EOW Bot — Team Weekly Digest Automation
45
employees in the weekly reporting workflow
3
team leads previously coordinating submissions
1 day → ≤1 hour
end-to-end preparation and coordination time
1 flow
submission, storage, dashboard and draft generation
Description
What it does
Previously, 45 employees posted weekly results in free form, three leads tracked missing submissions and rewrote updates, and one lead consolidated a report for joint approval. The process consumed a full working day. The new workflow collects structured Telegram submissions, persists them in Google Sheets, visualises them in a dashboard and generates a sectioned management draft that AI helps edit for clarity. Leads retain final review and approval; end-to-end preparation and coordination now take no more than one hour.
Reliability and operational continuity
I defined the operational requirement that submissions should remain recoverable during temporary service or API disruption. Using AI-assisted development, I configured and tested safeguards for concurrent submissions, retries and pending updates, as well as reminder rules that exclude colleagues who are on holiday or have already submitted. After an availability incident, I completed an unplanned migration and validated that reporting could continue without losing submitted data. A subsequent security and continuity review strengthened deployment and recovery controls without exposing internal configuration details.
Key decisions & trade-offs
Keep the data path independent from AI editing
Submissions remain stored and recoverable even when the AI editing step is unavailable.
Fail closed for reminders
Reminder logic excludes colleagues who have already submitted or are unavailable rather than sending duplicate requests.
Treat availability as a release outcome
After an availability incident, recovery and real user access became part of release acceptance.
My Role
Owned the product end to end: defined the reporting workflow, submission structure, editing rules, dashboard and management digest. Used AI-assisted development tools to create and validate the working automation, then managed incident response and an unplanned server migration when the original environment failed.