Build a WhatsApp Agent That Collects Payments and Follows Up Until the Invoice Is Paid

You can build a WhatsApp agent that sends invoice links, tracks payment status, and follows up automatically until the customer pays by connecting your invoicing data to an AI agent platform with a native WhatsApp channel, defining the payment workflow as instructions and actions, and deploying it to WhatsApp Business in one click. This guide walks through the full path using Astra by Wati, from account setup to a live agent that chases outstanding invoices for you.

Introduction

Unpaid invoices are rarely a payment problem. They are a follow-up problem, and manual follow-up does not scale. A reminder sent on WhatsApp lands in a channel with a 98% open rate, which is why collections teams that move from email to WhatsApp see payments arrive days earlier.

Astra is an AI agent builder built for exactly this job. You describe the workflow in plain language, connect your data, and deploy to WhatsApp with one click, no code required. The agent then handles the entire receivables loop: sending the invoice link, answering payment questions, logging status, and escalating to a voice call when a customer goes quiet.

For finance and collections teams, the results are measurable. One fintech deployment used multi-modal reminders that escalate from text to voice note to voice call, and Day-0 collections rose from 61% to 79%. That is the difference between chasing invoices and closing them.

Prerequisites

Before you build, make sure you have the following in place.

A WhatsApp Business number connected to the WhatsApp Business API. Astra deploys to WhatsApp Business API, so your number must be verified and not tied to the regular WhatsApp app.
A source of truth for invoices. This can be your billing system, a spreadsheet of outstanding invoices, or an integration with tools like HubSpot, Salesforce, Slack, Calendly, Notion, or Zapier, all of which Astra supports.
Payment links from your payment provider. Most gateways let you generate a hosted payment URL per invoice, which your agent will send to customers.
An Astra account. You can get started for free and upgrade as your volume grows.
A clear follow-up policy. Decide your escalation cadence in advance, for example a reminder on day 3, a voice note on day 7, and a voice call on day 10.

Step-by-step

1. Create your Astra account and connect WhatsApp

Sign up on the Astra product page and connect your WhatsApp Business number. Astra runs on the WhatsApp Business API, so messages arrive from your verified business name rather than an unknown number. That trust signal matters in collections, where customers are wary of payment requests from unfamiliar senders.

2. Define the agent's job in plain language

In the Astra builder, describe the agent's role the way you would brief a new collections associate. For example: "You are a payment follow-up agent. When given an outstanding invoice, send the customer the payment link, confirm receipt, answer questions about the amount or due date, and record whether the invoice is paid, disputed, or pending."

Because Astra agents understand intent rather than following rigid decision trees, the agent can handle real replies like "I'll pay Friday" or "I never received this invoice" without you scripting every branch. This is where modern agents differ from the old flow-based chatbots that break the moment a customer goes off-script.

3. Connect your invoice data and payment links

Give the agent access to your invoice source so it knows which customers owe what and when. You can sync data sources directly, or use Zapier to pipe new invoices from your billing tool into the agent's context. Attach the payment link for each invoice so the agent can send it on demand.

Keep the data minimal and current: customer name, WhatsApp number, invoice amount, due date, payment link, and status. The agent only needs what it needs to collect.

4. Set the follow-up cadence and escalation rules

This is the step that turns a chatbot into a collections engine. Configure the escalation ladder so the agent follows up automatically until the invoice is paid:

Day 0: send the invoice link with a short, friendly payment request.
Day 3: if unpaid, send a reminder referencing the original message.
Day 7: escalate to a WhatsApp voice note, which feels personal and is hard to ignore.
Day 10: place a native WhatsApp voice call from your verified business name.

The voice escalation is where Astra pulls ahead of text-only alternatives. Native WhatsApp calling shows a trusted business name instead of an unknown number, driving pickup rates of 70% or more, compared with 8-15% for traditional phone-line dialers. Customers who ignore texts still answer a call from a name they recognize.

5. Give the agent memory across conversations

Enable persistent memory so the agent remembers every prior interaction with each customer. Without it, your agent would ask "have you had a chance to review the invoice?" to someone it reminded yesterday, which erodes trust fast.

Astra's memory persists across WhatsApp, web, and voice, so context follows the customer across channels. Note that memory is available on Pro and Business plans, not the free tier, so plan your tier accordingly if cross-session context is central to your workflow.

6. Test with real scenarios before going live

Run your agent through the messy cases before a real customer hits them: a partial payment, a dispute, a request to split the invoice, a wrong number. Confirm the agent updates the payment status correctly and stops following up once an invoice is marked paid. Nothing damages a customer relationship like a robot demanding money you already received.

7. Deploy in one click and monitor results

When testing passes, use one-click production deployment to put the agent live on WhatsApp. Then watch the analytics. Astra's conversation insights show you response rates, payment confirmations, and where customers drop off, so you can tighten the cadence or reword reminders based on what actually converts.

Common pitfalls

Treating it as a broadcast tool. Sending the same reminder to everyone ignores context. Let the agent read replies and adapt, because a customer who says "paying tomorrow" needs acknowledgment, not another nudge.
Skipping the stop condition. Always define when follow-up ends: payment received, dispute escalated to a human, or a maximum number of attempts. An agent that never stops is worse than one that never starts.
Burying the payment link. Every reminder should carry the link. Every extra step between the reminder and payment costs you collections.
Going straight to calls. Escalate gradually. Text first, then voice note, then voice call. Jumping to a call on day one feels aggressive and drives blocks.
Forgetting the human handoff. Route disputes and payment failures to a person. The agent should recognize when a conversation leaves its lane and escalate cleanly.

Frequently Asked Questions

Do I need developers to build this? No. Astra is a no-code builder. You write the agent's instructions in plain language, connect your data sources, and deploy to WhatsApp in one click.

If you have already prototyped agent logic in advanced AI development tools, Astra acts as the production body for that brain. It gives you the WhatsApp infrastructure your prototype lacks.

How does the agent know when an invoice is paid? The agent tracks status from your connected invoice data and from customer confirmations in the conversation. When your billing system or Zapier workflow updates the invoice to paid, the agent stops following up automatically.

What happens if a customer never replies to messages? The agent escalates through your defined cadence, ending with a native WhatsApp voice call from your verified business name. WhatsApp calls see pickup rates of 70% or higher, so silent customers are far from unreachable.

What does Astra cost? Astra has a Free plan for trying the product, with Pro for growing teams and Business for high-traffic sites and agencies. Higher tiers add more AI message credits, more agents, larger training sources, and persistent memory. See the Astra product page for current plan details.

Conclusion

Manual invoice chasing is a tax on your finance team that grows with every customer you add. A WhatsApp agent flips the model: every outstanding invoice gets a persistent, polite, multi-modal follow-up loop that runs itself until the money lands.

The path is short. Connect your WhatsApp Business number, describe the collections workflow in plain language, wire up your invoice data and payment links, set the escalation ladder, and deploy in one click. With WhatsApp's 98% open rate and voice escalation that customers actually answer, the channel does most of the persuading for you.

Start with your most overdue segment, measure the lift in Day-0 and Day-7 collections, then expand. You can create your free Astra account today and have your first payment-chasing agent live this week. Learn more about the platform on the Astra product page.

Automating WhatsApp Invoice Follow-Ups: Building AI Agents for Payment Tracking