The Business Headache

Let’s be blunt: every first working day of the month, PolskieBiuro24’s team lost over 8 solid hours copying phone numbers, pasting e-mails and praying they hadn’t missed a debtor. Mistakes crept in, tempers flared, and nobody wanted to own the process. They didn’t need fancy SaaS. They needed their spreadsheet to do the grunt work for them – while keeping the rest of the workflow exactly the same.

My Brief, My Roles

“Don’t change how we work — just delete the pain.”

Challenge accepted. I stepped in as UX Strategist, Automation Engineer and API Integrator in one. My plan: a minimal-invasion design that your accountant can grasp and your CTO would still sign off.

Spreadsheet before workflow change
Spreadsheet before workflow change
Spreadsheet after workflow change
Spreadsheet after workflow change

Designing for Zero Friction

  • One extra column per month, that’s it. Each month now has a WINIEN flag. Tick it and walk away.

  • Templates live where the data lives. Two new tabs — SMS and Mail — carry four message variants (2-, 3-, 4-, >4-month debt). Anyone can rewrite copy without calling me.

  • Cron-driven workflow. At 11:00 on day 06 the bot:

    • Authenticates via Google Sheets API, reads the last six months.

    • Maps debt age → correct template.

    • Sends personalised HTML e-mail (SMTP) + SMS (SMSAPI).

    • Pushes a signed log to a secure dashboard.

  • Security basics done right. All API keys, SMTP creds and sender IDs live in ENV files — never hard-coded, never pushed to Git. Rollbacks are a one-liner.

Python code snippet
Python code snippet
Mailing history log
Mailing history log

Impact in Plain Numbers

MetricBeforeAfterΔ
Manual labour per month>8 h< 5 min-95 %
Typos / missed debtors15 %0 %-100 %
Avg. debtor response14 days48 h4× faster
Team morale“Billing day… sigh”“Already done?”+ 😊

“I finally stop dreading the first of the month—our sheet basically writes reminders in its sleep.”

Ready to reclaim your time?

Let’s automate!

The Process

Building a “ghost-in-the-spreadsheet” sounds like wizardry, but it’s really a disciplined loop of research, design and engineering. Here’s the quick-and-dirty roadmap that turned a manual eight-hour chore into an invisible assistant:

Discover

Mapped every click in the legacy sheet, interviewed three frontline users.

Design

Prototyped column tweaks in Google Sheets, validated with stakeholders in 30 min.

Build

Crafted a 180-line Python script: pandas, gspread, requests. ENV for secrets.

Test

Ran dry-runs with anonymised data, simulated API failures.

Deploy

Docker container on a tiny VPS, cron schedule, uptime monitor ping.

Wanna more?

Check All My Works