Qiita — A small n8n workflow can make daily reporting repeatable
Sangam Team
Nov 5, 2026
The Qiita walkthrough shows a compact automation: collect the day’s Google Calendar events, format the schedule, ask Gemini to produce a Japanese daily-report draft, and post the result to Discord. The value is not a claim that every report can be automated perfectly. It is the clear separation of steps. The workflow uses a scheduled trigger, a calendar reader, a code step for formatting, an AI agent for drafting, and a final code step before the Discord post. That last step is important. The author removes unwanted model preambles and keeps the output in a predictable report structure instead of sending the raw response directly. The prompt also includes a useful guardrail: do not assert information that is not present in the calendar. This is a small but meaningful boundary. The AI can organize and phrase the source data, but it should not invent work, progress, or decisions. For teams adopting a similar pattern, the workflow should begin in draft mode. Review a sample of generated reports, check how missing descriptions are handled, and decide whether Discord is the right destination for potentially sensitive schedule information. The automation becomes more reliable when source data is structured before it reaches the model. Source attribution: This editorial draft is based on the original source and is an independent summary, not a reproduction. Original source: https://qiita.com/wan-code/items/1ba7e323398808fa74e2 Discussion questions: 1. Which source system could provide reliable inputs for your first automation? 2. What claims should the model be forbidden to add? 3. Would your notification channel be appropriate for the data being summarized?