The Exchanges

Every argument clarity score on this site is built from rows on this page. Each question and answer was assessed with names hidden, the host's own answers included, on four things from 1 to 5: directness (does it answer the question asked), coherence (do the ideas follow), precision (concrete details and clear references), compression (says a lot per word). The weighted mix (30/30/25/15) is the exchange score. A person's published score averages their exchange scores on raw tape only, at least 8 of them, shrunk toward the cohort mean. Full method →

Shreyas Parab no published score: only 2 usable exchanges on raw tape, and a fair score needs 8+ record → ← everyone

Every exchange below was scored with names hidden, four dimensions each from 1 to 5. An exchange's score is 0.30·directness + 0.30·coherence + 0.25·precision + 0.15·compression. The published score averages the raw tape exchange scores and shrinks small samples toward the cohort mean, so five great answers can't beat twenty good ones. Produced feed rows count only toward coarse estimates, never toward a full score.

clear all ✕
2exchanges match
2on raw tape
0redirected or not addressed
Answered raw tape D 5 · C 5 · P 5 · Cm 4 4.85

Q around here, so it's okay. So in general, obviously, there's more things the machine can do than it could before, but what you're saying is it still can't do everything. Is it getting to a point where maybe you can do, like, almost everything, or is there still, like, you have a ton of these billers. In 10 years, will it be doing everything? How do you think about it?

A I think in a couple years. Um, what I would say is that dental billing, like, the difficult part is it's very unintuitive. So it's, it's also not, like, written down. It's not really in, like, the training data, right? Um, and that, like, Unintuitiveness is why we need the billers to, like, sort of correct the system. Also, like, for the billers to teach the engineers, like, why is this the way that it is? Um, and I can maybe give an example of, like, one edge case. So it's like, let's say, um, you got an EOB back from the insurance company, and EOB is basically a receipt on how they adjudicated and or paid the claim, right? This usually comes in the mail, um, along with a check. And it could be that, like, you got this EOB back for this claim, and you're like, I never even submitted this claim. And it turns out that they sent it back Um, under the parent's name instead of the child's name because they're on the same insurance plan, right? And so, like, you'd have to go into the EHR and go figure out, like, oh, actually, I'd never submitted this claim. Let me check the family file. Let me, like, go look at this. And then it turns out you submitted four x-ray codes, and they returned back, oh, we paid you for one code. And you're like, why is this the case? And it turns out, oh, they combined the different codes together, um, in this, like, very unintuitive way, in a way that, l…

AI assessment note: “I think in a couple years.”

Answered raw tape D 5 · C 5 · P 4 · Cm 4 4.60

Q So take us through what you guys are doing. Like, obviously, AI is gonna change this and make it a lot easier. AI still makes mistakes, too, though. It's not perfect. So, so, so, so what's possible here? What are you guys building?

A Yeah, I think, um, At its core, we're, like, a tech-enabled service, so what we actually have is a team of roughly 30 billers, um, and we're really prioritizing human in the loop. Um, and so our AI does a lot of the, you know, like, boring manual processes, um, and does, like, to the extent that it can, but obviously sometimes there's, like, some higher-order decision-making need to happen, and to do that, we rely really on our team of, like, expert in-house billers, like, each one of these, um, billers are US-based, they have eight years of experience in a dental office doing dental billing, And they really help us, like, train, not just, like, our own AI systems on how to do this process, but also, like, the engineers work with them very, very closely and, like, really, really understand the different workflows that are required to do this entire process.

AI assessment note: “At its core, we're, like, a tech-enabled service, so what we actually have is”

page 1
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 150 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.