I fired my one AI copywriter and hired a writers' room
Seven roles, ten stages, a show bible and a blind audience panel. How I write my talking-head videos with Claude, and the skill you can steal.
Where it came from: The producer behind my talking-head videos: seven roles, ten stages, and a weekly AI launches show.
The first time I asked an AI to write my short videos, I got exactly what you'd expect: smooth, confident and forgettable. Every hook was "Here's the thing nobody tells you about AI." I could have filmed fifty and no one would have stopped scrolling.
So I stopped asking one model for one draft and built a room.
The cast
Every role is the same model with a different job and the same show bible:
- Scout pitches twice as many angles as I need, and every angle has to react to something real: a trending post, a launch, a commit I made that day. No link, no pitch.
- Provocateur scores for tension. Would anyone argue?
- Strategist scores for fit. Does this help the business owners my products serve?
- Skeptic scores for truth. It's the only role allowed to say "must."
- Showrunner picks the episodes and writes the brief.
- Copywriter drafts, and then has to answer every note.
- The panel is four people who don't exist: an HVAC owner scrolling Reels at 9pm, a credit-union marketing VP, a sales VP who wants something she can forward to her team, and a lender's ops lead who's been burned by vendors. They score hooks blind, shuffled, with no idea which one the room liked.
Ten stages, every one saved to disk
Scout → pitch meeting → pick → draft → critique → revise → red-line check → blind panel → final cut → "own the take."
Two rules do most of the work. First, the copywriter can't ignore a note. Every one gets "accepted" (and changed) or "rebutted" with a reason. Second, only the Skeptic can block. Everyone else's "must" gets downgraded to "should." That keeps the room from sanding every idea down to nothing while still guaranteeing nothing false gets through.
On the first real run, five of six drafts failed the first red-line check. The copywriter accepted 134 notes and rebutted 4. That's the point. Most of the value happens in the round you'd skip if you were writing alone.
The bible
The bible is one file that every role reads. It covers who I am, how I talk, the panel personas, the pillars, the spice dial (1 is useful, 5 picks a fight) and the red lines: no facts beyond the sources, a post on X is "someone claimed," nothing that identifies a customer, no legal conclusions, no punching down.
When I give feedback, it goes into the bible, so the next run inherits it. "These aren't from what's trending." "This is for LinkedIn, not TikTok." "The teleprompter just says the sentence. Give me the whole script." "The music is too loud." All of that is in there now.
The last stage is the one I read
"Own the take" rewrites the final cut into 150 to 220 words I read aloud on a voice-following teleprompter. No hedges, no placeholders, no @handles. It has to sound like me, because I'm the one on camera.
There's also a weekly variant. The scout finds AI launches from the last seven days with working links, and the room builds a roundup with a use / watch / skip verdict for each one.
Steal it
The skill is writers-room in my John Lab plugin. Write your bible first. It's the part that makes the room yours.
/plugin marketplace add johnarndt/john-lab
/plugin install john-lab@john-lab/plugin marketplace add johnarndt/john-lab /plugin install john-lab@john-labDownload writers-room.zipView on GitHub
View SKILL.md
---
name: writers-room
description: "Use when a creator or founder wants short talking-head video episodes written by a room of critics instead of one draft. Triggers include 'pitch me video ideas for this week', 'write my LinkedIn videos', 'react to what's trending', 'give me a teleprompter script', 'make episodes from these sources', 'weekly AI launches roundup', or 'run the writers room'. Runs scout pitches, a scored pitch meeting, a showrunner pick, draft, critique, accept-or-rebut revision, a red-line check against a show bible, a blind audience panel, and a final spoken script."
---
# Writers' room
One model writing one script produces the average take. A room of roles that disagree, scored by a panel that has never seen the script, produces the take worth filming. Every stage writes a file so the room can be resumed, audited and reused.
## Before the first run: the bible
Write `bible.md` with the user and keep it in the show folder. It is every role's system prompt.
- **Host:** who they are, what they have actually done, how they talk (sentence length, words they never use).
- **Audience:** 3-5 named personas with role, age, platform, moment of viewing, and what makes them stop or scroll. Example: "Dana, 46, owns a 14-person HVAC company, scrolls Reels at 9pm, distrusts anyone selling AI."
- **Pillars:** for example build-in-public, hot take, story, how-to, news reaction.
- **Red lines:** no facts beyond the sources (a post on X is "someone claimed", not a fact); nothing that identifies a customer or prospect; no legal conclusions; no punching down; any product names or numbers that must never appear.
- **Banned phrases** and a **spice dial** from 1 (useful) to 5 (picks a fight).
## Roles
| Role | Job |
|---|---|
| Scout | Pitches twice as many angles as needed, each tied to one real source (a trending post, a launch, a commit) with its link. |
| Provocateur | Scores for tension and shareability. |
| Strategist | Scores for fit with the audience and the host's business. |
| Skeptic | Scores for truth and red lines. The only role allowed to say "must". |
| Showrunner | Picks the episodes and writes each brief. |
| Copywriter | Drafts and revises. |
| Panel | The bible's personas, scoring hooks blind. |
## Stages
Write each stage to `shows/.work/<date>/NN_<stage>.json`.
1. **Scout.** 2N angles: `{id, pillar, raw_take, source_url, source_text, facts[], audience, so_what}`. Drop any angle without a source link.
2. **Pitch meeting.** Provocateur, Strategist and Skeptic score every angle 1-10 with a blunt `note` and a `direction` or `kill`.
3. **Pick.** Showrunner picks N. Disagreement is signal: a high Provocateur score with a low Skeptic score can still be the best episode if the brief fixes the truth problem. Brief: `{thesis, direction, spice_target, avoid}`. Every kill gets a one-line reason.
4. **Draft.** Six hooks of 14 words or fewer, one of each type (confession, contrarian, number, stakes, you-are-doing-X, live test); 3-5 beats `{cue, nudge}`; three closes; a 110-200 word script; the facts used.
5. **Critique.** Each critic quotes the exact words they would change, with severity `should` or `could`. The Skeptic may use `must`, only for red lines.
6. **Revise.** The copywriter answers every note `accepted` (and changes it) or `rebutted` with a reason. Skeptic `must` notes cannot be rebutted.
7. **Red-line check.** Skeptic passes or lists must-fix items. On fail, one fix round that changes only those items.
8. **Blind panel.** Shuffle the hooks without labels. Each persona scores `stop` and `comment` 1-10 with a 12-word reaction.
9. **Final cut.** Lead hook, two A/B alternates, script, risk notes, production notes.
10. **Own the take.** The exact words the host reads aloud, 150-220 words, in their voice: no hedges, no placeholders, no handles. This is what goes on the teleprompter.
## Variants
- **Trend reaction:** every angle reacts to one specific trending post and carries its link.
- **Weekly launches:** the scout finds 10-12 launches from the last 7 days with working links; the room picks 4-5 for a roundup (250-360 words, each with a use / watch / skip verdict) and up to two for their own episode.
## Rules
- Keep a `history.txt` of produced topics and reject repeats.
- Facts come only from the sources. If a claim matters and cannot be checked, the script says who claimed it.
- The host's real feedback overrides the room. Record it in the bible so the next run inherits it.
- Audio, when the room hands off to editing: voice around -16 LUFS, music about 20 LU under the voice. Hosts consistently notice music and effects that are too loud.
- Produce scripts. Do not post anything.