Rules are how you teach 0ct two things: your team's definitions — like how ARR is calculated, what counts as a discovery meeting, your R/Y/G thresholds — and the AI tools each brief can use, like web search and code execution.
Every team has its own ARR formula, its own definition of a discovery meeting, its own thresholds for what counts as on-pace. Rules capture those decisions in plain English so they stay visible — not buried inside a prompt.
Decide which tools the AI can reach for on a given brief — web search to look something up live, code execution for real math, or custom integrations for your own systems. Each brief only sees the tools you attach.
A definition like the ARR formula, or a tool the brief should be allowed to use, like web search.
Rules are scoped per brief. Bianca's brief gets ARR, R/Y/G, discovery rules, and code execution. Another brief gets a different set.
Update the rule once and every future delivery uses the new version. No prompt rewrites, no drift across briefs.
Examples drawn from a real Daily Sales Brief.
What counts toward ARR, whether expansion is included, and how to handle MRR-to-ARR conversion.
The pace targets that drive the status badge at the top of every brief.
Booked vs completed, reschedule policy, and which calendar invites count toward the number.
How to resolve conflicts when two systems disagree on the same fact.
Capabilities a brief can reach for, only when you allow it.
Let the brief look something up live — competitor news, public earnings, FX rates, or anything not in your stack.
Run real calculations in a sandbox: pipeline coverage, conversion math, weighted forecasts, FX conversions.
Update the ARR formula in one place and every brief that uses it picks up the new version on the next delivery.
Rules are auditable artifacts your team can review — not paragraphs hidden inside a prompt.
Sales briefs see sales rules. Finance briefs see finance rules. No leakage between contexts.