Workflow platforms get all the attention, but a lot of valuable automation is smaller than that: a Python script that runs on schedule and does exactly one job. Here are seven I've built for clients that paid for themselves quickly — each one is a day of work or less.

**1. Price and stock monitoring.** A scraping script checks competitor or supplier pages daily and diffs yesterday: new products, price moves, out-of-stocks. One client used it to catch a supplier quietly removing their best-selling line two weeks before stock ran out.

**2. Lead list enrichment.** Before a sales push, a script merges your CRM export with public directory data — company websites, recent news mentions — so the team calls with context instead of cold. The value isn't the data; it's that nobody spends a day building the list by googling.

**3. Invoice folder wrangler.** PDFs arrive from five email addresses; the script reads attachments, normalises file names, sorts them by month and vendor, and flags duplicates. Accounting closes the month faster because "find the missing invoice" becomes a search instead of an archaeology project.

**4. Report generator.** Pulls data from your tools' APIs overnight and writes a one-page summary — numbers plus a delta column — into a shared doc. The Monday report meeting survives on it, and nobody builds spreadsheets by hand.

**5. Data cleanup bot.** Contacts with typo'd emails, duplicated names, wrong phone formats — a script re-exports the CRM, applies normalisation rules, and reports what it would change before writing anything back. Boring, but every marketing email sent to a dead address is money spent on nothing.

**6. Webform-to-database bridge.** That old form that writes to a mailbox somewhere? A script with an webhook receiver turns submissions into rows in whatever database you actually use — often the first step before a fuller Make/n8n automation.

**7. Backup and archive job.** Database dumps and media folders, compressed, versioned and shipped to an offsite bucket with an alert if it silently fails. It's the least glamorous script on this list and the only one your business structurally depends on when a disk dies.

**When a script beats a platform:** the job is a single transformation on a schedule, the data is small, and you'd rather not pay monthly for it. **When it doesn't:** anything a non-technical team member must edit themselves — that's when the visual canvas of n8n or Make wins.

Each of these starts with the same conversation: what file, spreadsheet or inbox do you dread opening? A free automation audit is the cheapest way to have it.