Part 3 · Operate
Writing rules
The patterns every writer avoids, and the rules of your own that outrank them.
01What it is
One list, read by every agent that writes for you: posts, captions, cold emails, ads, video scripts, follow-ups on deals and call briefs, the plan, the weekly report, the few words in an answer from Ask your team, and every variant of a winning post. It holds the built-in patterns that make writing read as machine-made, and the rules you add under Settings, Writing rules: phrases never to use, and lines every writer follows.
02Why it exists
A buyer decides in the first two lines whether a person wrote to them. Machine writing has habits a reader now spots at once, and each one costs a little trust whatever the rest of the message gets right. So the habits are named, one by one, and every writer is told about them before it writes rather than after.
Your rules sit above the list because you know things no list can. A word your market finds grating, a spelling your customers expect, a topic you never raise in public: say it once and every writer follows it from the next draft.
The list is also checked after the writing. A draft dense with tells goes back to be written again before you see it, and a draft that uses a phrase you banned is written again once. The checker on the same screen runs the same list, so what you see there is what your drafts are held to.
Two more checks apply to every draft a buyer reads. Money is written in the currency of the markets you sell in, US dollars unless they all share another, and no draft invents a price. And nothing an assistant says to the person prompting it survives: "it is important to note", or asking the reader to correct you, sends the draft back, and anything left after the second attempt is cut. What is written to you about your own account, such as the weekly brief on Home, uses the same words as every screen, never the names the code keeps for itself: channels, ideas being tested, the companies you watch.
03How to use it
- 01Open Settings, Writing rules.
- 02Add a phrase you never want to see, or a line every writer should follow. It applies from the next draft.
- 03Open a category of the built-in list to read each pattern with an example and a better version. Switch one off if it is how you actually write.
- 04Paste a draft into Check a draft to see each tell highlighted where it sits, worst first.
- 05Writing agents link here from their cards on the team screen.
04What good looks like
- You give it
- The generated post the checker offers as a sample: an opener about the times we live in, a not-this-but-that line, three emoji bullets, a staged reveal, a one-word question to farm replies and four hashtags.
- You get
- Banded as machine-made, with every tell highlighted in place and listed worst first: the emoji bullets, the hashtag wall and the inflated verbs at the top.
- Why
- Density is weighted tells per hundred words. Past SLOP_GATE_DENSITY the content gate sends a draft like this back instead of showing it.
Derived from the rules in the code
05What weak looks like, and what it costs
- You give it
- You switch off "Journey and embark" because your product maps a customer journey.
- You get
- Writers stop being told to avoid it and the checker stops counting it. The content gate still counts it toward a draft’s density.
- Why
- The gate is shared by every writer and does not read each account’s switches, so a switched-off pattern can still add to the total. On its own it never sends a draft back.
Derived from the rules in the code