WhatsApp customer service automation with AI, 24 hours a day
An AI agent answers customers on WhatsApp at any time, checks the order status in your system and hands the conversation to a person when the subject calls for it.
How it works today, without automation.
A customer asks about their order at 10 p.m. and only gets an answer the next day. By morning the WhatsApp queue is full, and most messages are the same question: "where is my order?".
- 01Outside business hours, nobody answers, and the customer buys from a competitor.
- 02To answer, the agent opens the ERP, looks up the order and copies the status into the chat.
- 03Repeated questions about delivery times, shipping and returns take up the team's day.
- 04The history is scattered across phones, and nobody knows what was already agreed.
The flow, step by step.
-
01
Customer sends a message
At any time, to the company's WhatsApp number.
-
02
AI understands the request
Identifies the customer and the subject: status, delivery time, return, quote.
-
03
Checks the system
Looks up the order in the ERP or e-commerce platform and answers with real data.
-
04
Hands over to a person
Complaints, negotiations or off-script questions go to an agent, with a summary.
-
05
Everything is logged
The conversation and the outcome go into the CRM or help desk.
What would be delivered.
Deliverables
- AI agent on WhatsApp with the company's rules and tone of voice.
- Order, delivery time and status lookups straight from your system.
- Handover to a human agent, with a conversation summary.
- Conversations logged in the CRM and a dashboard of the main topics.
- WhatsApp Business API
- ERP (Bling, Tiny, Omie and others)
- E-commerce
- CRM
Stages
-
Discovery2 to 3 days
Mapping the most common questions and the data the AI can access.
-
Building the agent2 to 3 weeks
Integration with WhatsApp and your system, rules and answers.
-
Testing with the team1 week
The agent answers in a test environment and the team fine-tunes the answers.
-
Go-live and tuning1 week
Production launch, conversation monitoring and improvements.
The final timeline is set in the proposal, after discovery.
Questions about this use case.
Does the customer know they are talking to an AI?
Yes. We recommend making it clear from the first message, and the customer can ask to talk to a person at any time.
What if the AI doesn't know the answer?
It doesn't make things up. When a question falls outside what was configured, the conversation goes to an agent, with a summary of what the customer already said.
Do I need to change my WhatsApp number?
Not necessarily. Your current number can be migrated to the official WhatsApp API. We assess this during discovery.
Can the AI check my ERP?
Yes, if the system has an API or another way to access the data. Bling, Tiny, Omie and most e-commerce platforms offer this.
Want to see how it would work in your company?
You explain the problem. We build the solution. A straight answer, no generic proposal.
See also.
Automatic document reading
Invoices, freight bills (CT-e), contracts and receipts read by AI and posted to the ERP.
// 03Order and lead triage
Requests from several channels classified and sent to the right person.
// 04Administrative routines on autopilot
Bank reconciliation, registrations and collections done by AI, with human approval where it matters.
// 05Assistant for company documents
The team asks and the AI answers based on internal manuals and policies.
// 06Reports that build themselves
The weekly summary comes out ready, without anyone merging spreadsheets.