Everyone's raving about Rytr's fancy commands for long-form content, but let's talk about the one that actually matters when the rubber meets the road: `paragraph`. It's not for writing paragraphs. It's for taming the beast of a vague, sprawling prompt into something the AI can actually execute without hallucinating half your requirements.
Most people dump a five-sentence instruction into the main command field and wonder why the output misses key points. The `paragraph` command forces you to sequence your logic. Want a security policy that includes scope, exclusions, enforcement, *and* review cycles? Don't just say that. Break it down.
Example: Instead of "Write a GDPR data retention policy for a SaaS application," you'd structure it as:
- Paragraph 1: Define the scope and governed data types.
- Paragraph 2: List specific retention periods per data category.
- Paragraph 3: Outline secure deletion procedures and documentation.
- Paragraph 4: Assign roles and responsibilities for compliance audits.
This mirrors how you'd actually audit the thing—stepwise, with clear checkpoints. It stops the engine from glazing over and inventing a "seven-year retention period for user IP addresses" because it got confused. You get control, traceability, and an output you can actually trace back to a requirement. The real pitfall is thinking of it as a writing tool; it's a precision instruction set. Use it like one.
—Greg
Trust but verify