Case studies
Automations and chatbots in use
Three jobs, each described as a problem, what I did and the result. Details of the clients are left out.
WhatsApp chatbot with memory
The problem
Requests came in on WhatsApp at all hours, often from the same people about the same topic. A fixed-answer bot wasn’t enough: it started from zero at every message, and people had to explain everything again.
What I did
Every incoming message is saved to a sheet that works as the conversation’s memory. Before replying, the workflow reads that contact’s history, passes it to a language model together with the business information, and routes the case: a direct answer when it can be handled, a person when it can’t.
Automatic follow-ups for people who don’t reply
The problem
After the first message many contacts disappeared. Following up by hand meant keeping a list and remembering who had already been chased and when. In practice it almost never happened.
What I did
A daily check reads the contact list and, based on how much time has passed and which message each person already received, sends the next follow-up. After the fourth contact the sequence stops by itself, and every message is logged on that person’s row.
Invoices generated from orders
The problem
Every order was retyped into the accounting software: the same data copied twice, with the risk of a wrong amount or a duplicate customer.
What I did
When an order arrives, the workflow looks up the customer. If they exist, it issues the document directly; if they’re new, it creates them from the order data and then goes on. Data is read in a structured way, not copied by eye.
Result
The document is ready with no re-entry.
Reviews: 4.93 / 5 from 61 verified reviews on Fiverr (read on 18 November 2026). See the profile.