Field SOP
Field SOP

Daily & Weekly Report Automation Workflow Template: n8n Scheduled Fetch + LLM Generation + Group Push (Downloadable)

30 minutes a day on daily reports is the lowest-ROI grunt work. This n8n workflow pulls Git/task-system data on schedule, LLM generates a structured report (done/in-progress/tomorrow/risks), auto-pushes to Feishu/DingTalk. With .json template download and companion prompt.

Published July 26, 20264 min read
<!-- workflow-daily-report-automation | resource | Daily & Weekly Report Automation Workflow Template -->

Spending 30 minutes before EOD writing a daily report is the lowest-ROI grunt work. This n8n workflow template pulls your data sources on a schedule, uses an LLM to generate a structured report, and pushes it to Feishu/DingTalk. Import, configure, run.

Workflow Chain

Cron trigger -> fetch multi-source data (Git commits / task system / DB) -> merge -> LLM generates report -> push to Feishu/DingTalk.

Download Template

Setup Steps

  1. Import: n8n -> Workflows -> Import from File, pick daily-report-workflow.json
  2. Cron node: set to 18:00 daily (before EOD)
  3. HTTP Request (data source): fill your source URLs; chain multiple nodes (Git commits / Jira / DB query)
  4. Merge node: use Set / Code nodes to concatenate sources into one text block
  5. LLM node: fill API key (Kimi/DeepSeek/Qwen all work); 8k context is enough
  6. Push node: Feishu bot or DingTalk bot webhook
  7. Test run: trigger manually, check if the report lands in the group

Companion Prompt (LLM Report Generation)

Prompt
You are a project assistant. Generate today's work report from the raw data below, in this format:
1. Done today: 3-5 items, by importance
2. In progress: 2-3 items, with progress
3. Tomorrow's plan: 2-3 items
4. Risks/blockers: if any
Rules: one sentence per item, verb-first; don't fabricate anything not in the data; if data is missing, write "TBD"
Raw data: {{concat here}}

Pitfalls

  • Messy source data: add a Set node after each HTTP node to normalize fields-don't let the LLM guess the structure
  • LLM fabricates: stress "no fabrication" and "write TBD if missing" in the prompt-far more stable than free-form
  • Don't push too often: daily reports once a day, weekly once a week-real-time is just noise

This article is AI-assisted and human-edited. Last updated: 2026-07-26

Related