What Is the Golden Rule 7+1?
When you face an important decision, how do you usually handle it?
Some people think it through alone. Others consult a few trusted colleagues. Some bring it to the relevant team. There is no single right way. But when it comes to running a system powered by AI agents (AI built to carry out tasks on your behalf), each of these approaches leaves something to be desired.
The reason is that AI has a tendency to produce confident-sounding answers even when those answers are wrong. Without a human standing by to say "something's off here," an AI can keep moving in the wrong direction indefinitely.
In this series, we run a consultative framework (a structured process of gathering input before a major decision is finalized) called the Golden Rule 7+1 to address exactly that risk. This chapter gives you the full picture.
Q. Why do we need such an elaborate system in the first place?
A. Because both AI and humans share the same bias: they tend to produce answers that favor their own position.
When a major decision rests on a single point of view, blind spots appear. This happens with humans too. But with AI agents, each judgment feeds automatically into the next, which means errors tend to spread further and faster.
That is why we built a mechanism to collect input from multiple angles simultaneously, before any decision is finalized.
Q. What does "7+1" mean?
A. It refers to seven internal roles and one external auditor.
The "7" stands for seven internal roles, each staffed by an AI: Tech Lead (responsible for technical oversight), COO (responsible for overall operations), QA (responsible for quality assurance), Brand Voice (responsible for maintaining consistent tone and language), Task Dispatcher (responsible for routing tasks), Researcher (responsible for information gathering), and Content Director (responsible for overall content strategy). All seven contribute their perspectives in parallel, at the same time, rather than one after another.
The "+1" refers to a separate external audit AI that operates outside the group of seven. In this series, we call it Antigravity. The key point is that it runs on AI from a different vendor than the internal seven. By bringing in an independent eye that is not part of the internal team, we can surface problems that insiders might overlook.
Q. Once we collect input from all 7+1, is the process done?
A. Not quite. We run up to three full rounds.
There is a reason we do not stop at one round. The first round is the foundation-building pass. Everyone is contributing for the first time, which means this is when the most blind spots (overlooked issues) tend to surface.
The second and third rounds are the consolidation passes. We take the issues raised in round one, revise accordingly, and run again. Each iteration makes the organization's decision-making more resistant to failure. Rather than building an impenetrable wall in one shot, we find the weak points and reinforce them one by one. The result is a structure that becomes genuinely solid through repetition.
Q. What is the order of steps within a single round?
A. Each round follows a three-stage consultation structure (here meaning: a sequence of external perspective, internal deliberation, and external verification).
Each round has three stages.
First, the external auditor Antigravity provides first-pass observations, acting as a kind of map that marks what areas deserve attention. Next, the seven internal roles deliberate in parallel, each contributing their own perspective. Finally, Antigravity performs a second-pass check to ask whether anything important was missed in the internal discussion.
First-pass observations from Antigravity, then parallel internal deliberation, then second-pass verification. That three-stage flow counts as one round.
Summary
Here is the full picture of the Golden Rule 7+1.
- 7 = seven internal AI roles contributing their perspectives in parallel
- +1 = an independent external audit AI from a different vendor providing a separate view
- 3 rounds = up to three full iterations of the above
- Three-stage consultation = each round follows the structure: external first pass, internal parallel deliberation, external second pass
We will cover each element in detail in the chapters ahead. For now, keeping this overall structure in mind will make it much easier to follow what comes next.