BBomerang.ioAI agents and automation
All case studies
In productionMultimodal support + CRM

The store counter and WhatsApp, on the same customer record

An agent that answers on WhatsApp understanding text, voice notes and images, and a connector that brings every in store sale to the customer's record in the CRM.

formen.ar

How the data flows

  1. Sale at the counter
  2. Management system
  3. Local connector
  4. Customer record

1-2 minfrom sale to CRM record

Client
Formen
Industry
Designer tailoring shop
Where
Corrientes, Argentina
Integrations
Kommo CRM, Dragonfish, Telegram, Google Reviews

The context

A designer tailoring shop that sells in store and serves customers on WhatsApp. The sales pipeline lives in the CRM, where people are identified by phone number; invoicing lives in a management system installed on the store's PC.

The problem

The store sells and WhatsApp talks, but the two never touched. The invoicing system does not store the customer's phone number and the vast majority of sales are billed to a generic end consumer, so the purchase and the conversation lived in separate worlds. When someone wrote asking about a suit, there was no way to know if they were already a customer or what they had bought.

What we built

  1. 01A WhatsApp agent that answers questions about sizes, styles and availability, understanding text, voice notes and images, and hands off to a person when the case needs human judgment.
  2. 02A connector installed on the store's PC that reads sales from the management system and writes them to the CRM: a note with the receipt, the catalog items and the sale amount.
  3. 03The salesperson enters the customer in an iPad form with the last four digits of the receipt; that number, not the time of day, is what links the sale to the person. If two sales could claim the same contact, it writes nothing rather than write to the wrong record.
  4. 04A service survey that goes out on its own: if the answer is positive, it sends the link to leave a Google review, and if not, the owner gets an alert with an open task.
  5. 05A watchdog that alerts on Telegram if the store's WhatsApp session drops, which used to disconnect without anyone noticing.

What changed

  • The purchase shows up on the customer's record one or two minutes after invoicing, with the items and the amount, without anyone entering anything twice.
  • Asking at the counter stopped being a new habit: what used to be typed by hand days later is now entered on the iPad.
  • Feedback comes in on its own, and the good ones turn into public Google reviews.
  • If a sale cannot be attributed with certainty, it stays unlinked and gets reported. The system would rather not write than clutter a customer's record.

Stack

WhatsApp Businessn8nEvolution APIOpenAIPowerShellSQL Server

Connects with

Kommo CRMDragonfishTelegramGoogle Reviews

Design something similar for your operation

No system is copied and pasted: what works here came from how this business runs. Tell us about yours and we will see what fits.

Tell us about your case

Other case studies