Rhetorical Analysis
Dissects every model response for persuasive strategy — breaking down Ethos (credibility), Pathos (emotion), and Logos (logic) with exact quotes, an effectiveness score, and improvement suggestions.
How it works
Rhetorical Analysis applies the classical framework of rhetoric — developed by Aristotle and still taught in every college composition course — to AI-generated responses. After the primary models deliver their answers, the selected analysis model examines each one as a persuasive artifact, asking not just 'Is this correct?' but 'How is this trying to convince me?'. It evaluates three core dimensions:
- Ethos — Credibility: Does the response cite authoritative sources, use confident expert language, or acknowledge its own limitations to build trust?
- Pathos — Emotion: Does the response use emotionally resonant language, vivid examples, or narrative framing to create engagement or urgency?
- Logos — Logic: Does the response rely on evidence, structured reasoning, statistics, or logical argument to build its case?
- Rhetorical Situation: Who is the intended audience, what is the apparent purpose, and how well-timed is the response (kairos)?
- Overall Persuasion Score: A balance assessment showing which appeals dominate, where the response is weakest, and specific suggestions for improvement
When to use it
- Evaluating persuasive content — ads, political statements, sales pitches, motivational texts
- Understanding how different AIs are designed to influence you (media literacy)
- Education: AP Language, college composition, debate prep, argument analysis
- Business: reviewing AI-generated marketing copy or pitch decks for persuasive strength
- Legal and journalism: analyzing rhetoric in briefs, public statements, or news coverage
- Improving your own writing — use the weakness flags and improvement suggestions as a writing coach
This mode organizes model output for review. Its conclusions still require appropriate source checking and human judgment.
Sources
- NIST AI Risk Management Framework — National Institute of Standards and Technology. Accessed 2026-07-20.