AI and Data Processing
AI and data processing
Last updated: 26 August 2026. AI is an optional refinement step. It is never required to generate, edit, copy, or download a deterministic Markdown draft.
Processing by mode
Customer / deterministic
No AI provider is used. The app formats selected Jira results in the Forge/Jira execution path. This is the default mode for a customer-facing draft.
Customer / AI-assisted
The user explicitly selects this mode after generation. The app sends OpenAI a structured customer-facing payload for rewriting; issue descriptions are excluded. The payload may contain issue keys, summaries, and grouping fields only. Caller-supplied Markdown is not sent. The returned text is displayed as a draft for review.
Internal / technical
The user explicitly selects the technical mode and confirms the additional data choice. Because this mode is intended for internal notes, it may include sensitive issue descriptions. Confirm your team’s policy before using it with confidential or regulated content.
Limits and failure behavior
Forge quota counters bound AI usage, with site and global monthly limits. Counters are best effort under concurrent requests. If AI is not configured, unavailable, timed out, or quota-limited, the app falls back to deterministic generation where possible.
AI output may be wrong, incomplete, or influenced by text in the source issues. Review it before sharing. This page does not promise zero retention, a particular OpenAI region, or a particular use of data for training; consult the applicable OpenAI documentation and agreement for the service configuration in use.
For the broader handling summary, read Privacy. For questions, email info@lucapalonca.com.