Meeting Title: Robert Tseng’s Personal Meeting Room Date: 2025-04-28 Meeting participants: Robert Tseng, Uttam Kumaran, Daniel Saltus
WEBVTT
1 00:01:26.960 ⇒ 00:01:27.546 Robert Tseng: Hey! Dan!
2 00:01:27.840 ⇒ 00:01:29.459 Daniel Saltus: Hey, Robert, what’s going on today.
3 00:01:31.030 ⇒ 00:01:32.569 Robert Tseng: Long time, no see.
4 00:01:32.570 ⇒ 00:01:34.541 Daniel Saltus: No, it’s been a little while.
5 00:01:35.080 ⇒ 00:01:36.159 Daniel Saltus: Where have you been?
6 00:01:36.160 ⇒ 00:01:38.829 Daniel Saltus: Pretty busy? Pretty good. How about you guys?
7 00:01:39.930 ⇒ 00:01:42.109 Robert Tseng: Good. Yeah. Busy, busy, but good.
8 00:01:42.400 ⇒ 00:01:46.520 Daniel Saltus: Yeah, good. I know it’s sort of the story of life these days, I guess, or as long as you’re
9 00:01:46.630 ⇒ 00:01:49.730 Daniel Saltus: moving from thing to thing. That’s the that’s the goal.
10 00:01:50.360 ⇒ 00:01:51.399 Uttam Kumaran: Hey, Dan!
11 00:01:51.400 ⇒ 00:01:53.933 Uttam Kumaran: Nice to meet you. My name is Zootam.
12 00:01:54.827 ⇒ 00:02:01.679 Uttam Kumaran: Robert and I both run Brain Forge. So just just listening in curious about what you’ll think about the demo today.
13 00:02:01.980 ⇒ 00:02:05.999 Daniel Saltus: Yeah, absolutely thanks for for hosting it and doing, putting all together. I’m looking forward to seeing it.
14 00:02:07.710 ⇒ 00:02:22.770 Robert Tseng: I think what. So what we’ll do is I’ll I’ll kind of pull up a the demo. I have every URL here, and I can share with you. We also recorded like the video in case whatever we say goes over your head, you can review it in your own time as well.
15 00:02:24.170 ⇒ 00:02:28.030 Robert Tseng: Okay, give me a second to pull it up.
16 00:02:29.920 ⇒ 00:02:31.160 Robert Tseng: See?
17 00:02:34.700 ⇒ 00:02:35.670 Robert Tseng: Bear with
18 00:02:42.920 ⇒ 00:02:44.180 Robert Tseng: great.
19 00:02:47.555 ⇒ 00:02:47.940 Robert Tseng: hey?
20 00:02:47.940 ⇒ 00:02:49.209 Robert Tseng: Let me know where you can see that.
21 00:02:50.260 ⇒ 00:02:51.510 Daniel Saltus: All right. I can. Yes.
22 00:02:51.890 ⇒ 00:02:56.542 Robert Tseng: Okay, great. So basically like this is, you know, this ui is like
23 00:02:57.030 ⇒ 00:03:00.179 Robert Tseng: something through together. But under the hood, what we’ve done is, we’ve
24 00:03:00.330 ⇒ 00:03:23.690 Robert Tseng: uploaded those reacted medical records that you gave us. Yeah, we kind of just adapted something that we had built for another client before. And and also in the health kind of patient health record space. And yeah, we just wanted to show you the capability. So all these Pdfs, we just kind of click into it. It extracts this this information. So
25 00:03:24.068 ⇒ 00:03:32.400 Robert Tseng: obviously, we can tweak it as we as you see fit. But yeah, we we kind of already trained it to, you know, extract basic information.
26 00:03:32.790 ⇒ 00:04:01.379 Robert Tseng: You know, id states of service. We tagged specific. We found this to be helpful for specifically medical records, because there are certain, you know, keywords, that you’re looking for, that really influence the outcome of the you know, whatever the case is about and then, you know, this is pretty standard stuff. It’s just like the metadata for the file itself. If there was any diagnosis that was made good. And then, yeah, the the medical code, which I’m sure is
27 00:04:01.440 ⇒ 00:04:08.336 Robert Tseng: pretty important to what you’re looking for and then just summaries of like what was actually done.
28 00:04:08.820 ⇒ 00:04:12.389 Robert Tseng: and so, yeah, I think this is we. We tried to
29 00:04:13.510 ⇒ 00:04:18.719 Robert Tseng: trying to extract something that’s just 2 to 3 sentences. But obviously these documents are
30 00:04:18.899 ⇒ 00:04:32.360 Robert Tseng: very long. So if we wanted it to be more robust, we could definitely train it to do better there. Yeah. And so anyway, just walking you through the different sections. And so, yeah, we’ve kind of got into a place where this is kind of the standardized data format for.
31 00:04:32.360 ⇒ 00:04:32.720 Daniel Saltus: I mean that.
32 00:04:32.720 ⇒ 00:04:56.359 Robert Tseng: Records themselves. And yeah, I think what would be, you know, if you were to? I would like to just get your feedback, you know, 1st glance like, if this is a little close to what you had in mind, different. You know, areas that you wanted to double click into on like, how we, how we did something, or if you wanted to expand on any of these these sections, I think that would be helpful for us to consider in this conversation.
33 00:04:56.360 ⇒ 00:04:57.080 Robert Tseng: Sure.
34 00:04:57.300 ⇒ 00:05:05.980 Daniel Saltus: Yeah, I. So at 1st glance, it looks like you sort of have complete included. What I would mostly need to know.
35 00:05:06.330 ⇒ 00:05:16.440 Daniel Saltus: As a general rule. If we’re looking at a medical record, the the impression part is is probably the most important part
36 00:05:16.600 ⇒ 00:05:17.570 Daniel Saltus: got it
37 00:05:17.860 ⇒ 00:05:40.579 Daniel Saltus: for us just because we want, you know. Let’s say we get that. You know. I don’t know. I can’t remember the number of pages we sent to you guys right, but we sort of would like to be able to look at a sort of like snapshot of each record and understand, you know, like what care is provided, and so we sort of know it was totally in a certain doctor, and you know they were doing certain things as listed in here. But then the outcomes are are the kind of the critical part.
38 00:05:41.153 ⇒ 00:05:47.110 Daniel Saltus: So we do a lot of accidents injuries. So, for example, like, you know, somebody goes to the hospital, complaining about
39 00:05:47.770 ⇒ 00:06:01.699 Daniel Saltus: their low back after getting rear ended in car accident. Did they make any other complaints about either their neck, their head, or their shoulder stuff like that right where we’d have to review the entire record. And we’d like to be able to know, like that kind of a thing just at a glance.
40 00:06:02.165 ⇒ 00:06:14.440 Daniel Saltus: That so it looks like you sort of, you know. Got there? I mean, I would have to now cross, check the record and and see this, you know. Compare the 2 you know, to really see if it’s perfectly thorough.
41 00:06:15.420 ⇒ 00:06:26.950 Robert Tseng: Okay, got it? Yeah. And then, you know, so things like that where you’re like, hey, this impression section, that’s the most poor. Obviously, we’ll pull it up to the front and to the top. And then, yeah, if you wanted to list out like.
42 00:06:27.110 ⇒ 00:06:44.150 Robert Tseng: Hey, typically, what you’ve seen is especially if there are multiple things that are going on. You know I I kind of did a quick glance and looked at it, using my untrained eye, I suppose, and saw that I feel like we caught like there were multiple things that were mentioned. We’re able to extract distinct things. But sometimes it’s like.
43 00:06:44.300 ⇒ 00:07:09.170 Robert Tseng: Hey, maybe it’s like a general issue. There’s like a hierarch we need to like, have a general issue. And then there’s like multiple sub issues in it. So if there’s any like kind of more semantic training that we need to do there where? Like, yeah, like, I think that’s that’s that’s that’s probably where we would like to maybe take like an example of like a summary that you thought was great, that your team had written before. We would just basically plug that into the model so that they can pattern match against that.
44 00:07:09.170 ⇒ 00:07:10.559 Daniel Saltus: Got it. Got it?
45 00:07:10.670 ⇒ 00:07:13.519 Daniel Saltus: Yeah, yeah, that that we could certainly do.
46 00:07:16.370 ⇒ 00:07:20.700 Daniel Saltus: I’m trying to, you know, rather than try to think of one off the top of my head, which I know is not gonna work.
47 00:07:21.210 ⇒ 00:07:21.690 Robert Tseng: No worries.
48 00:07:21.690 ⇒ 00:07:27.060 Daniel Saltus: Or at least gonna result in us just sitting here while I, while I, you know, think so. There’s no reason.
49 00:07:27.060 ⇒ 00:07:29.770 Robert Tseng: You sent us different formats. So we can just kind of click into and show kind of.
50 00:07:29.770 ⇒ 00:07:30.430 Robert Tseng: yeah, go ahead.
51 00:07:30.430 ⇒ 00:07:35.623 Robert Tseng: Changed. So, so yeah, like this one, I think just like a surgery note.
52 00:07:36.520 ⇒ 00:07:53.000 Robert Tseng: Yeah, this one had a lot more historical date, patient history data that needed to be brought up. So yeah, I think that was interesting. Where yeah, like, there’s all these different types of history. And so something that we had to train. The model is like, Okay, well, don’t just like, list everything out bullets. But like, you know, there’s different categories.
53 00:07:53.660 ⇒ 00:07:56.540 Robert Tseng: So yeah, kind of getting with
54 00:07:57.100 ⇒ 00:08:07.839 Robert Tseng: that. This kind of the 1st pass that like what we think those categories are. Yes, hopefully, that’s a helpful way to present it and then kinda keep going through this.
55 00:08:08.130 ⇒ 00:08:09.080 Robert Tseng: And
56 00:08:09.270 ⇒ 00:08:26.539 Robert Tseng: I correct me if I’m wrong. But these are all kind of like separate cases. They’re not necessarily linked. But if we wanted to do like a linking of like, hey? There’s multiple records for the same you know, same patient or something. Then, you know, that’s that’s maybe like a like a v 2, that we could kind of do, yeah.
57 00:08:26.540 ⇒ 00:08:43.649 Daniel Saltus: Just to to answer that before we move on, I forget. So they they are for the same lady. But she’s she wants to know whether we think she has a good medical malpractice case. And that is something that’s pretty advanced. And I don’t think that that the AI model is going to be able to make any assessment about
58 00:08:43.909 ⇒ 00:08:47.540 Daniel Saltus: the the standard of care of medical care right? But for.
59 00:08:47.540 ⇒ 00:08:47.880 Robert Tseng: Yeah.
60 00:08:47.880 ⇒ 00:08:53.090 Daniel Saltus: Purpose. If we could get like a medical summary
61 00:08:53.230 ⇒ 00:09:15.809 Daniel Saltus: that that would really kind of be ideal. So if you’re talking about like a version 2, you know, can’t see the total dim now, but like the dates of service like if we could kind of have like an ordered well, no, I mean in the background, I mean like that, you know. You know. What day did she go to Ami surgery? So June? And then the next visit was July 11, th and the next visit was July 13, th and kinda you know what happened on each day
62 00:09:16.440 ⇒ 00:09:25.629 Daniel Saltus: that that would be the ideal thing for us. So we just handed over a whole stack of records. We’d get essentially back like a a summary
63 00:09:25.910 ⇒ 00:09:32.310 Daniel Saltus: of where you know where the client went, what treatment the client underwent at each of those places.
64 00:09:33.000 ⇒ 00:09:50.319 Robert Tseng: Got it. Okay? Yeah. I mean, this was kind of like, the yeah, we tried to like, list it out. Here. But I mean, it’s just the diagnosis. But yeah, just kind of be substituting this with like a clear summary of like what actually happened there. Maybe it even is just like above before this list. Because.
65 00:09:50.320 ⇒ 00:10:06.760 Robert Tseng: yeah, for every patient, maybe there are multiple records. And then we just create like a summary kind of section at the top. But yeah, so I think you know, hopefully, you can see that this is pretty customizable, like, we want to work with you to get into the the best possible place. So you’re able to, you know, quickly. Get the information you need.
66 00:10:07.230 ⇒ 00:10:23.109 Robert Tseng: Yeah, I guess as far as like kind of you mentioned something interesting, not able not able to predict or determine. If you know, she received adequate care. I think that’s definitely something that could be trained over time, you know. I think if there are
67 00:10:23.783 ⇒ 00:10:35.996 Robert Tseng: you know, I would even venture like it’s it’s worth like trying to make that assessment and make that part of the the the, you know. Yeah, the actual live product that we ended up building for you.
68 00:10:36.410 ⇒ 00:10:38.279 Robert Tseng: yeah, if we end up having like a
69 00:10:38.620 ⇒ 00:10:44.569 Robert Tseng: predictive score of like, I don’t know some, some, some something there, that’s like, okay, that’s like a
70 00:10:44.810 ⇒ 00:10:49.579 Robert Tseng: high likelihood that this, that there is actually malpractice here? Some.
71 00:10:49.760 ⇒ 00:11:03.129 Robert Tseng: Yeah. Like, if we if we could actually build some of that deterministic capability into into this. So it’s not just summarization. But we also get to showcase some of the like, you know, some some of the
72 00:11:03.240 ⇒ 00:11:13.089 Robert Tseng: the predict, the predictive capabilities of this the solution. I think that that would be that’d be a cool you know thing to run run with with you guys as well.
73 00:11:13.250 ⇒ 00:11:19.386 Daniel Saltus: Yeah. So I guess then, if you look, if if you think that’s you know not
74 00:11:19.870 ⇒ 00:11:26.289 Daniel Saltus: not too great a leap, or you know whatever the the right phrase would be. Then, then, yeah. Then, having it
75 00:11:26.630 ⇒ 00:11:30.660 Daniel Saltus: into like a summary, and what I can do is I can try to find
76 00:11:32.350 ⇒ 00:11:48.609 Daniel Saltus: I can try to come up with something like that, because I know we we do those summaries, you know. So that I can. Yeah, I can try to pull something together to give you that. And then the if you could do even the next step would be to kind of making a, you know, have the
77 00:11:48.820 ⇒ 00:11:52.920 Daniel Saltus: the cognitive work done to sort of start to dive into.
78 00:11:53.290 ⇒ 00:11:57.749 Daniel Saltus: What does it mean rather than just? What is it, then? Certainly that could be helpful as well.
79 00:11:58.760 ⇒ 00:12:12.859 Robert Tseng: Yeah, so I mean, like, typically, kind of how this kind of solution evolves is like, yeah, the 1st thing is the to get the what is it? The doing? The summarization? Because that’s the most helpful. You know. And obviously, you know, with this approach you can see that, you know, there’s
80 00:12:13.520 ⇒ 00:12:33.969 Robert Tseng: I, I think, maybe the pricing model that you’re used to from record retrieval service wouldn’t necessarily apply here. There’s no per page charges or anything like we would just, you know, agree on kind of like a usage, you know, base, pricing kind of thing. And then, like, I think, for anything that’s more predictive that like next level, like deeper dive. If we’re like recommending or we’re we’re signaling to you like.
81 00:12:33.970 ⇒ 00:12:45.110 Robert Tseng: Hey, this is actually like a valid case. And it turns out to be like, we’re right then, like, maybe that’s like a that’s like that’s like bonus that we end up kind of being able to to, you know, to
82 00:12:45.110 ⇒ 00:13:10.520 Robert Tseng: to to layer into like the the compensation type to type performance right? Because we want the solution to behave as like like your best employee, or like to to have your mind pretty much. And that’s you know. I think that’s why we enjoy building these types of solutions and thinking about like, what does it actually need to do to make the help you make the intelligent decision? But you know not to jump
83 00:13:10.530 ⇒ 00:13:13.540 Robert Tseng: ahead. I think that’s no, no, of course, asking those questions too.
84 00:13:13.540 ⇒ 00:13:21.710 Daniel Saltus: Yeah, no, look. I mean, I you know the the more that we can tap into the better. I just, you know, don’t know exactly what there is
85 00:13:22.140 ⇒ 00:13:26.965 Daniel Saltus: out there, so to speak. That’s that’s available. So alright.
86 00:13:27.870 ⇒ 00:13:31.950 Daniel Saltus: I’m just trying to look at my initial notes from the client conversation.
87 00:13:32.400 ⇒ 00:13:33.000 Robert Tseng: Yeah.
88 00:13:33.230 ⇒ 00:13:51.269 Daniel Saltus: To see what she’s complaining about, because basically, what happened is, you know. So the client went, had her band removed, as you can see in the the 3, rd the second, 3, rd and 4th item there about the band removal and so then apparently the surgeon, like, left a piece of the band in her abdomen.
89 00:13:52.080 ⇒ 00:13:52.610 Robert Tseng: Oof.
90 00:13:52.610 ⇒ 00:13:56.979 Daniel Saltus: Right. And so that’s so. This is. The question is is that you know an actionable kind of a thing?
91 00:13:58.315 ⇒ 00:14:01.175 Daniel Saltus: So that’s the
92 00:14:04.840 ⇒ 00:14:25.579 Daniel Saltus: that would be kind of like the focus if we could, you know, if if it seems possible to start to get into, you know. Oh, hey! Like this note says that there was, in fact, the you know, 2 cm piece left, or, you know, something showed up on imaging that looked abnormal or whatever. But that’s the ultimate question. So if if it can start to answer.
93 00:14:26.200 ⇒ 00:14:33.979 Daniel Saltus: you know, or get into analyzing that, that would be sort of getting past, that what is it? And into that? What does it mean?
94 00:14:34.680 ⇒ 00:14:36.260 Robert Tseng: Okay, interesting.
95 00:14:37.850 ⇒ 00:14:40.380 Daniel Saltus: In this, at least in this specific instance, the.
96 00:14:40.380 ⇒ 00:14:40.770 Robert Tseng: Yeah.
97 00:14:40.770 ⇒ 00:14:43.019 Daniel Saltus: You know, each instance presents a different
98 00:14:43.210 ⇒ 00:14:46.660 Daniel Saltus: medical question, or medical and legal question.
99 00:14:48.140 ⇒ 00:14:59.390 Robert Tseng: I mean, I think, to kind of like gather the context of how we would really train the model to be able to detect. That would probably just be to look at. You know, previous
100 00:14:59.990 ⇒ 00:15:05.780 Robert Tseng: cases of kind of how how you you know. Maybe there’s a pattern of of like.
101 00:15:06.200 ⇒ 00:15:21.873 Robert Tseng: and the most basic is just like keywords like, say things that signal like, like, Hey, like something doesn’t seem like normal here, like, you know. Make sure you, you know. We we can. We can kind of guide the way that it extracts information that way.
102 00:15:22.320 ⇒ 00:15:31.964 Robert Tseng: but yeah, I think that’s all kind of part of the maybe further discovery process that would take us once we actually try to build something that’s like working towards a live solution for you.
103 00:15:33.350 ⇒ 00:15:51.163 Robert Tseng: yeah. Okay. Well, I mean, if no other questions on this for now well, yeah, just you know, this is really just like a demo to show you the capability. Hopefully, it kinda give you like, a good sense of like, yeah, this is what we’re able to. We’re able to do. This is, you know. And
104 00:15:51.660 ⇒ 00:16:02.580 Robert Tseng: yeah, we would love to kind of explore, like, what? What next steps look like like, how do we kind of get this? Yeah. Like, how? How would you? How would you want to kind of proceed from here?
105 00:16:02.620 ⇒ 00:16:07.129 Daniel Saltus: Yeah, I think the next thing would be if we got you a medical summary.
106 00:16:07.340 ⇒ 00:16:10.370 Daniel Saltus: Excuse me so that you could see
107 00:16:11.340 ⇒ 00:16:22.819 Daniel Saltus: what we’re looking for as like an end product. Then, you know, this could could turn in, or in theory could turn into that, and that would you know that would be that, I suppose, like the next test, to see.
108 00:16:23.950 ⇒ 00:16:27.460 Robert Tseng: Got it. Okay? Yeah. I mean, as far as like.
109 00:16:28.000 ⇒ 00:16:41.469 Robert Tseng: yeah, would you feel comfortable in making that like part of like some real scope that we can actually like. You know, work with you to develop on like, would you? It’s like I, typically.
110 00:16:41.890 ⇒ 00:16:46.219 Robert Tseng: we just show this Demo. And then we actually try to like, you know, put some dollar to to the.
111 00:16:46.220 ⇒ 00:16:46.710 Daniel Saltus: Yeah, yeah.
112 00:16:46.710 ⇒ 00:16:47.970 Robert Tseng: To the ground. Yeah.
113 00:16:47.970 ⇒ 00:16:56.988 Daniel Saltus: Right. So I understand what you’re saying. Yeah, you’re not looking to, just, you know. Do do gratis demo work for for in perpetuity, you know, I of course.
114 00:16:58.380 ⇒ 00:17:12.570 Daniel Saltus: I yeah. So I guess. Let me do 2 things. One is, speak to my partner, who has to be my wife, so I can’t avoid her. And then the second is, try to find out like a good a good medical summary
115 00:17:13.130 ⇒ 00:17:16.863 Daniel Saltus: to, you know, so we can come back to you with with something real.
116 00:17:17.130 ⇒ 00:17:17.599 Robert Tseng: Yeah.
117 00:17:17.609 ⇒ 00:17:19.979 Daniel Saltus: And and talk about it all right.
118 00:17:20.500 ⇒ 00:17:48.040 Robert Tseng: Okay, yeah. I mean, that sounds good. Yeah, we’ve we can kind of send you. I mean, I’ll send you this demo. You can go click around and walk your partner through it. And also, yeah, we’ll have a loom to kind of re, kind of basically re recap the capabilities here. And then we’ve we’ve we can send you some sample like scope of work. Just so you get a better sense of like, what does it actually look like to turn this into? Like?
119 00:17:48.450 ⇒ 00:17:51.060 Robert Tseng: you know, a paid engagement. Right time
120 00:17:51.060 ⇒ 00:17:56.370 Robert Tseng: lines that you can expect like, kind of how we work the whole process of yeah.
121 00:17:56.370 ⇒ 00:18:02.599 Robert Tseng: just to get an idea of, you know, if if we were to make improvements to this, what would what would that actually look like to engage with us?
122 00:18:02.820 ⇒ 00:18:03.730 Daniel Saltus: Understood.
123 00:18:04.110 ⇒ 00:18:04.680 Robert Tseng: Yeah.
124 00:18:05.110 ⇒ 00:18:09.679 Daniel Saltus: Alright. So yeah, if you could send me that, then obviously it’ll arm you with a little more information to to discuss with Hillary.
125 00:18:09.680 ⇒ 00:18:10.190 Robert Tseng: Risk.
126 00:18:10.530 ⇒ 00:18:30.149 Robert Tseng: Okay, cool. I mean, while I have you like, I’m curious, like, yeah, I mean, clearly, you know, this is not really, maybe the medical medical record was kind of like the starting point. But yeah, we’ve we’ve been like, kind of deep in the AI space at this point, and are curious about like trying to build solutions across, like the.
127 00:18:30.290 ⇒ 00:18:49.340 Robert Tseng: you know, and anything to support kind of the clients that we work with. So curious if there’s any other use cases that you can think of. Maybe this is, maybe it’s just something that we can kind of circle back on another another conversation. But yeah, I I think you know, I would say that anything that involves
128 00:18:50.354 ⇒ 00:19:04.960 Robert Tseng: document summarization like kind of routine routine tasks, stuff like that, like, I think, is all pretty fair game for us to really kind of help. Help you think through like how we can, we can build build a solution to help you with that.
129 00:19:05.130 ⇒ 00:19:11.749 Daniel Saltus: Yeah, I think the biggest immediate items would be getting
130 00:19:12.350 ⇒ 00:19:16.349 Daniel Saltus: obtaining medical records, you know, sending out requests, or or even
131 00:19:16.460 ⇒ 00:19:21.750 Daniel Saltus: I don’t know about that, but but getting both the records and the bills, that’s, I think, as we had talked about, there were certain.
132 00:19:21.960 ⇒ 00:19:29.150 Daniel Saltus: you know, issues in our practice. You know things that take up too much time, and that continues to be one of them. Certainly.
133 00:19:29.270 ⇒ 00:19:34.355 Daniel Saltus: that if we can get a excuse me
134 00:19:35.840 ⇒ 00:19:43.110 Daniel Saltus: you know some idea if if there’s some way to to engage the AI to to do that, that would also be very useful for us.
135 00:19:43.880 ⇒ 00:20:05.069 Robert Tseng: Yeah. So I mean something that we use internally, that you know it was even how we got we reach out to you like we have like. We have internal automation that goes and finds. Once we, we give them the right contacts, you know. They’re able to drop the message. Send that 1st 1st touch. Do follow ups and stuff like that. Like, yeah, we we could orchestrate that whole like
136 00:20:05.778 ⇒ 00:20:34.409 Robert Tseng: Outreach and follow up for you, especially if you’re gonna you have to go and you know petition from these medical providers, and and other other Fo, and I don’t necessarily know if the follow up is all always with the same folk, a person. But anything that’s like there’s a set way of like how you typically have to go and make make these requests like that’s you could totally automate that with AI and I think that’s something that we we have a lot of experience doing as well.
137 00:20:34.430 ⇒ 00:20:38.270 Daniel Saltus: Yeah, so that that certainly. You know, we could
138 00:20:38.500 ⇒ 00:20:40.750 Daniel Saltus: make it kind of like routine.
139 00:20:41.720 ⇒ 00:21:02.719 Daniel Saltus: you know, to request, you know, right right upon intake of a new case. Okay, you know, took an ambulance to the emergency room. Well, rather than have to rely on somebody to say, Oh, now I know I need to request the ambulance in the emergency room records to just, you know. Have that be part of, you know. Just just happen, so to speak. You know.
140 00:21:02.720 ⇒ 00:21:03.230 Robert Tseng: Yep.
141 00:21:03.230 ⇒ 00:21:10.900 Daniel Saltus: Like. Review the intake. Oh, they went to Westchester Medical Center in Westchester County, or Stanford Hospital, or Bridgeport Hospital, or wherever?
142 00:21:11.040 ⇒ 00:21:23.990 Daniel Saltus: Yeah. Oh, and it says that they took the ambulance there, and it says it was Amr, or whatever company. And just, you know know. Oh, here are the authorizations like. We’ll fill them out and we’ll get them out the door. That that would be would be wonderful.
143 00:21:25.050 ⇒ 00:21:32.772 Robert Tseng: Totally. Yeah, no, I can already see, like that full workflow that we would probably look to look to help you replace more or less.
144 00:21:33.070 ⇒ 00:21:33.660 Daniel Saltus: Right.
145 00:21:33.820 ⇒ 00:21:52.770 Robert Tseng: Yeah, I guess my question would be, once you get that like specificity of like, okay, it went to the hospital. How how easy it, is it to get the right contact like? Do you already have that? And it’s just like a limited set of providers that you work with? I imagine that’s probably more limited than we what we have to deal with. So yeah.
146 00:21:53.000 ⇒ 00:22:00.619 Daniel Saltus: Yeah. So it’s funny, because there are a huge number of repeat things. Right? I said, Stanford Hospital. We’re based in Stanford, Connecticut. That’s like.
147 00:22:00.910 ⇒ 00:22:03.209 Daniel Saltus: you know, probably the most common.
148 00:22:03.390 ⇒ 00:22:03.970 Robert Tseng: Yes.
149 00:22:03.970 ⇒ 00:22:21.370 Daniel Saltus: Stanford Ems bringing somebody to Stanford Hospital, and so we have the information. We know the email for Stanford Ems. We know the fax number for records and for billing for Stanford Hospital, like we could provide all that as kind of a baseline if you get a random one like your Stanford Connecticut Resident, who was in Iowa.
150 00:22:21.814 ⇒ 00:22:27.400 Daniel Saltus: Or whatever you know what I mean. No, we do not have, you know, like Des Moines Memorial hospital.
151 00:22:27.510 ⇒ 00:22:41.790 Daniel Saltus: you know, we no have no experience with them. So then it would just be, we would have to Google get on the phone. You know what’s who’s your medical records? How do I get them? That sort of a thing? So some are are baked in and and common, and some are just totally new.
152 00:22:42.770 ⇒ 00:22:51.411 Robert Tseng: Yeah. So for the ones that you have to go and like, actually basically set up a new pipeline is kind of the way I see it? Or
153 00:22:51.950 ⇒ 00:22:56.170 Robert Tseng: yeah, you have someone that just has to go and make those calls right now.
154 00:22:56.310 ⇒ 00:22:57.170 Daniel Saltus: Exactly.
155 00:22:57.450 ⇒ 00:22:58.330 Robert Tseng: Okay?
156 00:22:59.380 ⇒ 00:23:28.100 Robert Tseng: yeah. I mean, definitely kind of the the former of, just like within your existing network. Like, if you have all the information already mainly like stored somewhere. And we know that like, Hey, if it’s coming through Stanford Hospital like this is like contact like that that becomes more of like a messaging orchestration problem. And that’s that’s easy to solve. But yeah, like, how do we go after like the net? New like providers as well. I think that’s an interesting problem to solve. Yeah.
157 00:23:29.160 ⇒ 00:23:46.799 Daniel Saltus: I mean largely on their websites, you know the the hospitals will have. Oh, you know, call this number, or they kind of like. Maybe you’ll have like a phone tree. That isn’t really easy to find on the public website, or whatever get that information. I just don’t know how how well the AI could could poke around and get that information.
158 00:23:47.240 ⇒ 00:24:04.189 Robert Tseng: Yeah, I mean anything scraping off of like sites and stuff like it can do. Well, even making the calls, it could do yeah. So, but I guess depends on like, is it worth setting all that up into? If it’s low volume, you know, it would just say, like, what’s the split percentage wise between like?
159 00:24:04.410 ⇒ 00:24:08.639 Robert Tseng: Is it like 20% of like new of like cases that you get are.
160 00:24:09.510 ⇒ 00:24:09.910 Daniel Saltus: Yeah.
161 00:24:09.910 ⇒ 00:24:10.790 Robert Tseng: From that? Yeah.
162 00:24:10.790 ⇒ 00:24:15.640 Daniel Saltus: Evolve like a new provider, or something no probably less than than 20.
163 00:24:15.640 ⇒ 00:24:16.210 Robert Tseng: And that.
164 00:24:18.370 ⇒ 00:24:28.369 Robert Tseng: yeah. So if anything like, if we could even just like get the list and put it in front of whoever’s making the phone calls for you. They yeah, that could be like a
165 00:24:28.520 ⇒ 00:24:32.100 Robert Tseng: My friend, short term solution or something.
166 00:24:32.460 ⇒ 00:24:44.329 Daniel Saltus: Yeah. Or even if, like somebody, a human has to make the call, they make the call. They find the number. They punch the number into the our spreadsheet, or whatever of you know hospital and and their information, you know. Then then they, I would have it.
167 00:24:45.230 ⇒ 00:24:47.030 Daniel Saltus: you know, could in theory could run with it right?
168 00:24:47.210 ⇒ 00:24:48.680 Robert Tseng: Do do you have like a S
169 00:24:48.930 ⇒ 00:24:57.859 Robert Tseng: like a Crm, or like, I guess, a patient record, right like some sort of erp or subsystem that you’re storing all this like customer.
170 00:24:58.638 ⇒ 00:25:00.230 Robert Tseng: Kind of like data in.
171 00:25:00.230 ⇒ 00:25:18.619 Daniel Saltus: Yeah. So right now, what we have is is and I can. I’m not a big computer term guy. So it’s it’s a Nas and Network attached storage. That is cloud backed up. And then we have we’re in the middle of engaging a company that does a case management software. And they actually house documents internally.
172 00:25:19.124 ⇒ 00:25:24.089 Daniel Saltus: So once we have them online, everything should just be like one stop shopping.
173 00:25:24.090 ⇒ 00:25:24.570 Robert Tseng: Do that.
174 00:25:24.570 ⇒ 00:25:33.049 Daniel Saltus: Documents. Cloud, or you know, the the case manager, info client info. All that should in theory be in in one place.
175 00:25:34.030 ⇒ 00:25:46.480 Robert Tseng: Got it. Yeah, I mean, once having that having, that would probably be the easiest you can. Just we could just get that data via Api, and it’s very easy to work with. But yeah, okay, noted that. That’s the the
176 00:25:46.740 ⇒ 00:25:48.590 Robert Tseng: or how the data comes in and out.
177 00:25:50.630 ⇒ 00:26:04.590 Robert Tseng: yeah, I guess. Do you have any other questions for us? For now I think we kind of have some next steps here would love to kind of move this along. I don’t know. Kind of like timeline wise like how you’re. I know we’ve we’ve kind of. We had some delays here and there, but we’re we’re ready to kind of
178 00:26:04.760 ⇒ 00:26:06.930 Robert Tseng: keep. Keep, keep, keep, keep this. Keep this moving so
179 00:26:07.840 ⇒ 00:26:09.689 Robert Tseng: just kind of where where you’re what you’re thinking.
180 00:26:09.690 ⇒ 00:26:33.880 Daniel Saltus: Yeah, it’s so. That’s hard for me to answer like some things, you know. If you if you happen to rise to the top. I’ll get you an answer in a day or 2, but as you’ve experienced sometimes that you gotta prompt me a couple of times, or you can get me to answer you. So it’s it’s hard for me, you know. I’m not trying to avoid your questions. It’s an important one to keep things moving, but I just don’t. I don’t know how to answer specifically. You know.
181 00:26:33.880 ⇒ 00:26:52.010 Robert Tseng: Okay, yeah. No worries. I mean, we’ll we’ll stay on top of it. Just and keep that open communication line. Yeah. I mean, we’re we think that we could do. We could do a good job with with with you and your firm. So definitely, this is something that I’m interested in continuing to to pursue.
182 00:26:52.310 ⇒ 00:27:04.469 Daniel Saltus: Alright perfect. So let me I’ll start working on my end on on getting a summary and and talking with Hillary, and and I’ll wait to hear back from you guys with with some of the the term sheets or whatever. So we can, we can talk about that part of it. Okay.
183 00:27:04.840 ⇒ 00:27:06.160 Robert Tseng: Okay. Sounds good.
184 00:27:06.600 ⇒ 00:27:06.920 Daniel Saltus: Alright!
185 00:27:06.920 ⇒ 00:27:08.229 Robert Tseng: Yeah. Appreciate your time, Daniel.
186 00:27:08.410 ⇒ 00:27:11.379 Daniel Saltus: Of course. Thanks so much for your time, Robert. Thank you.
187 00:27:11.380 ⇒ 00:27:12.050 Robert Tseng: Alright!
188 00:27:12.050 ⇒ 00:27:13.579 Daniel Saltus: The presentation and everything.
189 00:27:14.480 ⇒ 00:27:15.299 Uttam Kumaran: Appreciate it.
190 00:27:16.060 ⇒ 00:27:17.059 Daniel Saltus: Are you topped.
191 00:27:17.330 ⇒ 00:27:17.850 Robert Tseng: Talk, soon.
192 00:27:17.850 ⇒ 00:27:19.269 Daniel Saltus: Everyone guys. We’ll talk to you soon.
193 00:27:19.270 ⇒ 00:27:19.870 Uttam Kumaran: Bye.