← Back to Quorum

Methodology

How Quorum works

What happens in a session, and what compounds over time

Every business process has been systematized. Judgment is the last one. Quorum exists to make judgment — not just intelligence — something that compounds the more you use it.

What happens in a session

1

Your decision is read structurally, first

Before any advisor responds, your decision is read for what kind of decision it actually is — not just the words you used. A financing decision and a co-founder split can share the same underlying structure even though nothing about the topics overlaps. This structural read is what everything after it is built on.

2

Six advisors respond from that read

The Contrarian, the Risk Architect, the Pattern Analyst, the Stakeholder Mirror, the Elder, and the Competitor each respond from their own angle on the same structural read — not six calls to the same generic model with different personas bolted on.

3

The Council synthesizes, it doesn’t vote

Where the six perspectives agree, that gets weight. Where they genuinely conflict, the synthesis says so directly rather than averaging it away — a decision that six advisors are split on is different from one they agree on, and you should be able to tell the difference.

4

Challenging one advisor reaches all six

If you push back on an advisor or add context they missed, that's treated as new information for the whole council, not a private exchange with one advisor — it reaches every advisor automatically. Each one reassesses independently through its own lens and may keep, strengthen, weaken, or reverse its position; being told the same thing never means reaching the same conclusion. The Council synthesizes once, after every advisor has had the chance to weigh in.

Why this isn't a chatbot

A chatbot routes your text straight to a model and returns whatever comes back. Quorum reads the structure of your decision first, and that structural read is what determines which advisors weigh in and how they relate your decision to your own history — a step a chatbot has no equivalent of. The six-advisor format isn't a stylistic choice either: disagreement between advisors is information. A single model answering as itself can't tell you that a decision is genuinely contested.

What Mirror adds over time

Patterns are only surfaced once they’re earned

A single decision can’t tell you much about a recurring pattern. Mirror only surfaces a pattern once there’s enough evidence across your own decisions to support it — and only from things you actually wrote, not from the Council’s own responses reflected back at you.

Structurally similar decisions, cited, not invented

When a current decision structurally resembles one you’ve made before, that gets cited directly — which past decision, not a generic "you tend to do X." You can always open the original.

Confidence compounds, it doesn’t appear instantly

Anything Mirror shows you carries an implicit confidence level based on how much evidence supports it. Early on, that means Mirror says less. That’s deliberate — a false pattern shown with confidence is worse than no pattern at all.

You don’t have to start from zero

On Elite, you can import context you’ve already built elsewhere — a description of yourself, or a ChatGPT/Claude export — and Quorum extracts a handful of distilled insights for you to review before anything is saved. Raw content is never stored.

Data handling

Your decisions are encrypted at the field level, and nothing is ever used as a case study or shared externally without your explicit, opt-in consent. Full detail on what's implemented today is at Security & Trust, and our data practices are covered in the Privacy Policy.

Have a more specific question? The FAQ on the home page covers the most common ones directly.