expert-panel

Marketing copy is scored by an auto-assembled panel of 7-10 domain experts. The evaluation applies rubric selection by content type and iterates through revision rounds until reaching a 90+ quality threshold.

AI writing detection works through a 'Humanizer' mechanism that applies a 1.5x weighting penalty to known AI writing patterns and banned phrasing. This ensures marketing content scores accurately reflect human-like quality and conversion readiness.

Yes, you can evaluate landing pages and email sequences for conversion optimization. The scoring system selects appropriate rubrics for various content types including ads, CTAs, strategy documents, and charts to provide line-level fixes.

The iterative revision loop runs up to 3 rounds of scoring and editing. It keeps improving drafts with actionable edits until every expert on the panel reaches a 90+ target score or the maximum iterations are exhausted.

Yes, the evaluation requires YAML-frontmatter activation and the Anthropic dependency. It uses weighted aggregation across rubrics to output transparent scoring and line-level fixes tailored to the specific content type being evaluated.

Rubrics used for content scoring include content, strategic, conversion, visual, and evaluation rubrics. The system selects the appropriate rubric based on the content type, evaluating everything from titles to recruiting and vendor evaluations.