Intro to the "unslop" framework
This resource is based on A copy-paste prompt that edits the slop out of any document, featuring Tim Metz of Animalz, published on the AI Lab by ActiveCampaign.
Get the one-pager
This one-pager distill’s Tim Metz’s anti-slop framework into something you can scan in 30 seconds, pin to your desk, or pull up to reference when reviewing AI-generated content. This framework captures how Metz’s system works at a high-level glance.
The big idea
“Workslop” is AI-generated content that seems valuable on first read but contains no real substance — needlessly verbose, redundant, or vague. Metz’s “unslop” command is a custom Claude prompt built around three editing principles that systematically cut slop without removing what matters.
The framework
| Principle | Rule | Watch for |
|---|---|---|
| MECE (Mutually Exclusive, Collectively Exhaustive) | Each section covers a distinct topic. Together, all sections cover the full subject. No gaps, no overlaps. | Sections that repeat each other’s points from slightly different angles |
| DRY (Don’t Repeat Yourself) | If you’ve said it once, don’t say it again. Eliminate anything that rephrases earlier content. | Elegant-sounding sentences that add no new information |
| Essential (As simple as possible, but not simpler) | Cut filler and complexity, but don’t remove necessary context or examples. | Output that “reads like shorthand” or cuts examples the reader needs |
Where slop hides
- External content: Blog posts, social copy, customer-facing docs
- Internal docs: Emails, meeting notes, strategy memos, briefs (“workslop”)
- System inputs: Brand kits, system prompts, reference docs that feed AI workflows (slop here compounds into every output)
When to use this
- Reviewing any AI-generated document before sharing or publishing
- Cleaning up brand kits or system prompts that feed into AI workflows
- Editing internal docs (emails, briefs, memos) for clarity
- Calibrating how much AI involvement fits each content type
Common mistakes
- Cutting too aggressively: If the text reads like shorthand, you’ve gone too far. Restore context.
- Removing examples: AI often flags examples as unnecessary. They’re usually important—keep them.
- Skipping the feedback loop: Don’t just fix the output. Tell the model what you changed and why.
- Unslopping outputs without unslopping inputs: Clean your brand kits and system prompts first. Sloppy inputs produce sloppy outputs.
Quick-start
Start with one internal document you recently wrote or received that feels bloated. Run it through the unslop command (see the prompt template asset). Compare before and after. If the rewrite is better, apply the command to the reference docs that feed your AI workflows—that’s where the compounding value kicks in.
Ready for the full story?
Read A copy-paste prompt that edits the slop out of any document, featuring Tim Metz of Animalz, published on the AI Lab by ActiveCampaign.
Related
More data from the AI Lab.