Module 1.0 · Lesson 06 of 07
LSN06

Prompts for content, email, and social

Channel-specific prompt patterns for long-form content, lifecycle email, and organic social.

Runtime
14 MIN
Format
Read · Self-paced
Access
Email required
Track
7 LESSONS
The premise

Three working prompt patterns cover most of a marketing team's daily output: long-form content, email, and social. Each is the Briefing Method tuned to what that format punishes, because a blog post, an email, and a LinkedIn post fail in three completely different ways.

Why each function needs its own pattern

Everything so far applies to any prompt. But formats have personalities. They reward different moves and punish different sins, and a brief that ignores the format's failure mode produces technically correct output that still does not work.

How long-form content, email, and social each fail, and the key move that fixes each.
FunctionHow it failsThe key move
Long-form contentEmpty calories: structure without substanceOutline first, then write section by section with your data injected
EmailDivided attention: three messages, zero actionOne job per email, one call to action, subject lines via examples
SocialCorporate costume: reads like a press releaseNative format rules in the task layer, hooks from your real winners

The three patterns below are starting briefs for each. Combined with your Voice Spec Block, they cover most of what a marketing team produces in a week.

Does “act as an expert copywriter” actually work?

The most popular prompt advice on the internet deserves an honest answer before we build the patterns: mostly, no. On modern models, telling the AI to act as a world-class copywriter adds a costume, not competence. The model does not gain skill from the title, and the measurable quality difference on current systems is small to none.

What the role line sometimes does is smuggle in context: "act as a B2B email specialist" hints at an audience and a format. But that is the audience layer's job, done badly. One orienting sentence does no harm. A role line used as a substitute for the four layers does real harm, because teams write the costume, skip the brief, and then blame the model. If a role seems to help, it is telling you which layer your brief was missing. Write the layer.

Free · No spam

Unlock the rest of this track

Free. Join the newsletter and all 7 lessons open instantly. Unsubscribe anytime.

FAQ · Operator questions

Frequently asked questions

01How do I prompt AI to write blog posts?
Use a two-step pattern: first brief an outline with question-based H2 sections and one takeaway per section, then write one section per message, injecting a real data point or example into each. Substance injection is what separates useful content from filler.
02How do I prompt AI to write marketing emails?
Brief one email with one job: under 150 words, exactly one call to action, audience and context stated, your voice rules pasted in. Generate subject lines in the same prompt by giving two or three of your best past subject lines as examples.
03Does role prompting like "act as an expert copywriter" work?
Mostly no on modern models. The title adds no skill, and any improvement usually comes from context the role implies, which the audience layer of a proper brief does better. One orienting sentence is harmless; a role used instead of a brief produces generic output.