Round Two and Round Three: Bulletproofing
After round one surfaces and resolves the rough edges, round two begins.
The key shift from round one is this: the focus moves away from "finding problems on first sight" and toward "checking whether the fixes introduced any new problems."
What Happens in Round Two
In a round-two consultation (here meaning: a structured review session where each subagent examines the work from its assigned perspective), the internal seven subagents review the output after fixes have been applied.
They check whether the corrected sections are actually fixed. At the same time, they check whether a fix in one place broke the consistency of another. Fixing a paragraph can make the connection to the previous chapter feel off. Changing a phrase can disrupt consistent use of a term. These side effects of a fix (problems that arise as a result of making a correction) are easy to miss unless you look at the work through post-fix eyes.
The external auditor's second-pass review also tracks "whether the points flagged in round one were properly addressed." Because the round-one observations and the round-two verification are linked, nothing falls through the cracks.
The number of issues raised in round two is smaller than in round one. The big structural problems come out in round one. What surfaces in round two are small inconsistencies introduced by the fixes, and the final checks that were deferred from round one.
What Happens in Round Three
Round three is the final check after the round-two fixes are applied.
At this stage the number of issues becomes smaller still. The major gaps have been closed over two rounds, so what remains is minor variation in phrasing, small inconsistencies noticeable from a reader's perspective, or a check of overall consistency when several chapters are viewed together.
The external auditor's focus also shifts, from "confirming the direction" as in round one, to "verifying the fine details and giving final sign-off." When the external auditor gives a GO in round three, that is the point this series treats as "the decision being settled."
The Sense of Bulletproofing
As round two and round three accumulate, there is a sense that judgments and designs become bulletproofed (here meaning: hardened to the point where they no longer collapse when challenged from a new angle).
A judgment made in only one round is fragile. When a different perspective is applied later, it tends to break down. The gaps remain: "we had not looked at that angle" or "we had not verified that assumption."
A judgment that has gone through round two and round three has been examined from different angles across multiple rounds. Points that passed round one but were flagged in round two have been fixed, and the side effects of those fixes have also been checked. When a new perspective is applied later, there is less room for the judgment to collapse.
Bulletproofing is not about the designer's judgment becoming sharper. It is a structural process: the same judgment is reviewed repeatedly by multiple independent sets of eyes, and the room for overlooked gaps is progressively reduced.
The Basis for "Three Rounds Is Enough"
Here is what the actual operation of this series can say to the question of whether a maximum of three rounds is sufficient.
After round three is complete, it is rare for significant gaps to remain. Round one draws out the rough edges; round two verifies the consistency of the fixes; round three locks down the fine details. Following this sequence makes it unlikely that a situation arises later requiring a fix from the ground up.
That said, if a fundamental design review enters during round one, that is a different matter. If the assumptions underlying the design change, you restart from round one with the revised design. This does not mean "re-counting within the same three rounds." It means running three fresh rounds with the revised design.
What Remains After Three Rounds
When three rounds of consultation are complete, a log remains.
What was flagged in round one. What problems appeared as side effects of fixes in round two. What fine-detail issues were raised in round three. What perspectives the external auditor brought, and how the internal seven subagents responded.
This log serves as a reference when building the next design or the next chapter. It accumulates a track record: "last time, this kind of angle came up in round one" or "content in this category tends to produce variation in phrasing."
Running three rounds of consultation does not only settle the current decision. It also produces data that reduces blind spots in future rounds.
Bulletproofing is not something completed in a single consultation. It grows stronger gradually, through repetition.