Content Engine
Multi-channel copy. One source, many outputs.
Content Engine produces channel-specific content from a single source. The source is usually a brief from Strategy Studio. The output is one piece per channel, calibrated to channel rules.
Media placeholder. Demo video, ~45 seconds. Path:
/docs/media/content-engine-walkthrough.mp4. Shows: brief loaded, sequential
drafts per channel, drift detector flagging tone divergence between email and
social.
What Content Engine runs
| Workflow | What it produces |
|---|---|
| Pillar to derivatives | One pillar piece, eight channel adaptations. The 1:8 target. |
| Channel set | Email, social, SMS, ads tuned per platform. |
| Bilingual variants | Arabic and English copy from the same brief. |
| Cultural calendar | Channel-specific content tied to regional moments. |
Sequential by default
Content Engine drafts one channel at a time by default. Parallel drafting is available, but sequential is the recommended pattern. Two reasons:
- The operator reviews each output before the next runs. Tone drift is caught at the source.
- The drift detector reads completed channels to calibrate the next one. The third draft reads what the first two said.
The drift detector
The drift detector compares channel outputs and flags divergence:
- Voice drift. The email leans aspirational, the social copy leans transactional.
- Claim drift. The Twitter post promises faster than the LinkedIn post.
- Audience drift. The SMS targets a different segment than the email.
A flagged drift produces a named diff. The operator resolves it by editing one side or accepting the variance.
Channel constraints
Content Engine knows what each channel allows:
- Twitter/X: 280 characters, threading rules.
- Email subject lines: visible character count varies by client. Default 50.
- SMS: 160 characters Latin, 70 characters Arabic UCS-2.
- LinkedIn: longer-form, formal default.
- Instagram caption: 125 characters before truncation.
The constraint engine refuses to draft outside the limit. The operator can override with explicit consent. Brand Guard records every override.
Reading from Brand Brain
Content Engine reads Voice, Channels, Guardrails, Audiences. Every channel draft cites the canon sections it pulled from.
Handoff to Brand Guard
Every output produced runs against Brand Guard rules before the operator sees it. A red verdict surfaces inline with the issue named. A yellow verdict surfaces as a Suggestion. A green verdict shows the rule set passed.
Tone
Formality 35, Warmth 45, Certainty 30, Brevity 30. Slightly warmer than Strategy Studio, still terse.