How I turned an aluminum company into a connected, documented process
Case study by Achiya Cohen, founder of Achiya Automation
The project built and deployed four n8n automations connecting the phone, website, WhatsApp and Airtable. From a customer request through closing the deal and producing the invoice, the steps were designed as a documented flow with a path for exceptions.
This project, and the evaluation method used to test the system before it went live, were featured in MarketingSherpa. Read the coverage (opens in a new tab)
Every deal — hours of coordination From a closed deal to a documented process
- The secretary calls the surveyor by hand after every deal
- Customer details were copied manually from Excel into invoices
- End-of-month commission calculations required manual work
- There was no practical way to review every agent call manually
- Deal closed → invoice + surveyor notification according to the flow
- The systems sync updates and log exceptions
- AI transcribes and summarizes calls captured by the flow
- A management dashboard was included in the project
Here's how it works inside
Each of these flows was built in the project and is shown here without identifying details.
Deal closed → invoice → customer
The moment a deal is marked "closed" in the CRM, n8n pulls the customer's details, generates a tax invoice, saves it to Drive and sends it to the customer on WhatsApp with a thank-you message — in Hebrew, in the company's tone. The agent just clicks "close deal" and moves on to the next one.

Status change → surveyor alert
When a status changes in Airtable, the flow checks whether it is relevant to the surveyor, who the assigned surveyor is and what language to use. The WhatsApp alert includes the address, time and a confirmation link.

Status change → customer update
Same trigger, a different message. The customer gets a personal note, in natural Hebrew — not a canned template — explaining where their order stands. There's de-duplication and smart throttling so they don't get spammed.

Phone call → transcription → CRM
A call captured by the flow is transcribed and classified (new customer / technical question / spam), contact details are extracted and the CRM is updated. An exception remains visible for review instead of being silently discarded.

The numbers you feel every day
Want to measure how long your manual process takes?
An introductory call to map the process, constraints and whether there is a suitable project scope.
Want a system like this? Check out our WhatsApp bot service or our pricing & plans.