Skip to main content
A process is the playbook for one type of service: how your company resolves a request for a duplicate boleto, a refund, or a scheduling, step by step. Once mapped, it becomes the script Tino follows in real conversations.
Who sees this: Administrator and Company Manager. The entire Processes area requires the Scale plan.The Tartini interface is in Portuguese. Screen and button names appear here exactly as on screen, with the translation in parentheses.

What a process is

Each process brings together:
  • 5W2H: the what, why, who, when, where, and how much (priority) of that service.
  • Passo a passo (Step by step): the steps Tino follows, with the text it sends at each one.
  • Saídas (Outputs): where the conversation goes when the process ends, for each possible outcome (Concluído (Completed), Parcial (Partial), Falhou (Failed), or Faltou dado (Missing data)).
  • Cobertura do Tino (Tino’s Coverage): how much of that process Tino resolves on its own, calculated step by step.
A mapped process only takes effect in service once it’s active. Processes created through automatic discovery start out inactive, unless you choose to publish them already active.

The Processes submenu, in screen order

Where to find it: Cérebro > Processos (Brain > Processes).
  1. Mapeamento (Mapping): lists every process already mapped and is the entry point for creating a new one, by talking with the mapping agent.
  2. Descobertas (Discoveries): queue of process candidates that Tino identified on its own, from content already indexed in Cérebro or from documents uploaded specifically for that analysis.
  3. Grupos (Groups): sets of processes with their own output path, used by campaigns.
  4. Catálogo (Catalog): an audit of the process records, independent of conversations, showing structural problems and how to fix them.

How Tino uses a mapped process

During service, Tino follows the step-by-step of the process that matches the customer’s request. At each step, it uses the text set in the mapping, or writes the message on its own when the script doesn’t define fixed text. At the end, it follows the output set for that outcome, or returns to triage if no output was defined. The one who puts the process in writing with you is Quattro, Tartini’s fourth agent. You explain, in your own words, how service happens today; Quattro asks what’s missing and builds the process, with steps, transitions, and outputs. A step that uses one of your connectors, for example to check a system, is included too: you choose which tools are available for that process. Nothing it builds takes effect without your approval. Once published, the process becomes the playbook Tino follows in real conversations, and the yardstick Scout uses to measure each one.

How Scout tracks adherence

Once a process is live, Scout audits real conversations against what was mapped and shows what Tino resolved on its own, where it needed help, and where the planned process didn’t match what actually happened. This execution analysis lives in Scout, outside the Cérebro.

Where each task points to


Frequently asked questions

Quattro is the one who leads the mapping conversation with you. Tino is the one who actually serves the customer, following the process once it’s published.
It shows as “Sem playbook”, different from a process with 0% coverage. Finish the mapping so Tino can run it.
In Scout, on the process adherence screen, outside the Cérebro.