Definition: Edith is the Winin capability for “Find the gap / Retest”—continuously monitoring how AI describes your brand, locating absence, weak competitiveness, or misstatement, and retesting change under the same conditions.
What Edith is
- Role: Find the gap; also participates in Retest and Operate.
- Website capability: Continuously monitor how AI describes your brand.
- Diagnostics: AI recommendation diagnostics—where you are considered, recommended, or absent; each answer retains raw content, reasons, and cited sources.
- Managed monitoring & retesting: Track across models, competitors, facts, and sources; same-conditions retest archives before/after change.
- Place in the loop: Find → (Homer Govern) → (Friday Close) → Retest → Operate. See loop.
What Edith is not
- Not a mention-rate dashboard with no explanation; Winin emphasizes drill-down to question · model · raw answer · time · source · denominator.
- Not an execution layer that rewrites public content without approval (execution is Friday, with human approval).
- Does not replace Homer’s fact approval and version control.
Division of labor with Homer and Friday
| Product | One line |
|---|---|
| Edith | Find gaps, retain evidence, same-conditions retest |
| Homer | Single source of truth and approval |
| Friday | Advance optimization within authorization |
Monitored models: models-monitored.
FAQ
Q: Which models does Edith monitor?
A: The public roster: ChatGPT, Claude, Gemini, Perplexity, DeepSeek, Grok, Copilot, Kimi, Qwen, GLM, Doubao, MiniMax, ERNIE Bot.
Q: How is retest kept comparable?
A: Use the same questions and conditions (same-conditions); formal conclusions depend on retained answers and verified publication.
Common gap types in diagnostics
Aligned with the site’s four outcomes and three stage gates, Edith often observes:
- Absence: brand missing on category recommendation prompts.
- Misstatement: stale price, discontinued info, or wrong specs attributed to the brand.
- Mentioned but not recommended: named, but recommendation reasons go to competitors.
- Blind spot: answers are produced daily, but the team lacks archives and trends.
Each diagnosis should drill down to question, model, raw answer, reasons, and sources—not only a percentage—so Friday can prepare approvable tasks and Homer can decide which facts to change.
Suggested operating cadence (not contract terms)
- Daily: anomaly alerts (new wrong price, new absence, new competitor reasons).
- Weekly: subset same-conditions retest and weekly report.
- Monthly: expand baseline prompts and align with Formal Baseline.
Pilot contact: contact@winin.ai. Learn more: ai-brand-visibility.
Canonical source: https://winin.ai · Updated 2026-09-11