The problem: 3 hours lost every Monday
Every week I spent an average of 3 hours preparing the project status report: extracting Jira data, calculating KPIs, writing the executive summary, formatting, sending to the right people. A time-consuming, repetitive, yet indispensable process.
In February 2025, I automated the entire process with Gumloop in just under 2 hours of configuration. Here's exactly how.
The workflow: 6-step architecture
Step 1 — Trigger: weekly schedule
The workflow fires automatically every Monday at 7:00 AM. Node: Schedule Trigger, cron: 0 7 * * 1
Step 2 — Jira data extraction
Node: HTTP Request to the Jira REST API. I extract: all current sprint tickets with their status, tickets created last week, tickets blocked for more than 48h, velocity over the last 3 sprints.
Step 3 — KPI calculation
Node: Code Executor (JavaScript). I calculate: sprint completion rate, story points done vs planned, open critical bugs, internal team NPS.
Step 4 — Narrative report generation
Node: LLM Call (Claude 3.5 Sonnet). System prompt:
You are a senior project manager. From the sprint data provided,
generate a concise status report for management.
Format: executive summary (3 sentences), highlights (5 bullets),
identified risks, required actions.
Tone: professional, factual, no technical jargon.
Step 5 — Audience personalisation
Node: Condition + two LLM branches. Management version (executive summary) vs technical team version (full detail). AI automatically adapts the level of detail.
Step 6 — Multichannel distribution
Node: Send Email for management, Post to Slack for the technical team, Create Confluence Page for archiving.
Results after 3 months
- ⏱️ Time saved: 2h45 per week (≈12 days/year)
- ✅ Quality: rated "better or equivalent" by 4/5 stakeholders
- 📊 Punctuality: 100% of reports delivered before 8 AM Monday (vs 70% manually)
- 🔄 Maintenance: 15 min/month to adjust prompts
"This isn't just automation. It's a transformation of my relationship with reporting: from a chore to an intelligent system working for me."
