.390 00:35:59.190 Uttam Kumaran: and then I want to show a demo of what we have. That’s it, and then just have it be as much of a discussion as possible about like the next steps.

368 00:36:00.620 00:36:05.950 Miguel de Veyra: Okay, okay. I mean, the current one is deployed so they could check it out, and they could play around with it.

369 00:36:06.120 00:36:07.680 Uttam Kumaran: Yeah, so we can walk through that.

370 00:36:12.372 00:36:14.420 Miguel de Veyra: Do you wanna try asking it now.

371 00:36:15.670 00:36:20.360 Uttam Kumaran: I have to jump to a design meeting, but I’ll poke around with it. In the meeting.

372 00:36:23.370 00:36:24.010 Miguel de Veyra: Okay.

373 00:36:24.870 00:36:34.230 Uttam Kumaran: Okay, cool. And then I know we didn’t get a chance to talk about stuff around okrs. But I think we’ll talk a little bit more in the team meeting me, and Robert will discuss.

374 00:36:36.960 00:36:37.620 Miguel de Veyra: Okay.

375 00:36:37.970 00:36:41.840 Uttam Kumaran: Okay, okay, thanks, guys.

376 00:36:42.310 00:36:43.180 Miguel de Veyra: That’s overly.

377 00:36:43.420 00:36:44.000 Uttam Kumaran: Bye.

378 00:36:44.260 00:36:44.780 Janna Wong: Thank you.

379 00:36:44.780 00:36:45.530 Casie Aviles: Thank you.