Case study · in production · Brazil
Dolce Vita is a decorated-cookie studio. Every order was born, negotiated, and closed inside WhatsApp conversations — and died there. No system, no reliable spreadsheet, no view of the month. It's the hardest context scenario there is: layer zero, everything in conversation and memory. Which is exactly why we chose to start with it.
Simple questions had no answer without digging: how much did we sell this month? Who hasn't paid yet? Which customers buy the most? How big is next week's production queue?
The information existed — all of it. But it lived scattered across hundreds of conversations, in the memory of whoever handled each customer, and in loose notes. No AI, no system, and no report can work with what was never recorded.
The diagnosis: it wasn't a tool problem. It was a context problem — the business's knowledge had never become a record.
The method's three fronts, applied to the case — without forcing anyone to abandon WhatsApp.
The system learned how the business operates: products, margins, customers, production cycle, and goals. Capture happens where the conversation already happens — a parser reads orders from WhatsApp and turns them into structured records.
Loose orders became live data: a queryable order ledger and a weekly report with revenue, receivables, and top customers. Questions that used to require digging now have immediate answers.
An assistant that answers with the business's real context — margins, history, seasonality — instead of generic AI responses. The owner asks in plain language and decides on data.
Stack: WhatsApp parser + live dashboard on Google Apps Script · in production since 2026
You probably don't make decorated cookies. But if your operation has information that only exists in conversations, decisions made from memory, and questions that require digging — your problem has the same structure as this one. Only the scale changes.
If the method works where information only existed in WhatsApp threads, it works where you already have a CRM. The degree of chaos changes; the method doesn't.
The same method, at another scale: currently applied in AI operations advisory for a global travel marketplace — structuring operational context and building team capability.
Compass answers in five minutes: describe your problem and leave with a diagnosis and a three-move plan.