A swimming team captain handed ChatGPT three scheduled tasks and cut her routine work by a multiple
Maria coaches an amateur swim team spread across Russia, and she got tired of collecting race results by hand. Now three scheduled ChatGPT Plus tasks prepare the weekly announcement, gather registrations and results from organizers' websites, and push the data into a Google Sheet. By the author's estimate, the work takes several times less time than before, although she has no precise measurements. For a small club, a spreadsheet and manual checking are enough without a database.
AI-processed from Habr AI; edited by Hamidun News
Imagine volunteering to run an amateur team, where every race weekend turns into an evening of copy-pasting: who registered, who swam what, who scored how many points. Maria, an IT specialist and captain of a swim team scattered across Russia, described on Habr how she got rid of this routine. She gave ChatGPT Plus three recurring tasks that run on their own according to a schedule, and she keeps all the data in an ordinary Google Sheet. By her own estimate, the whole setup takes "several times less time" than even some of these chores took last season. The text contains no precise measurements, so this is a subjective impression rather than measured savings.
What problem they were solving
The team rarely meets offline: you can count the major gatherings of the season on one hand. What holds people together is online content: who is going to which race, which distances they are swimming, how to arrange transport and lodging, and where to meet on site. Everyone is also interested in team and individual statistics, but nobody wants to collect them by hand.
In early 2026, Maria built a small "content factory" around a Google Sheet, powered by GPT. She didn't bother building a database: the data volume is small, and extra engineering would only complicate maintenance. The sheet holds the results of all races: who swam where, how far, and with what time. It also calculates points for the intra-team standings. Swimmers can open the sheet and check their numbers, although, according to the author, they rarely do.
How the weekly sprint works
Three automated tasks are set up on ChatGPT Plus:
- In advance. The bot pulls the upcoming week's races from the calendar and drafts a poll for the team chat. All that's left is to copy it and paste it into Telegram.
- Before the weekend. It collects registration data for the team from the organizers' websites and writes a draft announcement: locations, participants, distances.
- After the weekend. It retrieves results from the organizers' protocol pages, prepares rows for the spreadsheet, and drafts a weekly results post: who swam how, and whether the team has placements in age groups and overall.
The data verification is especially valuable. ChatGPT flags discrepancies: a swimmer forgot to list the team at registration, swam for another club, or switched teams entirely, or the protocol has an error in a name, time, or place. Maria messages the swimmer, and the swimmer decides what to do about it. If an organizer runs its own statistics system, errors can creep in during transfer in either direction, and data can fall out of sync because of delays. Such cases also end up in the bot's report.
In the end, the manual work left is to quickly review the results, add to the post, and publish it. At the end of the season, they compiled rankings of organizers and the intra-team standings, and each swimmer's personal statistics were put into cards that are easy to open on a phone, save, and forward.
Where automation hits a ceiling
Maria openly lists the weak spots:
- The AI sometimes plays up with runs and makes mistakes. She catches some of the errors, and the swimmers catch others. The author considers this an acceptable price for getting rid of the routine.
- Even large organizers, whose data is well structured, make mistakes, and protocols are sometimes corrected after the fact.
- Local organizers offer less automation, so protocols have to be fed in manually. There are not many such races.
- Some sites can't be handled by the agent at all. The workaround: send the bot screenshots or PDFs, and it adds the data to the overall results.
What this means for practitioners
The case shows a working scheme for small communities: a schedule of tasks in a ready-made AI service, a Google Sheet as the only storage, and a person who checks the result before publication. Here the AI not only writes texts but also cross-checks data from several sources and reports inconsistencies. The texts remain drafts, and responsibility for publishing rests with a human.
The limitations are clear. The scheme depends on the stability of the source websites and on how the task scheduler behaves. Scaling it to dozens of teams or to other sports without rework is unlikely to work: Maria has already tried collecting statistics for other sports but has not yet reached a working version. For next season, she is thinking of having the bot handle some of the publications in the team's Telegram chat, and she is asking readers for ideas.
For anyone who administers a club, chat, or amateur league, the takeaway is simple: start with one spreadsheet and three recurring tasks, keep manual verification, and only then think about a database and custom code.
- Nov 21, 2026FDE job postings grew 4.6x in a year: what a forward deployed engineer actually does at the client's site
- Nov 21, 2026Anthropic Mathematician Claims Riemann Hypothesis Proof with Claude: 1.3 GB of Code
- Nov 21, 2026OpenAI GPT-6 Cyber Found About 35 Flaws in iPhone, Mac and Other Apple Devices
- Nov 21, 202622 Claude Agents Spent Six Weeks on Black Hole Problems — and Solved Two
Key Questions Answered
What problem they were solving?
The team rarely meets offline: you can count the major gatherings of the season on one hand. What holds people together is online content: who is going to which race, which distances they are swimming, how to arrange transport and lodging, and where to meet on site. Everyone is also interested in team and individual statistics, but nobody wants to collect them by hand. In early 2026, Maria built a small "content factory" around a Google Sheet, powered by GPT.
How the weekly sprint works?
Three automated tasks are set up on ChatGPT Plus: In advance. The bot pulls the upcoming week's races from the calendar and drafts a poll for the team chat. All that's left is to copy it and paste it into Telegram. Before the weekend. It collects registration data for the team from the organizers' websites and writes a draft announcement: locations, participants, distances. After the weekend.
Where automation hits a ceiling?
Maria openly lists the weak spots: The AI sometimes plays up with runs and makes mistakes. She catches some of the errors, and the swimmers catch others. The author considers this an acceptable price for getting rid of the routine. Even large organizers, whose data is well structured, make mistakes, and protocols are sometimes corrected after the fact. Local organizers offer less automation, so protocols have to be fed in manually. There are not many such races. Some sites can't be handled by the agent at all.
Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews
Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.