Meeting Title: Yvette - Pranav: Intro + Sync on Workstreams! Date: 2026-03-09 Meeting participants: Pranav Narahari, YvetteRuiz, read.ai meeting notes, Yvette’s Notetaker (Otter.ai)
WEBVTT
1 00:01:52.160 ⇒ 00:01:53.210 Pranav Narahari: Hey, Yvette.
2 00:01:54.710 ⇒ 00:01:56.069 YvetteRuiz: Hi, how are you? How are you?
3 00:01:56.070 ⇒ 00:01:58.419 Pranav Narahari: I’m good, I’m good, nice to see you again.
4 00:01:58.730 ⇒ 00:01:59.750 YvetteRuiz: Yes!
5 00:02:00.730 ⇒ 00:02:02.100 Pranav Narahari: So,
6 00:02:02.500 ⇒ 00:02:08.009 Pranav Narahari: Yeah, just kind of… I had this meeting, like we said, like, this morning with, like, Janiece, just to kind of, like…
7 00:02:08.460 ⇒ 00:02:15.140 Pranav Narahari: you know, not try to just, like, get all the context from, like, Amber, Sam, and, like, Utam. I think it’ll also be,
8 00:02:15.650 ⇒ 00:02:22.870 Pranav Narahari: beneficial for me to kind of just, like, talk to you directly. What I noticed from, like, my conversation with Janiece was just, like.
9 00:02:23.170 ⇒ 00:02:25.800 Pranav Narahari: I think it’s good for me to just kind of, like.
10 00:02:26.220 ⇒ 00:02:30.549 Pranav Narahari: I’ve already gotten a ton of context, but just, like, start a little bit more from just, like…
11 00:02:30.890 ⇒ 00:02:50.239 Pranav Narahari: like, ground zero, and ask you, like, maybe specific questions that just, like, haven’t been, like, talked about in a while, and maybe that’ll just, like, spark certain things to get, like, initiated again that maybe haven’t been, like, talked about in some time. And so… yeah, there’s, like, a few things that I want to talk about, but…
12 00:02:50.430 ⇒ 00:03:00.840 Pranav Narahari: you know, maybe I can just kind of give you just, like, a quick, like, intro, like, of myself, of just, like, George. Not, like, a ton of time, because we’ve already kind of done this a little bit.
13 00:03:00.880 ⇒ 00:03:12.370 Pranav Narahari: But yeah, at Brainforge, like, what I am mostly doing in my day-to-day is, like, working with clients specifically, and then, helping orchestrate, like, okay, what does that look like
14 00:03:12.370 ⇒ 00:03:32.059 Pranav Narahari: internally to us… for us to actually deliver on these, like, specific things that we’ve, we’ve talked about, like, our problems for the clients. And so, for right now, like, with, with you, like, I want to be able to be, like, the main person that you can kind of come to, to talk about, like, okay, this is going well, this isn’t going well,
15 00:03:32.220 ⇒ 00:03:35.999 Pranav Narahari: And this is what we want to do, like, going forward.
16 00:03:36.010 ⇒ 00:03:42.509 Pranav Narahari: And it’s… I feel like when it’s, like, most successful is when, like, you kind of just, like, tell me your problems, and then I can kind of, like.
17 00:03:42.510 ⇒ 00:04:02.229 Pranav Narahari: talk to you about, like, the solutions that we have, that we’ve built in the past. Usually, we’ve built something similar, and so, from there, we kind of just, like, refine what that exact product’s gonna look like, and then I’m, like, I tell all those smart engineers, like, on my team to just be like, hey guys, XYZ, let’s, like, let’s ship these things.
18 00:04:02.450 ⇒ 00:04:14.710 Pranav Narahari: I’m guessing that’s kind of how it’s already been with, like, Amber, but just, like, wanted to let you know, like, this is how I felt, like, operating. It’s just, like, it’s, like, the most efficient, and, like, okay, cool.
19 00:04:15.010 ⇒ 00:04:17.400 Pranav Narahari: Yeah, and I’d love to know, like…
20 00:04:17.670 ⇒ 00:04:35.740 Pranav Narahari: kind of with you, too, like, my understanding right now between, like, you and Janiece is, like, Janiece is kind of, like, she’s more in the weeds with, like, the day-to-day, like, tasks that are going on, but you kind of, like, help orchestrate, like, okay, where do we need to be, like, in a week, in a month?
21 00:04:36.200 ⇒ 00:04:41.959 Pranav Narahari: Is that kind of accurate, or would you say it’s a little bit more, there’s some detail that I’m missing there?
22 00:04:41.960 ⇒ 00:04:56.820 YvetteRuiz: Yeah, no, it… that’s the way it should be, you know, because I… I just don’t have time to be in the weeds, you know what I mean? But I… I want to… I need to be in the know on, okay, if this is the goal, right, if this is everything we’re working to.
23 00:04:56.820 ⇒ 00:05:04.149 YvetteRuiz: How close are we to get there? What are the pain points that are keeping us from getting there? And where do I need to step in to really help
24 00:05:04.150 ⇒ 00:05:10.419 YvetteRuiz: navigate, find out where the issue’s been, and so… you know, I, I…
25 00:05:10.420 ⇒ 00:05:29.409 YvetteRuiz: I’ve had, you know, in some past conversations, it’s kind of like really trying to identify, okay, what is the responsibility of BrainForge, and what’s the responsibility of ABC, right? So then that way, you know, if it’s on our end, okay, what is it specifically that we need to get to? So, to answer your question, I don’t want to go too,
26 00:05:29.650 ⇒ 00:05:45.099 YvetteRuiz: you know, I don’t want to digress here. But yes, Janiece is the customer service trainer, so her responsibility is supposed to be driving this down and working with all the leaders here to push down, okay.
27 00:05:45.290 ⇒ 00:05:58.920 YvetteRuiz: making sure that it is happening, right? And then, of course, filtering it back up to me. But there’s still broken communication at some end, and I’m hoping, this is where I’m getting excited, that maybe you’re going to be the point person to really, okay.
28 00:05:59.130 ⇒ 00:06:06.539 YvetteRuiz: kind of guide me through that and make sure that we are kind of hitting those levels, those points, where are we at? Does that make sense?
29 00:06:06.880 ⇒ 00:06:18.229 Pranav Narahari: Yeah, so kind of, like, what I can help with is, like, when Janice is kind of, like, in the weeds with, like, the trainers and, like, using Andy, like, kind of helping with…
30 00:06:18.630 ⇒ 00:06:28.030 Pranav Narahari: getting you context on, like, okay, what are, like, the main chunks of, like, information that I should know, right? It doesn’t make sense for you to know every single tiny detail,
31 00:06:28.290 ⇒ 00:06:43.489 Pranav Narahari: Yeah, that’s… I think that’s kind of, like, how I see, like, this, like, triangle, like, type of communication working. And so, like, if we kind of maybe just talk about the specifics, too, and how I see the specifics to be, and maybe we can talk about how it may differ for you, is like…
32 00:06:43.570 ⇒ 00:06:59.099 Pranav Narahari: there are 3 main things that we’re working on right now, which is the migration, which we talk about it during every call, every conversation, basically, the central dock restructuring, and then also the zip codes updates. Yep.
33 00:06:59.100 ⇒ 00:07:04.610 Pranav Narahari: So those are the main three things. Let’s just start with, like, the central dock.
34 00:07:04.640 ⇒ 00:07:05.550 Pranav Narahari: Okay.
35 00:07:05.860 ⇒ 00:07:08.039 Pranav Narahari: What is, like, the type of, like…
36 00:07:08.170 ⇒ 00:07:23.239 Pranav Narahari: granularity that you would like updates on, in terms of… because, like, we give a lot to Janiece, we also talk even more about the technical stuff internally. What do you think is, like, important for you to understand in terms of, like, where we’re at in terms of status, anything in that.
37 00:07:23.240 ⇒ 00:07:29.010 YvetteRuiz: Perfect. Yeah, so as you heard earlier, I was a little bit confused, because I meet with my leaders
38 00:07:29.010 ⇒ 00:07:52.489 YvetteRuiz: When I do my one-on-ones, right? And I measure them based off of the success of things, you know what I mean? Our KPIs, you know what I mean? The phones, the developments and all. Andy is one of them. And so I am always going in there saying, okay, why are your percentages down this week? So-and-so’s not using this, right? Why is that, right? And they’ll come, they’re like, it’s not working, we’ve already put X amount of…
39 00:07:52.490 ⇒ 00:08:08.649 YvetteRuiz: And so, I go in there blindly versus prepared, you know? And so, one of them was last week with Patricia, you know, saying, hey… and I’m not in every meeting, so I don’t know the answers to some of those, but I’m gonna get to your question, what you’re asking. So, she’s like, you know.
40 00:08:08.920 ⇒ 00:08:24.079 YvetteRuiz: the… and I know that mitigation has something to do with it, because of the speed and the accuracy of what the… how Andy’s responding to them. So, one of the big things is it’s not giving us the right answers. That’s what we were saying. It’s not, you know, you can ask it, and it gives us
41 00:08:24.110 ⇒ 00:08:35.399 YvetteRuiz: different… different ways. It’s giving us the answers, right? So, they understood that Amber was going to be meeting with them to really get them the data, so they can start really structuring how that should be. So…
42 00:08:35.840 ⇒ 00:08:47.340 YvetteRuiz: somewhere that got lost in translation, or something happened, right? So, those are the things that I need to know. No, we’re here, you know what I mean? This is what you need to be working on.
43 00:08:47.430 ⇒ 00:09:01.759 YvetteRuiz: Does that make… does that give you a little bit of… it’s kind of the overview of what I really need to make sure that I’m pushing and I’m holding them accountable for, that there is no delay on our end, but I’m also getting that information on you guys, if there is a delay on y’all’s end.
44 00:09:01.760 ⇒ 00:09:13.359 YvetteRuiz: what’s the timeline? What’s the expected turnaround to get that going, so then that way I could build the confidence in it, because that’s what I’m trying to do right now. Like, this is a great tool. ABC’s making a big investment of us, but…
45 00:09:13.380 ⇒ 00:09:21.509 YvetteRuiz: If it’s not working, where is… where is it not working? Because at the end of the day, I have to answer to Matt and Bobby and let them know, hey, here’s where we’re at.
46 00:09:21.510 ⇒ 00:09:26.230 Pranav Narahari: Yeah, and totally, and also what I think is gonna be helpful from you, too, is to, like.
47 00:09:26.660 ⇒ 00:09:37.059 Pranav Narahari: be critical, and be, like, critical often, too. Just, like, let us know, like, which departments, you know… and Janiece does a good job of, like, letting us know, and, like, she is, like.
48 00:09:37.060 ⇒ 00:09:37.600 YvetteRuiz: I don’t know.
49 00:09:37.600 ⇒ 00:09:52.499 Pranav Narahari: she really understands how beneficial this could be to the trainers, but I think what information would be, and you kind of already gave it to me right now, to, like, a higher level, which departments are having issues with it, which ones are not confident in using it?
50 00:09:53.370 ⇒ 00:09:55.890 Pranav Narahari: Like, giving them the wrong answers.
51 00:09:55.890 ⇒ 00:09:56.510 YvetteRuiz: Yep.
52 00:09:56.510 ⇒ 00:10:01.040 Pranav Narahari: And that should be our North Star, right? Like, right answers. So…
53 00:10:01.670 ⇒ 00:10:21.310 Pranav Narahari: if we’re not… like, I just kind of… it’s a good, like, pulse check for me to you, kind of, like, you’ll be talking to them much more frequently than I’m talking to them, and so, like, if we can have that constant communication of just, like, oh, like, they’re definitely… and this can be quantitative, too, you know, they’re increasing their usage. We already have visibility of that, so you don’t need to really give me that information.
54 00:10:21.410 ⇒ 00:10:27.269 Pranav Narahari: But maybe the details that are lost in that… in those numbers will be super useful.
55 00:10:27.460 ⇒ 00:10:29.789 Pranav Narahari: Okay, so that’s really good to know, and I.
56 00:10:29.790 ⇒ 00:10:30.890 YvetteRuiz: Yeah, and with…
57 00:10:30.890 ⇒ 00:10:31.260 Pranav Narahari: Yep.
58 00:10:31.260 ⇒ 00:10:38.500 YvetteRuiz: Oh, I’m sorry, no, no, and the other thing with Janiece, and this is what I have tasked her with, because, you know, she… she wasn’t…
59 00:10:38.870 ⇒ 00:10:50.539 YvetteRuiz: she wasn’t being… she wasn’t able to dedicate the time to Andy as we needed to, and I shifted her focus, making sure, like, look, your responsibility as the trainer, this is one big tool, and it is making sure that
60 00:10:50.550 ⇒ 00:11:08.690 YvetteRuiz: everyone is aligning, getting the same message. Everyone is aligning, getting the same message, and we’re doing… because if… if I’m looking at KPIs, and I’m seeing… because she’s… I’m… her KPIs are over everything. If I’m seeing Lawn is not doing that, okay, what specifically, again.
61 00:11:08.690 ⇒ 00:11:22.229 YvetteRuiz: I’m holding the manager accountable, but I’m also holding her accountable, like, what conversations, because she’s meeting with them on a regular… she’s meeting with you guys on a regular. What specific conversations are we having to improve that? So I just want to kind of add that context as well.
62 00:11:22.230 ⇒ 00:11:24.649 Pranav Narahari: Perfect, yeah.
63 00:11:25.130 ⇒ 00:11:32.139 Pranav Narahari: One other thing that I kind of want to talk about before we get into, like, zip codes and, migration as well is,
64 00:11:33.380 ⇒ 00:11:53.100 Pranav Narahari: like, how has, like, the velocity seemed for you in terms of, like, how we’ve been moving forward with Andy and getting better? I ask that because, like, I know we’re working very hard here, like, I can see the engineers working every single day, and we have, like, great engineers, but what can sometimes become misaligned is, like, okay, if we’re picking up, like, new things that are actually, like.
65 00:11:53.400 ⇒ 00:12:12.220 Pranav Narahari: we’re not all aligned on is, like, a priority, then that’s just gonna slow us down, right? And so… how has that, like, looked for you? Have you felt like, okay, the project has kind of, like, veered in a weird direction, where we’re working on things that you didn’t initially think we should be working on? And you know, sometimes that’s fair, but, like, we should just be aligned on, like.
66 00:12:12.440 ⇒ 00:12:23.449 Pranav Narahari: it makes sense what we’re working on, you know? And it’s not just like, oh, they’re saying these things, they’re working on this, why is that? Does that… that question make sense?
67 00:12:23.660 ⇒ 00:12:26.930 YvetteRuiz: Yeah, so yeah, so thank you for that, and I think,
68 00:12:27.060 ⇒ 00:12:34.750 YvetteRuiz: getting that out there, because I do feel like for… I’ve lost my focus as far as, like, okay.
69 00:12:34.970 ⇒ 00:12:42.660 YvetteRuiz: if the original goal was to build Andy to be our AI assistant, to be able to help us, get us the answers.
70 00:12:42.660 ⇒ 00:13:00.349 YvetteRuiz: we’ve done… we moved the needle on that, but then somewhere down the line, we became… something kind of just… we hit a wall somewhere, right? And it was between the speed, between the data that it was spitting out to us, so then we identified, okay, the format’s not right, this is not, you know, right in this…
71 00:13:00.350 ⇒ 00:13:20.299 YvetteRuiz: And then that kind of plateaued for a minute, you know what I mean? And there’s lack of clarity that was coming down from that. I know that we also shifted because we were working on another project with ABC as well, and then, of course, the transcripts, some of that stuff also got lost in translation, because those are all important to give us the KPIs on what is Andrew giving us, but…
72 00:13:20.320 ⇒ 00:13:25.059 YvetteRuiz: before we even get to that level, I feel like we need to make sure that we get Andy right first.
73 00:13:25.110 ⇒ 00:13:25.950 YvetteRuiz: You know?
74 00:13:25.950 ⇒ 00:13:50.929 YvetteRuiz: And that’s why I like the way you broke it down. Okay, we have 3 things right now that we’re working on, right? The migration, right? The anti-migration. What is the anti-migration gonna do for us, right? Yeah. Speed, right? And we know that the impact that speed’s gonna get us, and of course, the accuracy, which is everything, because that’s where the trust is, right? That… do we have… is it giving us what we need to, right? And then the central docs. Okay, well, obviously, we have to clean up the central docs.
75 00:13:50.930 ⇒ 00:14:02.250 YvetteRuiz: have to get that correctly in order for it to function correctly. So there are two huge things, and I feel like that has to… we’ve got to start seeing progress before we start tagging all these other things.
76 00:14:02.790 ⇒ 00:14:03.310 Pranav Narahari: Yes.
77 00:14:03.310 ⇒ 00:14:03.740 YvetteRuiz: So…
78 00:14:03.930 ⇒ 00:14:06.589 Pranav Narahari: So I’m super glad you said that, too, because, like.
79 00:14:07.200 ⇒ 00:14:17.889 Pranav Narahari: sometimes I’ll be working with, people, and, like, they’ll be mentioning, like, 10 different work streams, and it… it’s hard, you know, to be like, okay, there’s 10 work streams.
80 00:14:18.070 ⇒ 00:14:20.479 Pranav Narahari: There has to be one priority, you know?
81 00:14:21.370 ⇒ 00:14:22.730 Pranav Narahari: And so, I’m glad…
82 00:14:22.730 ⇒ 00:14:33.250 YvetteRuiz: I could get that way, too. I could get that way, too, because I want to go now to my KPIs. I want to do that, but I’m like, I can’t even get there yet until I get this portion fixed right here. Totally.
83 00:14:33.250 ⇒ 00:14:49.200 Pranav Narahari: Yeah, and so I’m really glad that you said that, because I feel like we are actually working on all the right things right now, and so… and I think you already mentioned it, with migration, what that does is that it’s going to increase accuracy, but then also that time issue that I talked to Janiece a lot about, and it’s just like.
84 00:14:49.200 ⇒ 00:14:53.450 Pranav Narahari: These conversations aren’t going to be natural if we’re waiting 25 seconds for a response from Andy.
85 00:14:53.450 ⇒ 00:15:02.190 Pranav Narahari: And so that is something that we’re working on a lot. We have a couple things in the pipes to help with that. One was just from…
86 00:15:02.480 ⇒ 00:15:17.529 Pranav Narahari: like, migrating a little bit of the technical toolset, like, we were reaching, like you said, like, a plateau with one thing that we were using before. Now Casey right now is doing a huge migration into, using a different type of infrastructure for the backend, which is helping us out a lot.
87 00:15:17.750 ⇒ 00:15:35.649 Pranav Narahari: Also, what we did with, Google, we were getting, like, access, and Mustafa and Casey, like, mentioned to me last week that we were finally able to, like, get the newest and latest models, for Gemini. That’s going to also help. So, accuracy and,
88 00:15:35.700 ⇒ 00:15:41.600 Pranav Narahari: latency are both going to be, like, tackled by doing these few things that I just mentioned.
89 00:15:41.600 ⇒ 00:15:44.550 YvetteRuiz: So, can I just kind of ask a little bit more, and this is…
90 00:15:45.010 ⇒ 00:15:49.329 YvetteRuiz: How often are we going to be meeting, you and I? On a regular basis, or how…
91 00:15:49.660 ⇒ 00:15:50.939 Pranav Narahari: Like, these type of meetings?
92 00:15:50.940 ⇒ 00:15:51.720 YvetteRuiz: Yeah.
93 00:15:51.720 ⇒ 00:15:55.500 Pranav Narahari: I’m happy to do them whenever. That’s one thing I mentioned to Janiece, too. I was just like.
94 00:15:56.160 ⇒ 00:15:59.580 Pranav Narahari: ping me, I’m happy to hop on a call, because I feel like
95 00:16:00.080 ⇒ 00:16:10.900 Pranav Narahari: I don’t want to wait a week if there’s something on your mind, because if it requires us to course correct, like, I don’t want to waste days. I’d rather just, like, talk about it. Yeah.
96 00:16:11.340 ⇒ 00:16:14.189 YvetteRuiz: Okay, so thank you for that. So…
97 00:16:14.210 ⇒ 00:16:28.179 YvetteRuiz: going back to the migration piece of it, so I just want to make sure… so, on the accuracy piece of it, you know, we’re tapping into Gemini, and this is one thing I was… I had a meeting last week with Matt and Bobby, and I was trying to explain.
98 00:16:28.180 ⇒ 00:16:41.459 YvetteRuiz: the speed and how all that is. So, a couple things here, is how are… what is our speed today? Because I don’t even know what that is, right? I kind of pulled a number, and so I want to know what that is. What’s the average today?
99 00:16:41.650 ⇒ 00:16:50.139 YvetteRuiz: where are we headed to once this migration? What is that? Because that’s… that’s gonna be key for me to know. Do you know what that is today, and where… where we’re headed?
100 00:16:50.900 ⇒ 00:16:54.780 Pranav Narahari: Yeah, so I can definitely pull that number for you. I don’t have it top of mind.
101 00:16:54.780 ⇒ 00:16:55.180 YvetteRuiz: Okay?
102 00:16:55.180 ⇒ 00:17:11.930 Pranav Narahari: I have as part of our deliverables for this week, is I want to give you a report at the end of the week about, okay, what was our latency, what was that average execution time, and what is it now? And so, that is going to all be, and we’re going to do a ton of testing.
103 00:17:11.930 ⇒ 00:17:12.510 YvetteRuiz: Okay.
104 00:17:12.910 ⇒ 00:17:18.470 Pranav Narahari: for me to give you, like, an execution time right now, I don’t want to give you a number that isn’t been, like, rigorously tested.
105 00:17:18.950 ⇒ 00:17:19.560 YvetteRuiz: Okay.
106 00:17:19.560 ⇒ 00:17:24.100 Pranav Narahari: Because I can give you a time if it doesn’t show exactly… we’re gonna have a different problem if it’s not actually.
107 00:17:24.670 ⇒ 00:17:28.669 Pranav Narahari: showing up as accurate when, you know, the CSRs are using Andy.
108 00:17:28.670 ⇒ 00:17:29.090 YvetteRuiz: Okay.
109 00:17:29.710 ⇒ 00:17:40.809 Pranav Narahari: I will get that to you this week. That’s definitely, like, a deliverable. In the meantime, is it useful for you to know, like, what is just, like, the current execution time?
110 00:17:41.080 ⇒ 00:17:44.469 YvetteRuiz: Yeah, if you can give me that now, that’d be great.
111 00:17:44.470 ⇒ 00:17:47.139 Pranav Narahari: Sure, yeah, so I can’t give it to you in this meeting. I’m gonna work with.
112 00:17:47.140 ⇒ 00:18:04.330 YvetteRuiz: No, that’s fine. If you could just send it to me, that’ll be perfect. So then the other piece, tying to that, as you were… you were talking about Gemini, so the accuracy piece of it. So the… what is… what is that going to help us with? Help me understand that, because that’s… this is where I want to get a full understanding, so…
113 00:18:04.510 ⇒ 00:18:06.630 YvetteRuiz: you know how Andy was giving us…
114 00:18:06.980 ⇒ 00:18:15.759 YvetteRuiz: different answers, you know, even though we were asking it the same way. Some of it was tied to the center docs, right? And, you know, where it was kind of picking the stuff, right?
115 00:18:15.970 ⇒ 00:18:23.520 YvetteRuiz: But what is the Gemini and the things that we’re doing right now? How is that going to help? Can you kind of just give me a little bit of insight?
116 00:18:23.780 ⇒ 00:18:43.700 Pranav Narahari: Yeah, so I can… and keep asking more questions for, like, how deep technically you want me to go. These new models just do a better… they do a better job of, like, taking in the context that you give them, so the context being central docs, and assessing what is accurate information, what is inaccurate info, or irrelevant information.
117 00:18:43.700 ⇒ 00:18:44.340 YvetteRuiz: catch up.
118 00:18:44.340 ⇒ 00:18:56.420 Pranav Narahari: And so, how are we gonna test that? Is, we have a bunch of triage tickets that were, that came up, and they weren’t due to missing information, they were due to structuring issues.
119 00:18:56.420 ⇒ 00:19:09.790 Pranav Narahari: And so, what we’re gonna do is we’re going to create a script to regen… like, rerun each of these individual questions that previously gave wrong answers, or…
120 00:19:09.970 ⇒ 00:19:18.179 Pranav Narahari: Answers that said, like, okay, we cannot find this information for you, and now we’re gonna show you that using this new model, as well as, you know,
121 00:19:18.540 ⇒ 00:19:24.199 Pranav Narahari: migrate… just doing the full migration, we’re able to give you these answers now. And so.
122 00:19:24.330 ⇒ 00:19:42.029 Pranav Narahari: there’s, since we’re moving so fast, like, there’s a few different things we’re testing, right? We’re testing, like, the new models that we’re using, the new Gemini models, we’re also testing with, like, the restructured central docs, and then we’re also testing with, just the overall migration. And so.
123 00:19:42.220 ⇒ 00:19:46.319 Pranav Narahari: we were using, like I mentioned before, we were using a different service for, like, our
124 00:19:46.610 ⇒ 00:19:49.039 Pranav Narahari: the brain, you can think of it, of, like, Andy.
125 00:19:49.620 ⇒ 00:19:59.349 Pranav Narahari: and now we… you mentioned, like, there was a plateau. We were noticing that plateau as well with our… our backend service, and so now we’re… we’re completing, basically, the…
126 00:19:59.410 ⇒ 00:20:12.970 Pranav Narahari: the migration into something that’s much more scalable, something that’s going to allow us to use, like, the latest and greatest, like, models, like Gemini. And so with the combination of all three of these things, you should notice, like, with these triage tickets.
127 00:20:12.970 ⇒ 00:20:21.519 Pranav Narahari: We’re going to be able to solve a lot more of them more accurately, and then on top of that, what it’s doing is like… it’s like two birds with one stone, it’s also going to be much quicker.
128 00:20:22.160 ⇒ 00:20:25.619 YvetteRuiz: Gotcha, gotcha, okay. So, the,
129 00:20:26.030 ⇒ 00:20:29.630 YvetteRuiz: So, kind of piggybacking on that piece of it.
130 00:20:30.620 ⇒ 00:20:36.079 YvetteRuiz: because, again, it’s about speed, right? It’s about getting the CSR the answer that they need quickly, right?
131 00:20:36.850 ⇒ 00:20:44.659 YvetteRuiz: If we type in… so, like, let’s say, for instance, I’m using cancellation steps. If I just type in one word, customer moving.
132 00:20:45.250 ⇒ 00:20:58.069 YvetteRuiz: could it get me that quick to get me that information? Is this what this will help with as well? Or is there going to still be, you know, the questions the way they’re asked? Again, I’m asking this because, you know, when you’re thinking about
133 00:20:58.200 ⇒ 00:21:08.460 YvetteRuiz: how… how do you get the customer the answer? I just want to make sure that there’s no specific way to ask, or is there keywords that we can just go in there and type in that gives us a quick answer?
134 00:21:09.400 ⇒ 00:21:16.230 Pranav Narahari: So there’s different ways that we can build out this system. We can build, like, structured
135 00:21:16.610 ⇒ 00:21:31.460 Pranav Narahari: we can build, like, structured responses based on, like, keywords that are mentioned. So, like, let’s say, yeah, they just say two words, like, customer moving. If that is going to be, like, a common query that you notice that the CSRs are gonna say, then, yeah, we could…
136 00:21:31.460 ⇒ 00:21:38.919 Pranav Narahari: start talking about that triggering, like, a whole thing in, like, in our, like, our… the brain of, like, Andy.
137 00:21:41.090 ⇒ 00:21:48.889 Pranav Narahari: So what that would do, right, is that we would need from you probably a list of, like, okay, what are these most common phrases? .
138 00:21:48.890 ⇒ 00:21:51.670 YvetteRuiz: Most common keywords.
139 00:21:51.670 ⇒ 00:22:01.689 Pranav Narahari: And this can be, like, a joint thing that we work on together, is like, okay, we have a whole bank of, like, what are the prompts being asked to Andy currently, right?
140 00:22:02.040 ⇒ 00:22:16.989 Pranav Narahari: maybe do a first pass of, like, looking at what is, like, some certain categories, and maybe, like, short keyword questions that are being asked to Andy, and then you can let me know, okay, yeah, this is something that we want… that is definitely going to be asked a lot.
141 00:22:17.050 ⇒ 00:22:27.359 Pranav Narahari: we need it to be, like… we needed to be quick, we needed to, like… Andy should understand, basically, exactly what we’re asking it here, even though we’re only giving it two words.
142 00:22:28.290 ⇒ 00:22:46.569 YvetteRuiz: Okay, okay, all right. Well, thank you for that. These are just kind of the things that I’m kind of thinking of with the build, because, you know, when the CSRs go in, and I think you guys already have provided us, just like you said, the questions that are being asked today. We’re already kind of seeing some of those things, but, you know, again.
143 00:22:47.030 ⇒ 00:23:05.370 YvetteRuiz: I’m just, in my mind, you know what I mean? If I’m going to give the best customer experience, you know, what… I want to be able to train our agents confidently, let them know, like, how do you ask the questions, you know what I mean? There are certain things that you can ask for. And is also Andy going to be able… and this may be… this is something that I’ve been trying to get some answers to.
144 00:23:05.660 ⇒ 00:23:13.230 YvetteRuiz: is, like, the probing questions. Is it gonna get to that level to where if I do say, okay, moved, right?
145 00:23:13.340 ⇒ 00:23:23.630 YvetteRuiz: will Andy be able to respond something like, are you asking about a customer moving, or will it get to that level, or is that still something that we’ve got to work together to build?
146 00:23:24.520 ⇒ 00:23:28.970 Pranav Narahari: I think we need to work together a little bit on just…
147 00:23:29.300 ⇒ 00:23:33.899 Pranav Narahari: And it’s gonna be kind of hand-in-hand with what we… you actually just mentioned, with just, like.
148 00:23:34.360 ⇒ 00:23:38.059 Pranav Narahari: you… I’m giving you some context, go run with it, versus…
149 00:23:38.060 ⇒ 00:23:38.590 YvetteRuiz: scene.
150 00:23:38.590 ⇒ 00:23:44.990 Pranav Narahari: If we’re not confident on exactly the type of question you’re asking, then we ask additional probing questions.
151 00:23:46.060 ⇒ 00:23:51.629 Pranav Narahari: I tend to believe, like, that is based on what I understand about what Andy’s…
152 00:23:51.730 ⇒ 00:24:09.449 Pranav Narahari: what we need from Andy, that is… that’s kind of the experience we want. If from the… if from the prompt we’re not getting the exact kind of information that we need to answer the question, then we should ask a probing question to then further define what we need to provide to the user of Andy.
153 00:24:09.450 ⇒ 00:24:09.950 YvetteRuiz: Yeah.
154 00:24:10.590 ⇒ 00:24:16.139 Pranav Narahari: Okay, so in that way, though, it’s going to be a little bit slower, right? Because it’s going to come up with a probing question, and then.
155 00:24:16.140 ⇒ 00:24:16.840 YvetteRuiz: Yeah, yep.
156 00:24:16.840 ⇒ 00:24:21.059 Pranav Narahari: respond to that probing question, and then we have to respond back.
157 00:24:21.060 ⇒ 00:24:21.690 YvetteRuiz: Okay.
158 00:24:21.960 ⇒ 00:24:25.549 Pranav Narahari: And so, maybe what me and you talk about is, like.
159 00:24:25.990 ⇒ 00:24:38.360 Pranav Narahari: we kind of, like, we set a max on those probing questions, or if we feel like these probing questions are just becoming, like, yeah, of course that’s what it means. Then we can talk about, like, okay, maybe we don’t need the probing questions.
160 00:24:38.560 ⇒ 00:24:40.810 Pranav Narahari: The best scenario would be just, like.
161 00:24:41.180 ⇒ 00:24:45.210 Pranav Narahari: Andy’s brain understands what you mean, right? Yeah.
162 00:24:45.210 ⇒ 00:24:46.439 YvetteRuiz: Yeah, no, for sure.
163 00:24:46.440 ⇒ 00:24:58.280 Pranav Narahari: questions as possible. So what we should do first is just, like, okay, what is our baseline? Like, is our baseline that it’s, like, getting 70% of the questions right? Maybe for certain things, it’s getting it 99% right?
164 00:24:58.280 ⇒ 00:25:02.409 YvetteRuiz: then we don’t need a probing question there, but maybe for other things… Absolutely. Okay.
165 00:25:02.410 ⇒ 00:25:20.129 YvetteRuiz: That makes… that makes total sense, okay? Yeah, because you’re absolutely right. I don’t want to go granular down this rabbit hole, where it’s just kind of, what about this, what about that, you know what I mean? But I do want to be able… there are some things that make sense, you know what I mean? Like, residential versus commercial, right? You know, some of those things make sense, so I just want to see how…
166 00:25:20.220 ⇒ 00:25:21.810 YvetteRuiz: how,
167 00:25:21.930 ⇒ 00:25:27.530 YvetteRuiz: How accurate could we get him with getting us to, and quickly, to get us those answers?
168 00:25:27.800 ⇒ 00:25:32.110 Pranav Narahari: Yeah, okay. That’s great.
169 00:25:32.550 ⇒ 00:25:34.920 Pranav Narahari: we… yeah, so we talked about Central Doc.
170 00:25:35.070 ⇒ 00:25:41.749 Pranav Narahari: For zip codes, too, this… I feel like all of these kind of have the same, like, KPIs that you’re looking for.
171 00:25:42.160 ⇒ 00:25:53.289 Pranav Narahari: So, it’s more or less the same thing. Like, what we noticed was just, like, it’s really similar to Central Doc, it was just a restructuring thing. Like, we’re noticing that Andy wasn’t able to assess, like, the…
172 00:25:53.290 ⇒ 00:26:06.020 Pranav Narahari: or pull the accurate information because of the structure of the document. With zip codes, it’s actually a tiny bit different, that, like, when we were creating this document, since, like, the zip codes Google Sheet was
173 00:26:06.020 ⇒ 00:26:24.220 Pranav Narahari: quite, it was, like, the format was very, like, different from sheet to sheet. We were missing a few things. But now, with, like, this new iteration that Casey just created, we’re… and we’re able to test this against, like, all of the triage tickets, we’re noticing we’re missing way, way less.
174 00:26:24.270 ⇒ 00:26:29.489 Pranav Narahari: And so I think that’s what the update that Casey gave last week was.
175 00:26:29.490 ⇒ 00:26:30.140 YvetteRuiz: Yep.
176 00:26:30.420 ⇒ 00:26:35.099 Pranav Narahari: So… that seems to be moving in the right direction as well.
177 00:26:35.560 ⇒ 00:26:52.309 Pranav Narahari: Going back to the thing that you talked about, too, just, like, with constant meetings, like, I’m super happy to do that, like, you know, today we talked for 30 minutes, just because, like, it’s, like, one of our first ones. In the future, I can see them just being, like, 10 to 15 minutes every, like, once a week, twice a week. Yeah.
178 00:26:52.600 ⇒ 00:27:00.519 Pranav Narahari: And I think that’s just gonna help us just, like, better align on things, instead of being like, okay, I’ll ask him that on Thursday or Friday or Monday or whatever.
179 00:27:00.520 ⇒ 00:27:01.339 YvetteRuiz: Right, right.
180 00:27:01.340 ⇒ 00:27:04.839 Pranav Narahari: about it instantly will actually help us a lot, too.
181 00:27:04.840 ⇒ 00:27:05.330 YvetteRuiz: Yeah.
182 00:27:05.330 ⇒ 00:27:07.260 Pranav Narahari: So, yeah, that’d be my preference.
183 00:27:07.630 ⇒ 00:27:21.900 YvetteRuiz: Okay, yeah, then that’ll be… that’ll work then. Okay. Well, I’m… so you’re going to get back to me on the speed that it is today, and then kind of we’ll set some goals after, you know, we start testing that and kind of see how that…
184 00:27:22.690 ⇒ 00:27:26.809 YvetteRuiz: speed, comes down, or quicker, I’m sorry.
185 00:27:27.010 ⇒ 00:27:37.349 Pranav Narahari: Yeah, definitely. And so, I sent also just, like, an email earlier today, just for you to kind of, like, see, like, okay, what are the deliverables, for this week?
186 00:27:37.690 ⇒ 00:27:56.999 Pranav Narahari: that every week, just so, like, we’re just aligned. You can res… I’m not actually sure if you… if anybody responded to that yet, because I will be looking at that in the future, like, just to make sure, like, if you guys… you or Janiece have any comments, on that, like, we should be moving in a different direction, or you need more details on something.
187 00:27:57.000 ⇒ 00:28:00.720 Pranav Narahari: I think Monday’s a good day for us to just, like, get that sorted, so then…
188 00:28:00.720 ⇒ 00:28:09.999 Pranav Narahari: the rest of the week, we’re just, like, fully aligned on what we need to build, and, like, I don’t need to keep on telling Mustafa, Casey, Sam, like, oh, can you do this instead of that, you know?
189 00:28:10.000 ⇒ 00:28:22.299 YvetteRuiz: Right, right. So you are going to be joining our Monday meeting, then, the way you did this past Monday. We’re going to continue meeting, okay. So, is… so Amber is still going to be part of that, or she’s not going to be part of that, that’s going to be you?
190 00:28:22.810 ⇒ 00:28:34.330 Pranav Narahari: I’m going to ask her if, I think for the first few, she’ll be there. Okay. If it feels like, you know, she’s just kind of, like, she doesn’t really need to be there, you know, she’s, like, happy to kind of, like.
191 00:28:34.330 ⇒ 00:28:34.660 YvetteRuiz: pricing.
192 00:28:34.660 ⇒ 00:28:40.160 Pranav Narahari: then I can give her her time back on that. Okay. But for next week, she’ll definitely be there.
193 00:28:40.860 ⇒ 00:28:41.750 YvetteRuiz: Okay.
194 00:28:41.750 ⇒ 00:28:42.310 Pranav Narahari: Yep.
195 00:28:44.250 ⇒ 00:28:45.150 YvetteRuiz: Okay, great.
196 00:28:46.740 ⇒ 00:28:47.930 YvetteRuiz: Okay.
197 00:28:48.560 ⇒ 00:28:53.709 YvetteRuiz: What other questions do I have? So, right now, again, full focus are these three topics, which, again, I’m…
198 00:28:53.830 ⇒ 00:28:56.130 YvetteRuiz: perfectly.
199 00:28:56.520 ⇒ 00:29:15.139 YvetteRuiz: good with right here, and then, like I said, once we start seeing… getting a little bit more results on this, the next level, again, is just going to be tying that… more of the KPIs on the dashboard, because we really haven’t… that was the big piece in the beginning, you know, when we met with Utam, is really utilizing that dashboard and really using…
200 00:29:15.140 ⇒ 00:29:25.940 YvetteRuiz: training the leaders to be able to look at that and be able to see all that piece of it, but again, I don’t feel like we can move forward with that until we kind of get this foundation right here up and going.
201 00:29:26.130 ⇒ 00:29:30.580 Pranav Narahari: Totally, totally. And so it seems like a kind of our, like.
202 00:29:30.970 ⇒ 00:29:33.599 Pranav Narahari: Short-term, medium-term goal is to, like.
203 00:29:34.070 ⇒ 00:29:41.739 Pranav Narahari: Keep accuracy high, and make it even more accurate, and then keep the execution time low.
204 00:29:41.740 ⇒ 00:29:42.360 YvetteRuiz: Yes.
205 00:29:42.360 ⇒ 00:29:48.790 Pranav Narahari: And then once we feel like, okay, we’re at a point where, like, okay, this is really good, then let’s talk about that next piece, which is, like, that dashboard.
206 00:29:49.170 ⇒ 00:30:11.430 YvetteRuiz: Yes, yes, and then if… and then… then… will you be providing… we could talk more about that, those questions, I think that’s going to be the other big thing, is really doing the deep dive, or I don’t know how we’re going to pull that, as far as, like, okay, how many of them are giving the right answers, you know, and how many of them are not giving the right answers, so we can determine, should we start plugging in probing questions in there?
207 00:30:12.330 ⇒ 00:30:20.439 Pranav Narahari: Yeah, so I think, like, a couple weeks ago, or maybe even longer at this point, like, I created, like, a weekly summary. That’s what you’re talking about? We can do that on a weekly basis, for sure.
208 00:30:20.440 ⇒ 00:30:43.119 YvetteRuiz: Okay, that’ll be perfect, because if we can kind of get eyes on that, and kind of see, okay, look, this is what’s coming through. I mean, you guys see it on a regular basis, I don’t. So to be able to see that, that’ll be great, to be able to get view of, okay, what… okay, the percentage of our questions, all that is… seems to be great. It’s the thumbs up. This is what we’re… what’s being asked. These are still some of the ones that are not
209 00:30:44.370 ⇒ 00:30:47.259 YvetteRuiz: Generating, or questionable, or whatever.
210 00:30:47.540 ⇒ 00:30:49.740 Pranav Narahari: Yeah, okay, that sounds good,
211 00:30:50.100 ⇒ 00:30:55.080 Pranav Narahari: I do think, like, that will be really good for both you and for me to just kind of, like.
212 00:30:55.230 ⇒ 00:30:59.890 Pranav Narahari: iterate on Andy as quick as po- like, as quickly as possible.
213 00:31:00.060 ⇒ 00:31:12.089 Pranav Narahari: So, I think Thursday’s a perfect time for us to kind of, like, go over what did it come up with, what does that report look like, what are the ones that, like, it, like, solved accurately, which is the ones that it, like, maybe missed the mark.
214 00:31:12.090 ⇒ 00:31:20.390 Pranav Narahari: And then we can all discuss. You guys will have, like, the industry knowledge to, like, tell us, like, okay, this may be why, and we’ll be able to kind of give, like, the…
215 00:31:20.780 ⇒ 00:31:25.040 Pranav Narahari: the ANDI, like, technical knowledge for, like, okay, this is why I missed it. So I think that.
216 00:31:25.040 ⇒ 00:31:48.749 YvetteRuiz: Exactly, yeah, you’re spot on, and that’s where I went to in the beginning. It’s like, what… is it on our end that we gotta fix it? Is it on… so that’s what I really want to get down to, is where is the gap? You know what I mean? And what do we got to do? Because at the end of the day, if we can make it, you know, 98% accurate, and it’s getting us what we need to do, and the agents, the speed is there, our agents are really…
217 00:31:48.950 ⇒ 00:31:58.020 YvetteRuiz: you know, trusting it, because right now, I could tell you, I mean, I was listening to someone earlier, they’re going through the spreadsheets, I was running reports on hold time still, and I’m like.
218 00:31:58.270 ⇒ 00:32:07.160 YvetteRuiz: why do I have these agents on hold when they could be getting these answers over here? Is it still… what is the under… where’s the issue?
219 00:32:07.360 ⇒ 00:32:08.330 Pranav Narahari: Yeah.
220 00:32:09.420 ⇒ 00:32:11.149 Pranav Narahari: Yeah, and so…
221 00:32:12.000 ⇒ 00:32:15.219 Pranav Narahari: I understand that, too, and it’s really good for…
222 00:32:15.370 ⇒ 00:32:33.120 Pranav Narahari: I think what we’re doing right now is going to be… we’re working in the right direction. I see, like, the work that, like, Casey, Mustafa, and Sam are, like, giving me updates on, and I’m like, this is exactly what we need to be doing to solve this issue that you’re talking about. And so, I think in the coming weeks, like, once we kind of finish this migration.
223 00:32:33.680 ⇒ 00:32:35.730 Pranav Narahari: We’ll see a big difference.
224 00:32:35.730 ⇒ 00:32:36.200 YvetteRuiz: Okay.
225 00:32:36.200 ⇒ 00:32:40.340 Pranav Narahari: then we’ll continue refining it just from, like, those Thursday calls.
226 00:32:40.690 ⇒ 00:32:45.610 Pranav Narahari: Like, where we kind of talk about just, like, what was the… the past week’s, like, snapshot of, like.
227 00:32:45.780 ⇒ 00:32:54.319 Pranav Narahari: all the questions. That… that cycle, like, we’re gonna notice, like, after a few weeks that there’s less and less stuff to talk about, which is gonna be a good thing.
228 00:32:54.470 ⇒ 00:32:59.210 YvetteRuiz: Yeah, yeah, exactly, because then we can really start work… okay, what’s our next move?
229 00:32:59.210 ⇒ 00:33:00.630 Pranav Narahari: Exactly. Yeah.
230 00:33:00.880 ⇒ 00:33:01.960 YvetteRuiz: Okay.
231 00:33:01.960 ⇒ 00:33:02.620 Pranav Narahari: This was great.
232 00:33:02.620 ⇒ 00:33:07.630 YvetteRuiz: Yeah, no, thank you so much for taking the time to do this. I really, really appreciate it.
233 00:33:07.630 ⇒ 00:33:17.909 Pranav Narahari: Yeah, I appreciate it a lot, Yvette. And I told Janice this too, I think you guys are more comfortable, like, email, right? Like, in Slack, I haven’t seen too many communications.
234 00:33:17.910 ⇒ 00:33:19.170 YvetteRuiz: Oh, yeah.
235 00:33:19.170 ⇒ 00:33:26.000 Pranav Narahari: But I’m happy to, you know, communicate in either channel. Would you say email is better, or…
236 00:33:26.000 ⇒ 00:33:39.810 YvetteRuiz: Yeah, email’s better right now. I’m not… I’m just not in the habit of going to Slack. I use Google Chat a lot. So we have those chats that you guys have created, so if you want to communicate, I’m cool with that as well.
237 00:33:40.110 ⇒ 00:33:46.800 Pranav Narahari: Perfect, yeah, I think Google Chat just seems like the quickest way to do it. Email sometimes can get lost, right?
238 00:33:47.090 ⇒ 00:33:55.659 Pranav Narahari: Perfect. So yeah, whenever you want to meet one-on-one, and you honestly might hear from me more from meeting one-on-one in the beginning, for sure, I’ll message you through there.
239 00:33:56.130 ⇒ 00:34:17.270 YvetteRuiz: Okay, well, I appreciate you taking the time to do this, I really, really do, and I’m sure that I will have other questions, but I’ll reach out to you. You know, one of the things that I’m really, really working towards is really making sure, because there’s so much with the AI, with this backend stuff, not that I want to get too granular, but I also want to know, what does it mean? What does that mean, and how is it affecting
240 00:34:17.270 ⇒ 00:34:18.459 YvetteRuiz: what we do.
241 00:34:18.469 ⇒ 00:34:21.939 Pranav Narahari: For you, too, it’s probably, like, you want to be able to communicate to others, right?
242 00:34:21.940 ⇒ 00:34:22.770 YvetteRuiz: Yes.
243 00:34:22.770 ⇒ 00:34:36.529 Pranav Narahari: I want to set you up in a way where it’s like, this doesn’t seem like a black box. It probably also doesn’t make sense for me to tell you about our entire technical implementation and, like, all the… but there’s a middle ground there where I feel like we can definitely work on that.
244 00:34:36.780 ⇒ 00:34:39.629 YvetteRuiz: Yeah, for sure. Well, I appreciate that very much.
245 00:34:39.630 ⇒ 00:34:42.049 Pranav Narahari: Thank you so much Have a good rest of your day.
246 00:34:42.409 ⇒ 00:34:43.799 YvetteRuiz: You too, take care.
247 00:34:43.799 ⇒ 00:34:45.499 Pranav Narahari: Bye. Bye. Bye.