ElectroNeek

Use case · Customer-facing

Overdue invoices chased by email and WhatsApp with reminders written for each customer

A daily workflow pulls open invoices and their ageing from the accounting system, applies your sequence (reminder at due date, +7, +14, call at +21), and asks an agent to write each message for that customer — tone by relationship and history, language by customer, invoice details and payment link included. Email goes through your mailbox; WhatsApp through the Business API with approved templates. Replies land with a person; promises to pay are logged against the invoice.

Who it is for
Finance teams and accounting firms with hundreds of open invoices, customers who respond on WhatsApp more than email (Latin America especially), and a collections process that runs from a spreadsheet.
Systems involved
QuickBooks Online / Xero / Alegra / Siigo / NetSuiteGmail / OutlookWhatsApp Business (Meta API)AI Workflows agentTeams / Slack

How the run goes

5 steps · logged per run

  1. Trigger01

    Daily 09:00

    Open invoices and ageing

  2. Agent02

    Write each reminder

    Customer, language, tone, stage

  3. System03

    Send by email or WhatsApp

    Per customer preference

  4. Person04

    Handle replies

    Disputes, promises, questions

  5. System05

    Log to the ledger

    Promise date, dispute flag

Agent decidesDocument becomes fieldsBot does the clicksPerson approves

How it runs, step by step

  1. 1

    A schedule pulls open invoices with due dates and customer records from the accounting system, and the customer's channel preference and language from the CRM or a table.

  2. 2

    For each invoice at a sequence step, an agent drafts the message within your rules — factual, polite, with the invoice number, amount, due date and the payment link — in the customer's language. Customers on a dispute flag or a payment plan are skipped automatically.

  3. 3

    Email is sent from the collections mailbox; WhatsApp through the Business API using pre-approved message templates for the first contact and free-form replies within the 24-hour window. Each send is logged against the invoice.

  4. 4

    Replies are read by an agent that classifies them — paid, will pay on a date, dispute, question — and either logs the promise date, flags the dispute to the account owner, or hands the conversation to a person in Teams with the context.

  5. 5

    A weekly collections summary shows ageing movement, promises kept and broken, and the invoices that need a call.

Limits — what this does not do

  • WhatsApp first contact requires Meta-approved templates; free-text is only within the 24-hour reply window.
  • The agent writes within your rules; it never negotiates a discount or a plan — those go to a person.
  • Legal collection notices are templates you approve, not generated text.

Questions about this process

  • They get a message that reads like your team wrote it, at the stage you set, in their language, and a person answers when they reply. The sequence is yours; many teams stop automated reminders after the second and switch to a call task.

  • Yes, through the WhatsApp Business API with approved templates; Portuguese and Spanish messages are written natively, not translated.

  • It includes the payment link from your provider; it doesn't process payments itself.

Build this on your systems in a guided sandbox.

Tell us the process; after a 30-minute call the sandbox is ready from 24 hours, and we build the first run with your team.