Round One = Building the Foundation
Of the three rounds, round one is the round that surfaces the most feedback.
This is not because the quality of round one outputs is low. It is because we are presenting an early-stage draft to both the external auditor's perspective and the internal seven subagents at once, which makes blind spots easier to spot.
The Role of Round One
We position round one as "the round that draws out all the rough edges."
When a new design, a new chapter, or a new policy goes through consultation (here meaning: the act of gathering confirmations from multiple reviewers) for the first time, the person who built it always carries assumptions they have not made explicit. Something written as "this is obviously how it works" can look like an under-explained gap from a different area of expertise. The gap between what lives in the designer's head and what is actually written in the output is real, but it is hard for the designer to see from the inside.
Round one consultation is what makes that gap visible. No matter how many times a designer re-reads their own output, the knowledge they brought in as background will be skipped over as "obvious." Without a fresh outside view, that gap never surfaces.
In round one, the external auditor identifies primary areas of concern -- establishing where to focus the review -- and the internal seven subagents each run parallel checks from their own areas of expertise. Running the work through this flow in round one brings to the surface, one after another, the parts the designer had not noticed.
Why Round One Surfaces the Most Feedback
Here is a more concrete explanation of why so many points come up.
In round one, everyone looks at the draft or design for the first time. Round one is the only round in which reviewers see the original, before any revisions. On a first read, the review runs without the shared assumptions the designer took for granted. Basic observations tend to come up: "This term is not defined clearly," "The connection to the previous chapter is hard to follow," "The reasoning here is thin."
As a concrete example: even if a draft states in one line that "Kill Switch (an emergency stop mechanism) is assumed to be configured," everyone other than the designer has no shared understanding of "under what conditions does it trigger" or "who operates it." Because everything connects in the designer's head, the missing explanation is not noticed. Round one is the round that makes these omissions visible.
This is a good sign. "Many points raised" does not mean "many problems." Reading it as "many points raised means many blind spots made visible" is more accurate.
Building it as a "Foundation"
Only after the feedback from round one has been addressed does what we call "a foundation capable of withstanding consultation" actually exist.
Trying to move forward without that foundation tends to require fundamental revisions later. Instead of adjusting the branches of a design, we end up rebuilding the roots. Running round one carefully reduces the cost of revisions in subsequent rounds.
There is one more important meaning to round one. A record is created of how the internal seven subagents responded to the primary areas the external auditor identified. Information accumulates: "this area was picked up thoroughly," "this point was already shared internally," "this item went unnoticed."
That record becomes the starting point for the checks in round two and round three.
After Round One Ends
When the external auditor's secondary review for round one is complete, the result comes back as one of three outcomes: "send back for revision" (an instruction to fix and resubmit), "conditional go" (acceptable once specific fixes are made), or "overall restructuring needed."
Finishing with all approvals is rare. In fact, fewer approvals in round one is a sign that the consultation process is functioning correctly.
We revise what was sent back and prepare for the next round. What we revise and how is worth recording, because it directly affects the efficiency of round two. If it is clear what was raised in the previous round, the internal seven subagents in round two can run their checks while comparing against the prior version. Without that, the same scope gets reviewed from scratch. Without a record, round two becomes a near-repeat of round one.
Deciding up front that round one is "the round that draws out all the rough edges" means that when a lot of feedback comes in, we receive it without alarm. More is better, at this stage.