Why Global Businesses Are Switching to Custom ERP & CRM Software in 2026 | ThriveUp SolutionCase Study: AI-Assisted Monthly Billing Automation

Project type: an internal automation we built for ThriveUp Solution’s own monthly billing. Client names and amounts are deliberately left out, and no time-saving figures are claimed.

The challenge

Every month, each client needs an invoice and a message. Doing that by hand means preparing documents one by one, writing individual notes, and sending them across email and WhatsApp, which is repetitive and easy to get wrong.

What we built

A Node.js automation that reads the client list from a simple spreadsheet file (name, email, WhatsApp number, status, service and amount) and processes each row:

  • Active clients: generates a PDF invoice, has the Claude API write a short personalised cover message, and sends both by email and WhatsApp
  • Inactive clients: sends a friendly re-engagement message by email and WhatsApp instead

Built-in safeguards

  • Rows with an invalid email, a badly formatted WhatsApp number or missing service or amount are skipped and logged, not sent
  • Clients are processed one at a time, so one failure is logged and does not stop the rest of the batch
  • Credentials are kept in environment secrets, not in the code

Scheduling

The job runs on a monthly schedule through GitHub Actions and can also be triggered manually. That means billing runs without anyone needing to remember to start it.

Technology

  • Node.js
  • PDF invoice generation with PDFKit
  • Claude API for personalised message text
  • Email delivery through SMTP (Nodemailer)
  • WhatsApp delivery through Twilio
  • GitHub Actions for scheduling

What is next

The same modules are being reused to build a web dashboard, so the client list can live in a database and the monthly send can be managed through a screen instead of a file.

Why it matters

This is a small example of practical AI: the model writes a friendly message, while ordinary code handles the invoice, validation and delivery. Keeping the risky parts in predictable code and the language work in the model is how we approach most automation projects.

Want something similar for your finance or operations team? Explore our AI & Automation service or contact us.