Meeting Title: Brainforge x ABC Home and Commercial: Weekly Project Check Date: 2025-04-11 Meeting participants: Uttam Kumaran, Amber Lin, Yvetteruiz, Mattburns, Scott_Harmon
WEBVTT
1 00:01:25.610 ⇒ 00:01:26.760 Amber Lin: I’ve got.
2 00:01:26.960 ⇒ 00:01:28.190 Scott_Harmon: Amber. How you doing.
3 00:01:28.370 ⇒ 00:01:35.949 Amber Lin: I’m doing good. We actually had a company meeting here in La yesterday. So I met
4 00:01:36.160 ⇒ 00:01:39.340 Amber Lin: Robert in person for the 1st time.
5 00:01:40.480 ⇒ 00:01:44.519 Scott_Harmon: Oh, good for you! That’s always kind of cool, isn’t it? I mean, it’s so many zoom
6 00:01:45.750 ⇒ 00:01:49.999 Scott_Harmon: meetings and sessions. But when you meet people face to face. It’s always kind of.
7 00:01:50.000 ⇒ 00:01:50.970 Amber Lin: You’re surprised.
8 00:01:50.970 ⇒ 00:01:52.289 Scott_Harmon: By something right.
9 00:01:52.290 ⇒ 00:01:53.923 Amber Lin: I I was very surprised.
10 00:01:54.250 ⇒ 00:01:57.810 Scott_Harmon: Oh, you’re taller, you’re shorter. Oh, you know, like they always. It’s funny how
11 00:01:58.870 ⇒ 00:02:00.940 Scott_Harmon: there’s nothing like face to face.
12 00:02:00.940 ⇒ 00:02:01.920 Amber Lin: Hi, Matt!
13 00:02:02.310 ⇒ 00:02:02.720 MattBurns: Morning.
14 00:02:02.720 ⇒ 00:02:04.510 Scott_Harmon: Speaking of good looking people.
15 00:02:05.680 ⇒ 00:02:07.200 MattBurns: How are you guys today?
16 00:02:09.860 ⇒ 00:02:13.399 Scott_Harmon: We were just talking. Amber Brainforge is an all remote company.
17 00:02:14.480 ⇒ 00:02:18.900 Scott_Harmon: And Amber’s in like a lot of tech companies, as you know, and you know Amber’s in La and.
18 00:02:19.270 ⇒ 00:02:24.269 MattBurns: But they had a company meeting in La yesterday. So you got to meet a lot of her coworkers for the 1st time.
19 00:02:24.480 ⇒ 00:02:28.240 Amber Lin: Yeah. And I’m meeting Utam today in person as well.
20 00:02:28.240 ⇒ 00:02:29.989 Scott_Harmon: You’ve never met Utah in person.
21 00:02:29.990 ⇒ 00:02:33.820 Amber Lin: No, never! He’s flying into la right now, I think.
22 00:02:34.160 ⇒ 00:02:36.020 MattBurns: Oh, okay, yeah.
23 00:02:37.280 ⇒ 00:02:42.699 MattBurns: Oh, good. Yeah, that’s interesting. I didn’t know you guys were all over the country.
24 00:02:42.700 ⇒ 00:02:46.036 Scott_Harmon: All over the globe. Very. Yeah. Lots of people in
25 00:02:46.970 ⇒ 00:02:50.930 Scott_Harmon: yeah, all over the globe. Neat, real common in our business. But.
26 00:02:50.930 ⇒ 00:02:51.690 MattBurns: Yeah.
27 00:02:52.070 ⇒ 00:02:55.490 Scott_Harmon: Presents a lot of challenges from a human resources perspective.
28 00:02:55.490 ⇒ 00:02:56.290 MattBurns: Sure.
29 00:02:57.420 ⇒ 00:02:58.860 Scott_Harmon: Well, that’s.
30 00:02:59.780 ⇒ 00:03:06.089 MattBurns: You know, and and so many companies now are are kind of moving away from the all remote model where they’re
31 00:03:06.940 ⇒ 00:03:12.490 MattBurns: moving, place back in and requiring them to come in, and so on. So that’s.
32 00:03:12.670 ⇒ 00:03:14.342 Scott_Harmon: Yeah. Yeah. Yeah.
33 00:03:15.370 ⇒ 00:03:17.520 MattBurns: We’re still allowing most of our
34 00:03:17.920 ⇒ 00:03:25.410 MattBurns: staff to to work 3 days remote, 2 in the office and 3 mode, and I think it’s a good balance for us.
35 00:03:29.330 ⇒ 00:03:34.030 MattBurns: Just seems to be working out, I mean, I know the employees like it, so.
36 00:03:34.030 ⇒ 00:03:39.930 Scott_Harmon: Yeah, when I ran my last company. That’s kind of where I was at, and the only thing I insisted on was
37 00:03:41.170 ⇒ 00:03:44.649 Scott_Harmon: some facetime each week with your peers.
38 00:03:46.810 ⇒ 00:03:49.430 MattBurns: That’s what we’re doing. We we
39 00:03:49.540 ⇒ 00:03:53.419 MattBurns: like in person meetings like you, said Scott. So we actually had
40 00:03:54.190 ⇒ 00:04:00.229 MattBurns: our monthly company meeting yesterday, where he had all the office staff in, and we feed them lunch, and
41 00:04:00.610 ⇒ 00:04:08.379 MattBurns: seems to seems to really work well, because a lot of those folks sometimes don’t even see each other.
42 00:04:08.790 ⇒ 00:04:13.790 MattBurns: because they alter different days, and we have a few satellite offices at Bastrop and
43 00:04:13.970 ⇒ 00:04:23.809 MattBurns: Marble Falls, and these are the surrounding towns. Amber that we have some satellite offices. So those folks don’t even see each other all that often. So
44 00:04:24.180 ⇒ 00:04:30.750 MattBurns: yeah, yeah, we did at that meeting, guys. We did roll out to everybody
45 00:04:31.523 ⇒ 00:04:38.049 MattBurns: in terms of the all the office staff that we were working with Brainforge, and what the progress was and where we were.
46 00:04:38.410 ⇒ 00:04:43.359 MattBurns: and actually showed some live examples of
47 00:04:44.060 ⇒ 00:04:48.950 MattBurns: asking the AI bot a question, and it answered it
48 00:04:49.420 ⇒ 00:04:53.859 MattBurns: within about 5 seconds, and we heard a lot of Oo’s and ahs! When that happened. So
49 00:04:54.040 ⇒ 00:04:59.929 MattBurns: it was good. So that was kind of cool, and I think everybody realized.
50 00:05:00.820 ⇒ 00:05:06.150 MattBurns: Yes, this can put the protocols.
51 00:05:06.540 ⇒ 00:05:11.440 MattBurns: the services, everything at your fingertips, so that.
52 00:05:11.850 ⇒ 00:05:17.280 MattBurns: as Stephen put it, because Steven kind of did the intro to the to the meeting.
53 00:05:17.810 ⇒ 00:05:21.530 MattBurns: you know, you can concentrate on the customer. You don’t have to be thinking.
54 00:05:21.890 ⇒ 00:05:30.260 MattBurns: you know, all this other stuff. It just pops it right in front of you. So it actually the event from my perspective, I thought it went well. What did you get any feedback and such.
55 00:05:30.750 ⇒ 00:05:57.010 YvetteRuiz: Yes, I I complimented Steven, because Steven did a great job of of presenting to presenting it. But I did. I got my team members came up, saying, Are we next to be to get this done? Event? I had the billing department come to me and tell me, hey, this is great stuff. You know, I’m really excited about that. And then we had some service managers that said the same. So it was. It was a lot of positive that came out of that for sure.
56 00:05:57.010 ⇒ 00:06:01.309 Scott_Harmon: Really like the way you said that, Matt, that
57 00:06:02.330 ⇒ 00:06:07.590 Scott_Harmon: you can increase the customer, the focus on the customer, and this the
58 00:06:08.180 ⇒ 00:06:17.040 Scott_Harmon: the skills it takes to keep cut, you know, to to work with customers and kind of the diplomacy and all that kind of stuff, and have
59 00:06:17.360 ⇒ 00:06:24.830 Scott_Harmon: the right technical information more at your fingertips. That’s a really good way, I think, to think about
60 00:06:25.550 ⇒ 00:06:27.740 Scott_Harmon: partnering AI with people.
61 00:06:29.130 ⇒ 00:06:36.689 MattBurns: Agreed. It’s interesting. Our Hr. Director and I were chatting on a different subject. And she said she saw was listening to a podcast or something that
62 00:06:37.798 ⇒ 00:06:45.110 MattBurns: was involving customer service. I didn’t even tell you this event, but she basically said that the gist of it was
63 00:06:45.860 ⇒ 00:06:50.110 MattBurns: you have. You could have a standalone AI,
64 00:06:50.400 ⇒ 00:06:56.299 MattBurns: or you could have an a live person with an AI assistant, if you will, and
65 00:06:56.440 ⇒ 00:06:59.660 MattBurns: you know the combination of here was was a better
66 00:07:00.419 ⇒ 00:07:07.810 MattBurns: customer service experience, because they’re they were utilizing it kind of in the manner that we’re looking at that
67 00:07:08.050 ⇒ 00:07:11.039 MattBurns: it’s it’s there to help. It’s there to provide all the
68 00:07:11.510 ⇒ 00:07:15.310 MattBurns: technical info whatever. Yet you still got the
69 00:07:15.470 ⇒ 00:07:18.950 MattBurns: human voice that’s got the empathy, and so on and so forth.
70 00:07:18.950 ⇒ 00:07:24.370 Scott_Harmon: You, and the fullest of time. I think I think we will see both. I personally.
71 00:07:25.050 ⇒ 00:07:29.729 Scott_Harmon: with what I’m learning about ABC. Culture and brand prefer the human
72 00:07:30.210 ⇒ 00:07:34.140 Scott_Harmon: in the loop model. I think it plays to your strengths, and I think it.
73 00:07:34.460 ⇒ 00:07:42.799 Scott_Harmon: You know the with what I’ve learned just about your culture. It’s better to try and make some personal connections where that’s
74 00:07:43.650 ⇒ 00:07:48.679 Scott_Harmon: even them in. Even in those down periods and customers, when they’re mad, you know, you can keep them if they
75 00:07:49.240 ⇒ 00:07:54.779 Scott_Harmon: if they have a trusting relationship with the company. So I kind of like that model for you guys. But.
76 00:07:55.520 ⇒ 00:07:56.069 MattBurns: For sure.
77 00:07:56.070 ⇒ 00:08:04.520 YvetteRuiz: For sure, for sure. Steven is in another meeting today. Janice is out today. So I guess it’s just us. Okay.
78 00:08:04.520 ⇒ 00:08:06.360 Amber Lin: I’ll give us.
79 00:08:06.360 ⇒ 00:08:15.500 YvetteRuiz: Thank you, Amber. Thank you. Thank you. Thank you for putting that presentation together. We had the contest. That was super fun as well. We got a winner.
80 00:08:16.755 ⇒ 00:08:18.374 YvetteRuiz: I kinda it was
81 00:08:18.780 ⇒ 00:08:19.810 Amber Lin: I can share my screen.
82 00:08:19.810 ⇒ 00:08:21.018 YvetteRuiz: Yeah, if you could.
83 00:08:21.320 ⇒ 00:08:24.290 MattBurns: 3rd choice whatever that was. Yeah.
84 00:08:24.670 ⇒ 00:08:28.469 YvetteRuiz: Was bummed. Steven was bummed. We didn’t pick his.
85 00:08:28.850 ⇒ 00:08:31.330 MattBurns: Yeah, it’s the actually, the second one on the screen.
86 00:08:31.330 ⇒ 00:08:32.049 Scott_Harmon: Oh, that’s fun!
87 00:08:32.059 ⇒ 00:08:32.569 Amber Lin: Oh!
88 00:08:32.570 ⇒ 00:08:33.230 Scott_Harmon: Yeah, yeah.
89 00:08:33.580 ⇒ 00:08:34.280 YvetteRuiz: My favorite.
90 00:08:34.280 ⇒ 00:08:37.029 Amber Lin: My favorite 2, that these 2 were Stevens, for.
91 00:08:39.596 ⇒ 00:08:40.830 Scott_Harmon: Too bad!
92 00:08:40.830 ⇒ 00:08:47.439 YvetteRuiz: We told Steven. Maybe if we dress it up like we talked about it having different costumes or something, we may choose his down the line.
93 00:08:48.963 ⇒ 00:08:49.906 YvetteRuiz: Okay.
94 00:08:51.140 ⇒ 00:08:56.110 Scott_Harmon: Oh, yeah, Amber, can we drop that logo into the ux somewhere?
95 00:08:56.520 ⇒ 00:09:00.146 Amber Lin: Oh, totally! Now that we have the winner
96 00:09:00.800 ⇒ 00:09:06.849 Amber Lin: I’ll I think I can ask my team. If not, I’ll ask Tim. It will be pretty quick.
97 00:09:07.100 ⇒ 00:09:09.550 Scott_Harmon: Yeah, I think you just probably just drop right in.
98 00:09:09.550 ⇒ 00:09:10.910 MattBurns: Yeah, excellent.
99 00:09:11.200 ⇒ 00:09:13.890 Amber Lin: Let me start this slideshow.
100 00:09:14.890 ⇒ 00:09:17.770 Amber Lin: Good morning, everybody. Happy. Friday.
101 00:09:18.290 ⇒ 00:09:19.030 YvetteRuiz: Alrighty!
102 00:09:20.570 ⇒ 00:09:39.480 Amber Lin: So today we have a quite a few things to go over. So we’ll look at again at Andy’s performance this week. We’re wrapping up the rollout to 5 Csr, so we’ll look at lessons learned. And then we’ll look at. Look at the full rollout plan which I already sent you guys via email. So if you guys want to go back and reference it, it’ll be there.
103 00:09:39.480 ⇒ 00:09:51.539 Amber Lin: And we’ll look at the oh, by the way, measures as we talked about last week. I think that’s very important, and some really good news from the data team, and then I’ll hand it over to Utam and Scott to talk about the phase. 2. Proposal.
104 00:09:53.000 ⇒ 00:09:54.460 Amber Lin: So 1st of all.
105 00:09:54.780 ⇒ 00:10:09.260 Amber Lin: this is a screenshot of the last 2 weeks, and continuing from last week, we see that the total exchanges is going up a lot. We see that quality score is going up, and then the error rate is going down
106 00:10:09.300 ⇒ 00:10:26.059 Amber Lin: the average execution. Time is going went up a little bit because open AI’s model was down, and then we switched it really fast. So that was a outlier in a time, just so that if you’re looking at it, our bot is fine. Open. AI!
107 00:10:28.550 ⇒ 00:10:32.269 Amber Lin: So this is the uses this week.
108 00:10:32.745 ⇒ 00:10:53.180 Amber Lin: Joy! Again taking up the leaderboard number one Janice also tested it a lot. And then we can also see we have a lot of other people testing it, too. Yusuf is finally using it. I’m very, very happy. And we also see other people from the exact team also using it. So I’m very happy to see that
109 00:10:53.680 ⇒ 00:11:19.049 Amber Lin: in comparing this week from last week we can see that we definitely have a lot more exchanges. From 50 to 8, almost 90 exchanges. We have higher quality scores. We have a lot lower error rate. So even just from this 1, 2 weeks of testing, we have improved a lot. So I think it really fulfills our goal of why we even did the testing with the 5 Csrs. In the 1st place.
110 00:11:19.230 ⇒ 00:11:19.790 YvetteRuiz: Yeah.
111 00:11:20.160 ⇒ 00:11:41.470 YvetteRuiz: hey, Amber? I can. I ask a real quick question? I know you guys meet weekly but Janice was mentioning that when she’s doing some searches kind of what I would. I had sent an email on some of the feedback that I received from the new hires is when they are searching. The documents are on there if they’re on there, but they’re not populating, or they’re giving something like.
112 00:11:41.910 ⇒ 00:11:50.740 YvetteRuiz: so I don’t know if she’s mentioned it to you. Regarding it’s not locating the info. Sometimes it’s not locating the documents in there
113 00:11:51.030 ⇒ 00:11:52.720 YvetteRuiz: that we do have in there with.
114 00:11:53.260 ⇒ 00:12:22.959 Amber Lin: Yeah, yeah. We actually in our daily meetings, we Janice daily a lot of these. We find it in the Central Doc. But then there’s somehow the wording is a little bit off, and that’s why that’s where me and Janice sort of work together to test work. What works best for the for the bot, and usually we find a resolution, if not, we market, and then we escalate it to the team, and they do more in depth changes in the bot behavior.
115 00:12:24.000 ⇒ 00:12:31.229 YvetteRuiz: Okay, alright. So we are looking into. And that’s kind of what we’re. I was my concern early on. Was it the wording? Was it the way we’re asking it. That’s why
116 00:12:31.660 ⇒ 00:12:39.380 YvetteRuiz: populating. So okay, so I’m just, I was just curious about that. What we ended up finding because she was talking about that yesterday.
117 00:12:39.380 ⇒ 00:12:46.670 Amber Lin: Totally I’ll talk about it in a in a quick bit. But this is the Csr feedback. From
118 00:12:47.270 ⇒ 00:13:00.147 Amber Lin: from this testing period not all of them has filled it out. But overall we have a 4 out of 5 rating, which I think is pretty good, and because there is room for improvement. If we’re 5 out of 5, then we’re done
119 00:13:01.850 ⇒ 00:13:22.660 Amber Lin: as you also mentioned. This is Joy’s feedback, she said. Annie’s very apt at and at easy and quick answers, but does not go into enough depth about topics doesn’t specify specific instructions or a list pricing. And so that’s something that. Oh, now she’s brought it up. Oh, okay, we’re gonna look at it.
120 00:13:22.660 ⇒ 00:13:26.699 Scott_Harmon: Can I? On that particular point? Amber, is there a.
121 00:13:27.840 ⇒ 00:13:37.270 Scott_Harmon: you know? I know we we changed the prompting behavior so that it would give more summary, shorter answers like Early On, it, was being too verbose, and gave too.
122 00:13:37.270 ⇒ 00:13:38.159 Amber Lin: Much data.
123 00:13:38.380 ⇒ 00:13:45.029 Scott_Harmon: So some of this could just be training. Which is, she can ask a follow up question, saying, Give me more detail.
124 00:13:45.130 ⇒ 00:13:51.940 Amber Lin: And you know, maybe we just need to train them like, Hey, just tell Andy you want more detail, and it’ll come back and.
125 00:13:52.440 ⇒ 00:13:55.290 Scott_Harmon: You know. Maybe they just don’t know that, you know. They just think that.
126 00:13:55.290 ⇒ 00:14:08.150 Amber Lin: Yeah, totally. I think that aligns with the point that Yvet brought up as well. Of how do we ask them questions? Because a big part about AI is prompt engineering like it’s a separate job. So we do have to train everybody.
127 00:14:08.150 ⇒ 00:14:17.630 Scott_Harmon: It’s funny a lot of people don’t know. I’m sure you guys use this a lot amber. You use it all the time. You have to be really bossy with AI and tell it exactly how you want it to behave.
128 00:14:18.380 ⇒ 00:14:26.989 Scott_Harmon: You know, be more detailed, you know, or be more succinct, or give me more details on price, like you can just really boss it around.
129 00:14:28.210 ⇒ 00:14:31.249 Scott_Harmon: Sometimes I think people are a little too polite to it, so maybe.
130 00:14:32.130 ⇒ 00:14:35.379 Amber Lin: I think they they still feel like they’re talking to a person.
131 00:14:35.650 ⇒ 00:14:46.429 Scott_Harmon: Yeah, like, Oh, thank you very much, you know. But anyway, on this particular issue, I don’t know if that it’s just a matter of whether I’m I’m guessing. Andy would give more detail on these topics if asked. So
132 00:14:46.800 ⇒ 00:14:48.940 Scott_Harmon: that could just be a user training.
133 00:14:48.940 ⇒ 00:14:58.869 Amber Lin: Yeah. And we have all the detail there. So it’s just a matter of okay, what format works best for the reps. And through these feedbacks. We’ll know a lot better as we go.
134 00:14:59.280 ⇒ 00:15:00.650 Scott_Harmon: That’s great feedback, though. Yeah.
135 00:15:00.650 ⇒ 00:15:04.549 MattBurns: Hey? That! How? How long has joy been with us?
136 00:15:04.810 ⇒ 00:15:06.070 MattBurns: Just curious.
137 00:15:06.200 ⇒ 00:15:13.400 YvetteRuiz: She’s been with us. Probably this is her 3rd 4th week, Matt. Maybe a month. Just just yeah. She’s a.
138 00:15:13.760 ⇒ 00:15:14.530 MattBurns: Okay.
139 00:15:15.170 ⇒ 00:15:16.040 YvetteRuiz: I’m sorry.
140 00:15:16.950 ⇒ 00:15:20.020 MattBurns: So how’s how’s she doing overall? She pretty sharp.
141 00:15:20.020 ⇒ 00:15:48.769 YvetteRuiz: Oh, my goodness, yes, she is. I I gave her credit when I gave my feedback I mean her. The way she’s thinking is exactly the way we want everyone thinking. I mean, she came to the table with a lot of feedback when I met with her, but she was like, you bet, before I attempt to try to go reach Kenny or my train. I go here first, st and I start asking questions, so she’s extremely sharp. She’s been the one that has been giving us the most, as you can see by her asking questions and stuff. But she’s gonna yeah. She’s really good.
142 00:15:48.770 ⇒ 00:15:49.369 Scott_Harmon: That’s great!
143 00:15:49.700 ⇒ 00:15:54.719 MattBurns: That’s interesting. Yeah. And I I agree with what you said, Scott. I think
144 00:15:54.910 ⇒ 00:16:00.669 MattBurns: it’s interesting. You don’t in one respect you want a short, succinct answer
145 00:16:00.970 ⇒ 00:16:13.949 MattBurns: from the AI bot. But then you could also just say, Give me more detail and boom, then it’ll cause you what you don’t want is so much detail at 1st that it’s overwhelming. So I I like what it does in terms of
146 00:16:14.430 ⇒ 00:16:22.970 MattBurns: give me the quick answer with limited real detail, and then, if I want more detail, ask for it.
147 00:16:24.180 ⇒ 00:16:31.160 Scott_Harmon: And like I said, some of that could just be training her. There are some things you can do that are really easy
148 00:16:31.320 ⇒ 00:16:35.959 Scott_Harmon: where you can tell Andy. And again, you probably see this as you use AI bots
149 00:16:36.120 ⇒ 00:16:42.130 Scott_Harmon: at the end of its answer. It could go. Would you like more detail on this like it can ask kind of a hanging question.
150 00:16:42.595 ⇒ 00:16:45.520 Scott_Harmon: You know all the normal Chatbots do that.
151 00:16:46.380 ⇒ 00:16:50.839 Scott_Harmon: And they’ll even maybe suggest what they often do is they’ll suggest, follow up questions.
152 00:16:51.340 ⇒ 00:16:54.770 Scott_Harmon: Would you like me to show you more detail on pricing. Or would you like me to
153 00:16:54.930 ⇒ 00:16:57.560 Scott_Harmon: give you more about the territory assignment, you know.
154 00:16:57.810 ⇒ 00:17:09.069 Scott_Harmon: just to almost prompt them on what a follow up could be. That’s a i’ll say, trivial speaking for amber. Those are very easy things to get Andy to do so. A lot of that, I think, is just
155 00:17:10.170 ⇒ 00:17:13.559 Scott_Harmon: in this exact kind of feedback from her. Yvette.
156 00:17:13.680 ⇒ 00:17:17.770 Scott_Harmon: just go. Okay, how do we want to use it? How do we want Andy to interact.
157 00:17:18.270 ⇒ 00:17:22.559 Scott_Harmon: you know? Train her to use it better, and then give us hints about how to have Andy.
158 00:17:22.900 ⇒ 00:17:23.670 YvetteRuiz: Yeah.
159 00:17:23.670 ⇒ 00:17:24.020 Scott_Harmon: Convert.
160 00:17:24.020 ⇒ 00:17:50.360 YvetteRuiz: No, for sure. That was one of the things like after meeting with them, because I just met with them this week because we weren’t getting the feedback the 1st feedback. So I went ahead and jumped in. And so that was my goal is like, now we gotta really start shifting the more that we on board. We want to make sure that we are working with our agents. So then that way, they start feeling more comfortable. Going in there and asking it and really getting in there and stuff. So we will start focusing a lot.
161 00:17:50.360 ⇒ 00:17:54.580 Scott_Harmon: Again. Andy will just kind of evolve in its conversational
162 00:17:54.730 ⇒ 00:18:02.230 Scott_Harmon: behavior. We’re going to get to in a minute. The oh, by the ways! But the oh, by the ways are part of what’s called a wrap up a Csr wrap up so.
163 00:18:02.940 ⇒ 00:18:05.210 Scott_Harmon: In all Csr scripting.
164 00:18:05.410 ⇒ 00:18:11.119 Scott_Harmon: As obviously, you guys know, there’s always a set of things to be done at the end of a call.
165 00:18:11.559 ⇒ 00:18:19.010 Scott_Harmon: Can I help you with this? Would you like more detail about that? Sometimes there’s a survey you stick in there, you know. There’s just all kinds of stuff you put at the end
166 00:18:19.160 ⇒ 00:18:26.950 Scott_Harmon: so we can have Andy do more and more of that. Oh, by the way, fits perfectly, you know. Would you like more detail.
167 00:18:27.400 ⇒ 00:18:31.810 Scott_Harmon: It’s 1 of those things that customers really like when the Csr comes back and says.
168 00:18:32.470 ⇒ 00:18:34.889 Scott_Harmon: You know, can I give you additional stuff? So
169 00:18:35.240 ⇒ 00:18:40.289 Scott_Harmon: that’s that’s a real easy thing for us to add in this behavior, as as we keep
170 00:18:40.820 ⇒ 00:18:44.150 Scott_Harmon: just understanding how Csrs, you know, want to use it.
171 00:18:45.260 ⇒ 00:18:46.840 Amber Lin: Yeah, cool.
172 00:18:46.840 ⇒ 00:18:47.410 YvetteRuiz: Thank you.
173 00:18:49.450 ⇒ 00:18:50.780 Amber Lin: Next slide.
174 00:18:51.730 ⇒ 00:19:21.239 Amber Lin: And so this is kind of example of how we adapted Andy based on the feedback and adjusted the document so that it’s best for Andy to read, because we read differently than a computer. So, as mentioned before, we identified a gap of, there’s no pricing information. You also sent me this when you talk to the Csrs. And so before we only have a short response. When we even when we asked for the cost.
175 00:19:21.740 ⇒ 00:19:43.910 Amber Lin: And so, after we adjusted that, we kind of use our trader Bot as still in development, and it gave us a very clear structure of the prices based on the square footage, the type of houses. And then, when we asked again of Okay, we have this customer. That’s a term my customer. Actually, they have a crawl space and they have a square footage of 2,000.
176 00:19:44.010 ⇒ 00:19:53.009 Amber Lin: What’s the price? And then Andy gave a much more detailed response. There. So this is an example of how we’re able to adapt them. As long as we have feedback.
177 00:19:53.180 ⇒ 00:20:09.211 YvetteRuiz: Absolutely, absolutely. I loved it. When you guys came back with that because that was this is a common question that is asked, you know, because people, you know whether they’re refinancing their home. They’re buying a home. They need to get a price in it. So there’s qualifying questions. And of course we’ve got to get to the pricing, because these are the ones that do so. That was a really good one that they shared.
178 00:20:09.420 ⇒ 00:20:20.579 Scott_Harmon: I hate to keep cheerleading, but you know this is exactly the design we hope for right what way back, when you were thinking about. You know, knowledge bases and stuff. This kind of information knowledge loop
179 00:20:21.250 ⇒ 00:20:24.510 Scott_Harmon: where you or Andy identify a gap.
180 00:20:25.080 ⇒ 00:20:26.600 Scott_Harmon: The trainer bot.
181 00:20:26.860 ⇒ 00:20:38.540 Scott_Harmon: you know, creates. You make sure you get the information accurate in the in the knowledge base. And now the answer just gets better automatically. But the the knowledge is in sync.
182 00:20:38.850 ⇒ 00:20:42.860 Scott_Harmon: right? It’s like always the most recent knowledge, and you’ve almost got an organic
183 00:20:43.500 ⇒ 00:20:46.640 Scott_Harmon: way for your knowledge to evolve where you don’t
184 00:20:46.830 ⇒ 00:20:54.399 Scott_Harmon: have all this knowledge laying around the company that’s sort of old that nobody refers to right, and so.
185 00:20:54.840 ⇒ 00:21:03.359 Amber Lin: And I saw the old Wdi document. It was. It was very messy. So me and Janice work together with the bot to put.
186 00:21:03.360 ⇒ 00:21:09.530 Scott_Harmon: Yeah, that’s the that’s the problem we set out to fix was having knowledge, that is, is
187 00:21:10.750 ⇒ 00:21:21.220 Scott_Harmon: relevant and used in the business versus in a dusty room somewhere, you know, somewhere, and so that this is just a terrific example of that really exciting.
188 00:21:22.810 ⇒ 00:21:24.652 Amber Lin: That makes me feel so happy.
189 00:21:25.600 ⇒ 00:21:27.850 Amber Lin: I wish Denise was here because this is both.
190 00:21:28.162 ⇒ 00:21:31.100 Amber Lin: This is awesome, all of her effort, too.
191 00:21:31.860 ⇒ 00:21:33.260 Amber Lin: Okay.
192 00:21:33.600 ⇒ 00:21:51.290 Amber Lin: so just to wrap quickly, wrap up the rollout. I think it went really well. Everybody was using the bot. It gave the queries back under 10 seconds we locked all the gaps and we updated them, most of them and the other ones
193 00:21:51.860 ⇒ 00:22:07.040 Amber Lin: we’re finding actively find a solution to. And then we also have the baseline set. I’ll update you guys on a data team updates in a bit. And then we did get a 4 out of 5 score. So yes, 80% of them.
194 00:22:07.966 ⇒ 00:22:09.849 Amber Lin: So right on the line.
195 00:22:10.940 ⇒ 00:22:31.170 Amber Lin: And so here’s our timeline. We’re at the end middle of April, and we’re gonna start the full deployment. And on the other side we are training the trainer system. But, as you’ve seen, and we’re hoping to pilot at least have a few test users by the end of next week or in the next sprint.
196 00:22:33.760 ⇒ 00:22:34.660 Amber Lin: So
197 00:22:35.990 ⇒ 00:23:04.550 Amber Lin: I just wanted to address a few lessons learned, and I know everybody has. There is brewing your mind already, of how do we make it? Even better for the next next part? And we talked about as of how do we reinforce ideal behavior. Like joy of actively using it, going to it 1st before you go to your supervisor. And I was thinking, maybe if we have shout outs, we have small rewards, have them share the tips
198 00:23:05.020 ⇒ 00:23:17.540 Amber Lin: to reward those high usage people, and then reinforce that behavior. And then for lower usage reps. We can have check ins and have buddy systems to make sure that they get on track too.
199 00:23:18.350 ⇒ 00:23:46.640 YvetteRuiz: I really like this. Yesterday I did recognize her personally, and the on the other few, particularly her worth more. But I did give her a gift card, I mean, she was super thankful the other group was thankful. They were kind of like, what is this for? And I was like, you know, your feedback is valuable. I mean, this is only going to be as successful as you guys make it. You know what I mean. So again getting their buy in and stuff like that so absolutely love the cause. I know you shot this over to me, but I did walk around yesterday and give them the the.
200 00:23:46.640 ⇒ 00:23:47.750 Amber Lin: Thank you.
201 00:23:47.750 ⇒ 00:23:48.435 YvetteRuiz: Cards.
202 00:23:49.120 ⇒ 00:23:54.919 Amber Lin: Thank you that that’s gonna help all of us tremendously. So thank you for taking the.
203 00:23:56.005 ⇒ 00:23:57.090 YvetteRuiz: Absolutely.
204 00:23:57.090 ⇒ 00:24:19.169 Amber Lin: Yeah. And then also on. Because we mentioned Joseph didn’t have didn’t use it because he didn’t have access. So we wanna make sure that we have a dedicated owner to check access. And we have a centralized sheet that we check it off. Because now we’re gonna have 25 people. We really need to make sure everybody has it. So we don’t get stuck there.
205 00:24:19.520 ⇒ 00:24:42.360 YvetteRuiz: Yes, agreed 100, and that’s goes back to the conversation I was having, which is like, now we’re going to have to have check in systems, you know, because it was something very simple when you talk to you so he’s like, Oh, I didn’t, you know. No, but it’s just a matter. We have to give them the nudge. They get busy. And you know they go back to their own behaviors. Let me just ask the question versus, let me ask this. So we’re gonna work on that for sure. Amber cool.
206 00:24:42.360 ⇒ 00:24:56.009 Amber Lin: Totally. And we also want to encourage more feedback. So that’s how do you guys want to structure the weekly rituals? How do we have positive reinforcement. I’m thinking that we could
207 00:24:56.230 ⇒ 00:25:15.019 Amber Lin: for our end. Also, we can tell them, hey, we fix this based on your feedback and that that’s sort of a reinforcing behavior for the agents to actually give feedback. So I’ll coordinate with Janice. Of how do I? How do I share to the team that. Okay, we’ve changed this based on your feedback to make them feel good about it.
208 00:25:15.350 ⇒ 00:25:17.081 YvetteRuiz: Yeah, cause then they’ll feel heard.
209 00:25:17.370 ⇒ 00:25:41.429 Amber Lin: Yeah, totally. And lastly, I want to minimize the impact of errors because, we can talk about this later. But I know, Eva, you brought up this. I think I’m not sure if you also got the wrong response from us, or from just searching up documents. But I know that wrong answers will always have an impact, and I want to find a way to reduce that as early as possible.
210 00:25:42.120 ⇒ 00:26:10.379 YvetteRuiz: Yeah, no. I do. One of the things that we are gonna work on amber is some of these that are service related tech related. I do now. Starting next week, I have a meeting with the DM. Of the division, because these are the ones that are conflicting information. And Steven and I are gonna work to make sure that. Okay, is this correct? Is it not? So? Then that way we’re feeding it. And that’s where the miscommunication came on versus what we had. But then, what a service manager was telling us.
211 00:26:10.380 ⇒ 00:26:15.299 Scott_Harmon: So again, just hugely want to reinforce what you said.
212 00:26:15.810 ⇒ 00:26:22.370 Scott_Harmon: That’s an opportunity to use the trainer bot to when where there’s a conflict inside the company
213 00:26:22.610 ⇒ 00:26:32.720 Scott_Harmon: about the accurate, you know. Source of truth that you can say, Hey, look! We want to use Andy. We got to get the right knowledge in there. Let’s agree on this folks, and
214 00:26:32.930 ⇒ 00:26:36.989 Scott_Harmon: that’s that’s a way to get consensus using the trainer bot. You know, we could.
215 00:26:37.110 ⇒ 00:26:43.550 Scott_Harmon: As we roll the trader. Bot, you can actually put some things in it to force people to agree.
216 00:26:44.110 ⇒ 00:26:51.339 Scott_Harmon: Yes, this is the accurate pricing stuff like you could have Steven and Fred review it before, like, okay.
217 00:26:51.980 ⇒ 00:27:00.070 Scott_Harmon: you know, this is the pricing stuff. Do we agree? And then the trainer Bot publishes it in the knowledge base like. So there are things we can do
218 00:27:00.290 ⇒ 00:27:03.440 Scott_Harmon: as you discover, these information gaps
219 00:27:03.760 ⇒ 00:27:10.869 Scott_Harmon: that to kind of get rid of them, you know, one at a time. That’s that’s a real high value stuff.
220 00:27:12.300 ⇒ 00:27:32.969 YvetteRuiz: Well, that next week I mean, I have. I already have a list of stuff because Janice started mentioning me and setting those things aside that really need the next level approval. So I’ve gone through them. I’ve looked at some stuff through our agreement. And now I just need to partner up with the division manager to kind of, okay, are we? Are we on the same page with what we’re saying? So then that way we provide the.
221 00:27:32.970 ⇒ 00:27:41.740 Scott_Harmon: Yeah. And so not. This isn’t the right form to feature Engineer Amber. But what I would recommend you consider is the trainer bot
222 00:27:42.030 ⇒ 00:27:46.440 Scott_Harmon: publishing a draft knowledge. Note
223 00:27:47.300 ⇒ 00:27:59.659 Scott_Harmon: that then the trainer, or whomever can then circulate internally to say, you know, before it gets published to the real knowledge base. So it’s just like draft. And they can say, you know, email to Matt, or whomever and go.
224 00:28:00.120 ⇒ 00:28:03.120 Scott_Harmon: you know, comments and
225 00:28:03.660 ⇒ 00:28:09.690 Scott_Harmon: the comments can go even right back to the trainer bot like. No, I think you missed San Antonio, whatever.
226 00:28:09.910 ⇒ 00:28:16.509 Scott_Harmon: and by having a draft, then you can make sure people agree before it then gets pushed.
227 00:28:17.300 ⇒ 00:28:23.270 Scott_Harmon: And it’s just a real nice, simple way to do a workflow to get alignment. So one thing to consider is just a a draft.
228 00:28:23.270 ⇒ 00:28:27.080 Amber Lin: Cool step. I’ll write it down. Thank you, Scott.
229 00:28:28.040 ⇒ 00:28:32.339 Amber Lin: Here. So next up is the full rollout.
230 00:28:33.170 ⇒ 00:28:41.970 Amber Lin: We have this document. I’ll send this to you guys. But essentially, we want to have them using it by Monday. So
231 00:28:42.735 ⇒ 00:29:01.650 Amber Lin: I’m not sure if it cause you say you want to also give it to the division managers as well. Would you be able to send me a list of their emails, or you can also just send it directly to Tim. And I. I’m gonna send or I’m gonna ask him to give all the Csrs access today.
232 00:29:02.030 ⇒ 00:29:09.675 YvetteRuiz: Yes, I have it on my to do list. Today. I had printed out all the emails that you sent me that I haven’t gotten back to you. So that’s on my to do list today.
233 00:29:10.300 ⇒ 00:29:21.439 YvetteRuiz: I wanted to our branch managers. They they can provide some good feedback, and then Steven will definitely the division manager for pest because we’re going to be meeting with them next week, so I’ll send you their information so we can.
234 00:29:21.440 ⇒ 00:29:35.130 Amber Lin: Awesome, and I’ll I’ll ask him to give the Css. Csrs access first, st just because it takes a little bit longer. For for it to take effect, so I’ll get it as soon as possible.
235 00:29:35.530 ⇒ 00:29:38.512 YvetteRuiz: Matt, do you want access to.
236 00:29:39.010 ⇒ 00:29:39.790 MattBurns: Sure.
237 00:29:39.790 ⇒ 00:29:40.565 YvetteRuiz: Okay.
238 00:29:41.520 ⇒ 00:29:43.380 YvetteRuiz: We’ll add Matt to it.
239 00:29:45.460 ⇒ 00:30:01.849 Amber Lin: So this is just a quick overview. I have all of this in the document. So what we kinda wanna achieve for the full rollout, and here’s just a quick checklist of what we want to do. I think we have everything except for the access, and once we have that we’ll be good to go.
240 00:30:02.020 ⇒ 00:30:05.060 Amber Lin: And so
241 00:30:05.160 ⇒ 00:30:11.139 Amber Lin: to be conscious of time, we’re halfway through. I want to update you guys on the oh, by the way, measures. So
242 00:30:11.600 ⇒ 00:30:29.460 Amber Lin: Annie did a fantastic job. She’s the new addition to our data team, and she’s working with Brian very closely to get all the data in place to have all the measurements to improve our dashboards. So this is what we talked about last week. We wanted to look at. Oh, by the ways! So now we have
243 00:30:29.640 ⇒ 00:30:31.619 Amber Lin: the number of by the ways
244 00:30:31.950 ⇒ 00:30:37.159 Amber Lin: through time, and we have the division, the category of oh, by the ways.
245 00:30:37.290 ⇒ 00:30:50.300 Amber Lin: So now we know, okay, what are we pushing? Do we need to push this more? Are we actually pushing what’s in season? Because last time we wanted to push exterior cleaning window cleaning. So this will be really helpful.
246 00:30:50.500 ⇒ 00:30:54.789 Amber Lin: and we’ll also start to review this in our weekly meetings, going forward.
247 00:30:54.790 ⇒ 00:30:57.560 Scott_Harmon: What’s the definition of used? Oh, by the way.
248 00:30:57.560 ⇒ 00:30:58.959 Uttam Kumaran: Yeah? The same question.
249 00:30:59.280 ⇒ 00:31:00.610 Amber Lin: Oh, so that’s
250 00:31:01.040 ⇒ 00:31:13.749 Amber Lin: because when we have the answer from the bot, it doesn’t always do. Oh, by the way, because it will make it very, very long. We can adjust the wording, but that’s essentially the percentage of answers that had to upsell.
251 00:31:17.720 ⇒ 00:31:19.770 YvetteRuiz: So are you saying like.
252 00:31:20.330 ⇒ 00:31:39.269 YvetteRuiz: I’m sorry. So I know when we when we did some searches amber like we did yesterday, we did one for the testing that we when we were doing the meeting we asked what was covered in the signature Pest control service, and under there it says, Oh, by the way, we have $20 off the mosquito service is, is that what this is.
253 00:31:39.270 ⇒ 00:31:42.930 Amber Lin: Yeah, so that yes. So
254 00:31:43.040 ⇒ 00:31:51.649 Amber Lin: this is an example of when they actually suggested a Oh, by the way, right? So some of the answers don’t, because some of the answers don’t need it.
255 00:31:51.650 ⇒ 00:31:52.799 Scott_Harmon: Why don’t they.
256 00:31:53.643 ⇒ 00:32:10.399 Amber Lin: See if they if someone’s just asking a simple question, how do I add this to? How do I do this? In the system. The rep is asking a technical question. It doesn’t really make sense to give them. Oh, by the way, concern, it’s not really directly to the customer. So, depending on the question type.
257 00:32:10.510 ⇒ 00:32:11.190 Amber Lin: we will have.
258 00:32:11.510 ⇒ 00:32:15.079 Scott_Harmon: So utam is, that is, that the the right design.
259 00:32:15.990 ⇒ 00:32:21.829 Uttam Kumaran: Yeah, I think this is what we’re thinking about is, we just started sort of getting this. Oh, by the way, process
260 00:32:22.423 ⇒ 00:32:50.680 Uttam Kumaran: involved, I I’ve been thinking about different ux for this as well, which is basically like, let’s say, a Csr is having a great conversation. They can click a button that then prompts them with maybe the oh, by the way, that they should, they should suggest. I think at this moment we’ve just sort of had this through through this. But I want to think about a more structured way, because not all the time is it appropriate? To send that note. So still thinking.
261 00:32:50.680 ⇒ 00:32:51.030 Scott_Harmon: No, I.
262 00:32:51.030 ⇒ 00:32:51.670 Uttam Kumaran: Yes.
263 00:32:51.670 ⇒ 00:32:56.869 Scott_Harmon: I, we could take this offline. I’ve got a a an alternative thing for you to consider
264 00:32:57.080 ⇒ 00:33:01.660 Scott_Harmon: which is just a model what open AI is doing. And and again.
265 00:33:02.140 ⇒ 00:33:04.510 Scott_Harmon: what’s really common in call center
266 00:33:04.650 ⇒ 00:33:09.910 Scott_Harmon: systems is what’s called a closeout flow. So after the main question’s been answered,
267 00:33:11.650 ⇒ 00:33:19.380 Scott_Harmon: Csr systems and Acd systems. And Ivr, all these systems have a series of things to do at the end of a call
268 00:33:20.930 ⇒ 00:33:22.190 Scott_Harmon: that are
269 00:33:22.540 ⇒ 00:33:33.109 Scott_Harmon: may be related to the call, and maybe not. Sometimes they’re as simple as a customer. Satisfaction survey, but that’s where Upsell would go. I think it’s better to put them on.
270 00:33:33.220 ⇒ 00:33:34.610 Scott_Harmon: So, for example.
271 00:33:34.810 ⇒ 00:33:38.600 Scott_Harmon: the ui at the end of the call could go. Can I help you with anything else?
272 00:33:39.110 ⇒ 00:33:46.450 Scott_Harmon: Would you like more detail about the pricing. Or oh, by the way, would you like more information about our spring leaf cleaning?
273 00:33:46.590 ⇒ 00:33:49.819 Scott_Harmon: In other words, it’s there. Then the Csr
274 00:33:49.990 ⇒ 00:33:55.379 Scott_Harmon: can say yes or no, the Csr understands the context of the call. We don’t yet, because
275 00:33:55.950 ⇒ 00:33:59.429 Scott_Harmon: the the agent just doesn’t have that much contextual knowledge yet.
276 00:33:59.890 ⇒ 00:34:06.950 Scott_Harmon: And so the Csr can just use their judgment and go, man, they’re too mad. This is not the, you know. Time to tell them about leaf cleaning.
277 00:34:06.950 ⇒ 00:34:07.460 YvetteRuiz: No.
278 00:34:07.460 ⇒ 00:34:11.769 Scott_Harmon: But I might argue, it’s better to always in your closeout flow.
279 00:34:12.280 ⇒ 00:34:21.080 Scott_Harmon: Say, you know, can I help you? Oh, by the way, on this, or can I provide more detail about that? And I bet you’d get a lot better?
280 00:34:21.510 ⇒ 00:34:25.170 Scott_Harmon: I wouldn’t wait for the Andy to identify when it’s appropriate.
281 00:34:25.170 ⇒ 00:34:54.289 YvetteRuiz: Right, and I and I agree with that Scott 100. That’s kind of what I’m when I was meeting with Steven yet last week I we made the suggestion, which I thought it was great. You know what I mean, whether we have something that’s there for them to push a button, and it has these quick, all all, you know. Just kind of the quick offers that we have. So then they’re not digging through the sheet. They’re not having to remember. And there’s kind of quick phrases. So we can incorporate something like what you’re saying, Scott or Amber. I don’t know. I know, you guys were going to be working on
282 00:34:55.120 ⇒ 00:35:01.190 YvetteRuiz: the potential of us having that quick button that we can press and it populate that.
283 00:35:05.110 ⇒ 00:35:06.579 Amber Lin: It’s on our to do list.
284 00:35:07.840 ⇒ 00:35:08.920 YvetteRuiz: Okay.
285 00:35:09.100 ⇒ 00:35:13.540 Scott_Harmon: Yeah, just I mean, I at the risk of being pedantic, can I share my screen real quick?
286 00:35:13.540 ⇒ 00:35:16.119 Amber Lin: Yeah. Totally. Let me adjust this.
287 00:35:16.520 ⇒ 00:35:17.140 Scott_Harmon: Yeah.
288 00:35:18.000 ⇒ 00:35:19.220 Amber Lin: You should be able to share.
289 00:35:19.220 ⇒ 00:35:23.669 Scott_Harmon: And I, you know, just to, because I would just suggest we might have
290 00:35:24.560 ⇒ 00:35:28.229 Scott_Harmon: follow the chat bots best practice here. So
291 00:35:29.450 ⇒ 00:35:32.189 Scott_Harmon: this is just a question I was asking yesterday about
292 00:35:33.300 ⇒ 00:35:37.066 Scott_Harmon: trade, foreign trade, surprise, surprise,
293 00:35:38.260 ⇒ 00:35:45.700 Scott_Harmon: and you can see at the end it provided me the answer. But at the end it says, Do you want to go deeper on
294 00:35:46.220 ⇒ 00:35:49.879 Scott_Harmon: this part of the answer or that part of the answer? So there’s always.
295 00:35:50.370 ⇒ 00:35:54.670 Scott_Harmon: and offer to provide more depth. You could also just provide.
296 00:35:55.130 ⇒ 00:35:57.960 Scott_Harmon: And oh, by the way, suggestion here. So
297 00:35:59.011 ⇒ 00:36:01.519 Scott_Harmon: you know, you should always
298 00:36:01.670 ⇒ 00:36:04.800 Scott_Harmon: provide the answer. But then offer to provide more help.
299 00:36:05.640 ⇒ 00:36:08.119 Scott_Harmon: That’s kind of the default. Ui here.
300 00:36:08.811 ⇒ 00:36:15.280 Scott_Harmon: And I found those things to be very helpful a lot of times I didn’t even know. Well, yeah, I guess I do want more information about that.
301 00:36:15.280 ⇒ 00:36:16.120 Amber Lin: Me too.
302 00:36:16.408 ⇒ 00:36:21.880 Scott_Harmon: So that’s that’s something I think you might want to consider and see how it works with the Csrs.
303 00:36:22.240 ⇒ 00:36:23.429 YvetteRuiz: Yeah, makes sense.
304 00:36:24.050 ⇒ 00:36:29.450 MattBurns: Hey? That, would it? And again, this would be for way further down the road if we ever
305 00:36:29.660 ⇒ 00:36:32.720 MattBurns: integrated with evolve here. But it
306 00:36:33.230 ⇒ 00:36:38.889 MattBurns: it’d be interesting to see if the the AI could
307 00:36:39.050 ⇒ 00:36:43.810 MattBurns: kind of identify what they already have with us. What they’ve done in the past, whatever.
308 00:36:44.040 ⇒ 00:36:49.609 MattBurns: and make spots suggest, say, Hey! This is a question as to what might be the best.
309 00:36:50.350 ⇒ 00:36:50.750 Scott_Harmon: Absolutely.
310 00:36:51.060 ⇒ 00:36:51.370 MattBurns: Yeah.
311 00:36:51.370 ⇒ 00:37:02.119 Scott_Harmon: That that is, there is so much I mean, so much potential. That’s called just contextual. And you know information that if you provide it more context with
312 00:37:02.640 ⇒ 00:37:07.619 Scott_Harmon: all the customer information. What they’re subscribed to, you know where they live?
313 00:37:08.149 ⇒ 00:37:15.890 Scott_Harmon: What their history, their call history is. Have they called you 5 times or 10 times like AI can make so much more nuanced suggestions.
314 00:37:16.120 ⇒ 00:37:23.750 Scott_Harmon: Oh, I know you had pest service 2 years ago, you know, like you just get wildly smarter about.
315 00:37:23.750 ⇒ 00:37:26.400 MattBurns: Yeah, particularly with things like.
316 00:37:26.580 ⇒ 00:37:32.169 MattBurns: if they’re already an existing window customer. Don’t bring up window, bring up something else or.
317 00:37:32.170 ⇒ 00:37:32.960 YvetteRuiz: Exactly.
318 00:37:33.170 ⇒ 00:37:42.339 MattBurns: Alternatively. If they had window 3 years ago, and haven’t haven’t had it since. Like you, said Scott. Hey? I noticed you had it in the past. Would you be interested in doing it again? You know.
319 00:37:42.570 ⇒ 00:37:56.660 YvetteRuiz: Or the most popular one is the rodent, right like tree. And I really like I really liked with Steve presented yesterday, you know, with the gutter cleaning the ants, the pest control. You know what I mean like. There’s so many different ways that we can go about it.
320 00:37:56.660 ⇒ 00:38:06.240 Scott_Harmon: For sure the right way to do it. That’s you could get so targeted. And the nice thing is, AI will just figure out what it needs to know to suggest.
321 00:38:06.240 ⇒ 00:38:06.780 Scott_Harmon: Right?
322 00:38:07.257 ⇒ 00:38:18.719 Scott_Harmon: You know you. But you’re right. It will take us to connect a few more sources, you know, in, because right now, Andy’s really just looking at the the knowledge base it built
323 00:38:18.720 ⇒ 00:38:19.390 Scott_Harmon: right? Right?
324 00:38:19.636 ⇒ 00:38:20.130 Scott_Harmon: So it’s.
325 00:38:20.130 ⇒ 00:38:39.779 YvetteRuiz: And I think that’s the key piece that our Csr struggle with is like, when do I offer it? You know what I mean? And I I know a lot of times because I mean you really the the end of the time. The wrap up is is good, but even like when you’re hey? Well, like while I’m getting you scheduled, I mean, just when do I go in there and put that out there? But the that’s the true challenges.
326 00:38:39.780 ⇒ 00:38:49.129 Scott_Harmon: Now until we have that you’re going to have to use training. And Andy, I guess, is probably gonna suggest that in cases I would argue
327 00:38:49.390 ⇒ 00:38:54.880 Scott_Harmon: where it may or may not be appropriate, and you’re going to have to have the Csr. Train them on the judgment of.
328 00:38:55.590 ⇒ 00:38:57.460 Scott_Harmon: But but Andy should always kind of
329 00:38:58.010 ⇒ 00:39:03.580 Scott_Harmon: prompt them right? And then the Ca, you should just tell the Csrs. Look here.
330 00:39:03.870 ⇒ 00:39:07.269 Scott_Harmon: Andy’s always gonna offer it. It’s not you don’t always have to.
331 00:39:08.550 ⇒ 00:39:12.229 Scott_Harmon: You know, share it with the customer. Here’s when when you should, and here’s when you shouldn’t.
332 00:39:12.490 ⇒ 00:39:13.930 MattBurns: Sure. Okay.
333 00:39:16.180 ⇒ 00:39:19.190 YvetteRuiz: Okay, well, thank you, Miss Amber.
334 00:39:19.190 ⇒ 00:39:37.659 Amber Lin: Of course. Just other really really good good news we connected well, and Annie connected the with working with Brian. We connected the call data to the bot data. So that will be really, really helpful. Down the line of oh, this call went shorter because they
335 00:39:37.904 ⇒ 00:39:38.149 Amber Lin: really.
336 00:39:38.720 ⇒ 00:39:39.440 YvetteRuiz: Love it.
337 00:39:39.630 ⇒ 00:39:40.689 Scott_Harmon: Really cool.
338 00:39:40.930 ⇒ 00:39:43.520 Amber Lin: Yeah, I was like, how did you do that? That’s crazy.
339 00:39:44.296 ⇒ 00:39:45.850 YvetteRuiz: Yeah, so.
340 00:39:45.850 ⇒ 00:39:48.989 Scott_Harmon: So this is an example just quickly. So the call.
341 00:39:49.440 ⇒ 00:39:52.540 YvetteRuiz: These are both records from 8 by 8 the.
342 00:39:52.540 ⇒ 00:39:58.600 Amber Lin: The 1st one is from the 1st line. So the customer this is Janice’s call.
343 00:39:58.600 ⇒ 00:39:59.120 Scott_Harmon: Yeah, yeah.
344 00:39:59.120 ⇒ 00:40:11.320 Amber Lin: Made it on March 19.th And then the question she asked was the 1st line down here? So she asked, can we have bedbug. Do we have bedbug services and the outputs? Yes, we do.
345 00:40:11.320 ⇒ 00:40:13.859 Scott_Harmon: So that that’s the question she asked. The bot.
346 00:40:13.860 ⇒ 00:40:14.700 Amber Lin: Yes.
347 00:40:15.170 ⇒ 00:40:16.580 Scott_Harmon: That’s really cool.
348 00:40:16.580 ⇒ 00:40:17.310 Amber Lin: Yeah.
349 00:40:18.540 ⇒ 00:40:25.769 Scott_Harmon: Oh, that is really cool. Okay, alright. Yeah. You down the line that there’s a bunch of really wonderful things we can.
350 00:40:25.770 ⇒ 00:40:38.870 Uttam Kumaran: Yeah. And and really, this is the, this is the main way. We’re gonna be able to isolate like the impacts of the bot is to actually link the call logs to the chat logs, and then eventually the call logs to the customer outcome.
351 00:40:40.050 ⇒ 00:40:42.370 YvetteRuiz: Yeah, that’s amazing, impressive.
352 00:40:42.370 ⇒ 00:40:43.769 Scott_Harmon: That’s very nice. Yeah.
353 00:40:43.770 ⇒ 00:40:44.690 YvetteRuiz: Okay.
354 00:40:47.550 ⇒ 00:40:49.059 Amber Lin: Oh, Hi!
355 00:40:49.460 ⇒ 00:40:51.399 Amber Lin: It’s a little bit unstable.
356 00:40:52.089 ⇒ 00:41:01.969 Amber Lin: Here, I’ll hand it over to Utam. So essentially, we want to talk about the phase to proposal. We came up with a pricing model, and we wanted to share with you guys.
357 00:41:02.500 ⇒ 00:41:03.200 MattBurns: Good.
358 00:41:03.600 ⇒ 00:41:15.099 Uttam Kumaran: Yeah, maybe we can just sort of ha! This is probably the 1st week where we we actually saw some adoption. And we were sort of thinking about what’s the best model to actually price price. This, you know, we.
359 00:41:15.200 ⇒ 00:41:36.330 Uttam Kumaran: you know, worked working with Scott and talking a little bit about how typically software is price for sort of these sort of call center activities and sort of looking at the adoption. You know, one of the things that we noticed as we spoke with Brian, and we spoke with some folks on the team about you know. What are the total call volumes understanding like, hey, we want to limit having people like
360 00:41:36.420 ⇒ 00:41:40.330 Uttam Kumaran: come in and and take extra calls, things like that. And so one of the things that
361 00:41:40.380 ⇒ 00:41:51.511 Uttam Kumaran: we we thought about was like, Okay, what’s the best way to price this and sort of make sure that you guys can use the tool as needed and sort of drove over time. And then actually, the the price
362 00:41:51.800 ⇒ 00:42:10.410 Uttam Kumaran: the price goes down on average for every incremental usage. And so really, where we arrived on. And this is something that I’ll I’ll I’ll share in an email. Is at A at a model where you would prepay for basically chat sessions, and so there would be a model where you can prepay for a set number of sessions.
363 00:42:10.410 ⇒ 00:42:21.840 Uttam Kumaran: Your agents can use that up, and then, as you go over that. There’s just another tier of pricing. So really, what that what that helps do is sort of as the it gets used more and more and more.
364 00:42:21.840 ⇒ 00:42:33.849 Uttam Kumaran: the price per session actually goes down. And we would sort of tier that, you know, on a basically on a volume basis. I know it’s a little bit hard to to see that visually, but that’s sort of what we were thinking.
365 00:42:33.850 ⇒ 00:42:34.230 Scott_Harmon: Yeah.
366 00:42:34.480 ⇒ 00:42:36.989 Uttam Kumaran: Yeah, Scott, if you want to add any color there.
367 00:42:36.990 ⇒ 00:42:43.449 Uttam Kumaran: yeah, just to be more tangible, and we’ll send you something over. Post this call Matt and and that, but.
368 00:42:43.880 ⇒ 00:42:50.190 Scott_Harmon: You know, we talked about this a couple weeks ago. What? And this is, you know, we’ll get this on the table. Say, if something to react to.
369 00:42:50.470 ⇒ 00:43:08.570 Scott_Harmon: It’s kind of my recommendation to start simple and then get more complicated down the line. And what we’re going to propose is 3 bands or tiers, which would be 3 different monthly prices. The 1st one is $9,000 a month again. These are all negotiable, I’m just get, you know getting. And then there’s a couple of other steps above that.
370 00:43:08.690 ⇒ 00:43:11.329 Scott_Harmon: and the 1st band would be up to
371 00:43:11.940 ⇒ 00:43:16.450 Scott_Harmon: some number of Andy sessions in that month, and
372 00:43:16.900 ⇒ 00:43:26.529 Scott_Harmon: I think that number’s, you know, going to be around 2,500. But it’s a lot relative to where we’re at. It’s a lot right? And so the idea is, you just pay that amount.
373 00:43:27.040 ⇒ 00:43:28.680 Scott_Harmon: Yeah, for any.
374 00:43:28.810 ⇒ 00:43:32.410 Scott_Harmon: You know, it’s just one set fee for up to that many sessions.
375 00:43:32.750 ⇒ 00:43:39.330 Scott_Harmon: And then once you go over, it’s just a little bit like cell phone plans. Then once you go over that number, there’s a second band.
376 00:43:39.900 ⇒ 00:43:48.420 Scott_Harmon: and then we we actually put a 3rd band. If it’s just really wildly, heavily used, we set the numbers pretty high
377 00:43:48.750 ⇒ 00:43:51.820 Scott_Harmon: in terms of the number of sessions, so that.
378 00:43:52.430 ⇒ 00:43:55.569 Scott_Harmon: you know we’d be seeing an awful lot of adoption
379 00:43:55.820 ⇒ 00:44:04.589 Scott_Harmon: right? And the one thing it gives it gives brain forges that each, you know, you’re committed to at least the monthly.
380 00:44:05.080 ⇒ 00:44:22.100 Scott_Harmon: the bottom monthly tier. And just to be candid, that covers their call right, they’re they’re not gonna lose money, but they’re probably not gonna make much money either, right? So we tried to structure it where it’s sort of a shared risk reward thing. So the heart of it is these 3 bands
381 00:44:22.360 ⇒ 00:44:24.440 Scott_Harmon: that are based, and and I
382 00:44:24.610 ⇒ 00:44:35.439 Scott_Harmon: you’ll see the exact numbers, but I think we try to set them pretty high, so our hope would be that you’d look at them and go. Wow! Andy’s really being used a lot, and so you’d feel good about that.
383 00:44:35.650 ⇒ 00:44:45.169 Scott_Harmon: It is as as you mentioned, the math, for the tiers is, if you do the math, you’ll see that the per session price falls
384 00:44:45.620 ⇒ 00:44:53.789 Scott_Harmon: for each band. Right? So it’s just kind of like a volume discount. So that’s the concept. I’ve used it a lot and other kind of things. It’s just designed to be real simple.
385 00:44:54.020 ⇒ 00:44:56.900 Scott_Harmon: I’d like pricing models where
386 00:44:57.240 ⇒ 00:45:01.119 Scott_Harmon: you know, you can sign up to something. You know what it’s going to be. You feel good about it.
387 00:45:01.630 ⇒ 00:45:05.650 Scott_Harmon: and you’re incentivized to get people to use it more.
388 00:45:06.610 ⇒ 00:45:19.239 Scott_Harmon: and you don’t necessarily. It’s not like a sas license. Where? Oh, my gosh, I gotta pay for every incremental seat. I’m not sure people are using it. So it’s it’s a hundred percent usage base. So that’s the philosophy. We can add fancy stuff to it later.
389 00:45:19.470 ⇒ 00:45:25.019 Scott_Harmon: I frankly think if we can get the oh, by the ways thing figured out, we could add like a kicker
390 00:45:25.490 ⇒ 00:45:29.949 Scott_Harmon: like a little bonus like oh, you know, if you sold out oh, by the ways.
391 00:45:30.420 ⇒ 00:45:33.720 Scott_Harmon: And Andy, you know, was helpful on that. But we can add that later.
392 00:45:33.820 ⇒ 00:45:36.530 Scott_Harmon: So that’s where we’d like to start.
393 00:45:36.710 ⇒ 00:45:38.139 Scott_Harmon: But then that’s the concept.
394 00:45:38.140 ⇒ 00:45:40.180 Scott_Harmon: I didn’t. You know we’ll give you the numbers to react to.
395 00:45:40.180 ⇒ 00:45:45.169 MattBurns: Okay? Well, again, look, concept makes sense, because in one respect
396 00:45:45.700 ⇒ 00:45:50.860 MattBurns: you don’t want a pricing model that discourages usage. You want one that encourages usage.
397 00:45:50.860 ⇒ 00:45:51.830 Scott_Harmon: Exactly.
398 00:45:52.740 ⇒ 00:45:56.790 MattBurns: This would do, and I think.
399 00:45:57.190 ⇒ 00:46:02.290 MattBurns: Yvette, you would agree to get the full benefit of it.
400 00:46:03.390 ⇒ 00:46:10.540 YvetteRuiz: We’ve got to have high usage of it, because that’s where you can make an impact in your handling time.
401 00:46:10.870 ⇒ 00:46:17.290 MattBurns: And hopefully in your oh, by the way, because we know both of those
402 00:46:17.850 ⇒ 00:46:21.790 MattBurns: impact, the bottom line, the handling time in in particular.
403 00:46:21.920 ⇒ 00:46:22.360 YvetteRuiz: Yep.
404 00:46:22.360 ⇒ 00:46:27.510 MattBurns: Just from an employee. Staffing standpoint, hey? I don’t need as many people.
405 00:46:27.790 ⇒ 00:46:31.560 MattBurns: because now each agent can handle more calls
406 00:46:31.840 ⇒ 00:46:36.970 MattBurns: and handle them better and more efficiently. And also
407 00:46:37.350 ⇒ 00:46:43.760 MattBurns: I’m selling more services via, oh, by the way, appointments. And that’s really where
408 00:46:44.540 ⇒ 00:46:48.850 MattBurns: that’s a hard dollar where we could say, like again.
409 00:46:49.000 ⇒ 00:46:53.000 MattBurns: guys, we emphasize that a lot in our monthly meetings that, hey?
410 00:46:53.300 ⇒ 00:46:58.809 MattBurns: Last month we set what it what it was, Eva. 330 appointments.
411 00:46:59.180 ⇒ 00:47:00.409 YvetteRuiz: Yep, and oh! By the way.
412 00:47:00.750 ⇒ 00:47:06.990 MattBurns: Way. But that Scott is with really
413 00:47:07.100 ⇒ 00:47:14.579 MattBurns: participation. Level, I’d say, is modest or moderate. Wouldn’t you agree that I mean you still have some people that
414 00:47:14.990 ⇒ 00:47:19.259 MattBurns: there’s a few people that knock it out of the park, and there’s a few people that don’t do anything.
415 00:47:20.000 ⇒ 00:47:26.579 MattBurns: and it’s a little bit with like our field sales. It’s like if I can get the Zeros to
416 00:47:26.790 ⇒ 00:47:29.479 MattBurns: give me 3, 4, 5,
417 00:47:29.900 ⇒ 00:47:36.610 MattBurns: Really jumps a lot. And so that’s a tangible result. So yeah.
418 00:47:37.260 ⇒ 00:47:40.359 MattBurns: Scott, let’s evaluate that. And just kind of.
419 00:47:40.830 ⇒ 00:47:41.550 Scott_Harmon: Yeah, I think.
420 00:47:41.550 ⇒ 00:47:42.370 MattBurns: I think the time.
421 00:47:42.370 ⇒ 00:47:46.539 Scott_Harmon: Exactly insane are are. That’s, I do think those are the right 2
422 00:47:47.340 ⇒ 00:47:50.140 Scott_Harmon: bottom line things we want to stay focused on
423 00:47:50.581 ⇒ 00:48:04.729 Scott_Harmon: again. We can you give us feedback about whether the 9 grand per month for the base tier works or not after you stare at it a bit. Our thinking also was. I don’t know what your burden cost for a level 2 specialist would be, but
424 00:48:05.070 ⇒ 00:48:18.270 Scott_Harmon: the way I think about it is gosh! They can shorten. Call times. Let my Csrs do more more calls a day. Would I pay a, you know, a hundred $1,000 a year for an employee to do that, you know I probably would. Right.
425 00:48:18.270 ⇒ 00:48:23.090 Uttam Kumaran: Yeah. And additionally, Scott, it’s also like we’re trying to limit the overflow, you know. And that’s something that
426 00:48:23.250 ⇒ 00:48:40.389 Uttam Kumaran: I know you guys are coming up into season right now. And so we were thinking a lot about that, as like, Oh, yeah, if you have to bring on. And then, you know we learned from Brian that those folks are stretched across multiple divisions right? So their HD goes up. And you know, that’s that’s sort of how we were thinking about it.
427 00:48:41.110 ⇒ 00:48:43.069 Scott_Harmon: So think you’re right. You open up.
428 00:48:43.070 ⇒ 00:48:50.279 Uttam Kumaran: I think once we can agree on okay, when to give credit to Andy for them, then I think we can add that it’s because you’re right. That’s almost like.
429 00:48:51.330 ⇒ 00:48:54.059 Scott_Harmon: You should be happy to pay. You know that cause you’re like.
430 00:48:54.060 ⇒ 00:48:59.300 MattBurns: Yeah, I it’s interesting. I had a conversation with our lawn manager this morning, and
431 00:48:59.750 ⇒ 00:49:03.170 MattBurns: in in Austin, Texas in particular, when
432 00:49:03.840 ⇒ 00:49:07.279 MattBurns: it’s generally the 3rd week of March through
433 00:49:07.810 ⇒ 00:49:11.640 MattBurns: maybe the middle of May our lawn volume increases
434 00:49:12.770 ⇒ 00:49:26.640 MattBurns: 3 times what it is almost the rest of the year, and there’s no way you can staff for it. So we have the overflow groups that can handle on. But those overflow groups are not nearly as proficient.
435 00:49:26.640 ⇒ 00:49:28.180 Scott_Harmon: See? That’s that’s the problem.
436 00:49:28.180 ⇒ 00:49:39.729 MattBurns: Lawn people. So does Andy make those overflow people extremely efficient, or a lot more efficient than they otherwise would be. Because I bet you, Yvette.
437 00:49:40.640 ⇒ 00:49:55.180 MattBurns: the handling time, and maybe Brian could. I’d be interesting if Brian could run some data on this to say when those calls go to an overflow agent. What’s the average handling time versus quote unquote the expert in the law department handling it so anyway.
438 00:49:55.180 ⇒ 00:50:03.290 YvetteRuiz: We. We have that I mean it. It goes up with them, and it goes up with new hires. We and those are the 2 things that spike it up. Are are there
439 00:50:03.290 ⇒ 00:50:04.180 YvetteRuiz: virtually times.
440 00:50:04.180 ⇒ 00:50:07.870 MattBurns: And and the hope is obviously this is the solution for that. Yeah.
441 00:50:07.870 ⇒ 00:50:12.320 Scott_Harmon: Good. So so we’ll get it over like I said, we’re gonna start with 3 bands. I think.
442 00:50:12.550 ⇒ 00:50:13.060 MattBurns: Yeah.
443 00:50:13.060 ⇒ 00:50:32.209 Scott_Harmon: You know, we’re not gonna be able to get into. I think you want us also to be incentivized to to make it so successful, we can get to the second and 3rd band. So now you know, brain forge and amber, and they’re like, Hey, how do we get more people? You know, we are. Gonna wanna talk about rolling it to other lines because we’re only one now. So
444 00:50:32.580 ⇒ 00:50:36.440 Scott_Harmon: that way, you can just see. Okay, all the incentives are aligned.
445 00:50:36.920 ⇒ 00:50:46.040 Uttam Kumaran: Exactly. And and you know, as we roll this out beyond, we were thinking about, do we price this on a per division, and thinking through all that. I think this sort of solves for a lot of that where
446 00:50:46.160 ⇒ 00:51:13.560 Uttam Kumaran: we just want adoption across the entire org. As we start to take on more of that work like for us. We’re, you know, we’re bearing the expense of, you know the system. And so we’re always looking for ways to make the system cheaper and faster. But that way again, we we sort of focus on just the adoption. Right? And I don’t know. It seems like, after examining a couple of models like it seems pretty aligned. And so we’ll we’ll get that over to you today, and then we can poke at it a bit.
447 00:51:13.560 ⇒ 00:51:20.740 Scott_Harmon: The other thing just to give. You know, utam kudos, and to be maybe overly blunt.
448 00:51:20.950 ⇒ 00:51:28.090 Scott_Harmon: What I like about it is he’s running a services business, and he’s willing to have the base tier be a break even deal like.
449 00:51:28.600 ⇒ 00:51:34.899 Scott_Harmon: you know, if he never gets out of the base here, he’s not gonna make a ton of ton of profit on this thing.
450 00:51:35.380 ⇒ 00:51:42.989 Scott_Harmon: as you know, Matt won’t do that right. They have a baked in margin, so he’s he’s really. He’s really risking, you know.
451 00:51:44.060 ⇒ 00:51:45.199 MattBurns: Well, that’s why we
452 00:51:45.350 ⇒ 00:51:51.069 MattBurns: even preliminary discussions. We said, Yeah, some sort of base plus model makes sense. We know you have to
453 00:51:51.200 ⇒ 00:51:56.680 MattBurns: get a certain amount every month, to quote unquote cover costs or close to it. And then.
454 00:51:57.190 ⇒ 00:52:01.900 MattBurns: if it’s successful, we both win after that. So I think it makes sense.
455 00:52:01.900 ⇒ 00:52:07.489 Scott_Harmon: Okay. Well, we’ll send it over like I said, you’ll you’ll mark it up, and you know, feel free to suggest.
456 00:52:07.490 ⇒ 00:52:08.190 MattBurns: We will.
457 00:52:08.190 ⇒ 00:52:19.509 Scott_Harmon: Tweaks to it, and and then why don’t we just send over the numbers and to get the numbers square, and then we can send over a full agreement with T’s and C’s that Matt can
458 00:52:19.780 ⇒ 00:52:23.189 Scott_Harmon: true on. Better to get the 3.
459 00:52:23.190 ⇒ 00:52:26.550 MattBurns: Next week’s fine. We’re not going to do much on it today.
460 00:52:27.160 ⇒ 00:52:32.240 YvetteRuiz: I’ve got to jump off to go to a collections meeting so we can get some money in here.
461 00:52:32.240 ⇒ 00:52:33.309 Scott_Harmon: That later, too, I would have told.
462 00:52:33.310 ⇒ 00:52:35.821 YvetteRuiz: I was supposed to be in there. Too bad!
463 00:52:36.100 ⇒ 00:52:38.490 MattBurns: Yeah, we’ll jump in there. But
464 00:52:38.810 ⇒ 00:52:43.019 MattBurns: good good meeting, guys. Thanks. I appreciate what you do. And it looks like we’re
465 00:52:43.610 ⇒ 00:52:46.290 MattBurns: everybody’s positive on it. So good stuff.
466 00:52:46.490 ⇒ 00:52:49.699 YvetteRuiz: Absolutely. Thank you so much, Miss Amber.
467 00:52:49.700 ⇒ 00:52:50.349 MattBurns: Thanks guys.
468 00:52:50.350 ⇒ 00:52:52.499 Uttam Kumaran: Thanks, amber thanks everyone. Bye.