Letterbook Docs
Draft eagerness
Control when AI drafts are visible based on confidence.
Last updated June 25, 2026
What this guide covers#
Draft eagerness controls how readily Letterbook shows AI-generated drafts to agents. It is a confidence threshold behind a simple slider from Never draft to Always draft.
Configure it in Settings > Automations > Draft Eagerness.
How it works#
Letterbook can generate AI drafts with confidence scores. Draft eagerness decides which unreviewed AI drafts are visible in the conversation.
Lower eagerness hides more low-confidence drafts. Higher eagerness shows more drafts, including drafts that may need heavier review.
Discarded drafts are hidden regardless of this setting.
Choosing a setting#
Start conservative when your AI workflow is new.
| Team state | Suggested setting |
|---|---|
| New AI setup, limited knowledge base | Lower eagerness |
| AI answers are useful but need review | Middle eagerness |
| Strong knowledge base and reliable integrations | Higher eagerness |
| Team does not want AI reply drafts shown | Never draft |
The right setting depends on the quality of your sources. Better knowledge base articles, cleaner customer data, and clearer policies make higher eagerness more useful.
Tune with real tickets#
Review drafts that agents accept, edit heavily, or discard.
Increase eagerness when:
- Useful drafts are not appearing often enough.
- Agents trust the AI for common ticket types.
- The knowledge base covers the questions customers ask most often.
Decrease eagerness when:
- Agents frequently discard drafts.
- Drafts guess without enough source material.
- Drafts are accurate only for a narrow set of ticket types.
- The team is still building review habits.