So what have been the limitations of those components?

Yeah, let's go through them one by one.

The measurement piece has been vexing just because the early measures that we developed we knew were mostly based on claims data, mostly related to documentation that processes happened or didn't happen.

We weren't really getting quite to the health outcomes and health status of people, patients.

And then we had patient experience surveys which were meant to record how patients were experiencing their care.

It was always a hope, actually, I think as someone who participated in the development of measures, that we would get to a more nuanced, better set of measures through enhancements through health data, and we haven't really been able to deliver that.

Many of the measures in use today you would recognize them from twenty-five years ago, still claims-based, still relatively difficult to implement, and don't get at the real nuances of what patients are experiencing in their care or their health outcomes.

So what new approaches could be adopted? What current components could be eliminated under that new paradigm?

Yeah

... um, and I don't want PRDs to contain code blocks, which Claude already naturally starts to do.

Um, I want Claude Code or Opus or whatever I'm using in Cursor to make those decisions.

So what I started doing is, you know, for a client, I'll take, um, the transcript and then my notes, and I put it together and I build a brief.

Um, and the brief really contains like, um, data structures, uh, golden pathways, uh, what the MVP looks like, needs to haves versus nice to haves.

Really get all that right.

Um, and then I can take that over, uh, and I have a few agents, uh, through Agency Agents.

Hmm.