Skip to content

Key concepts

This page explains the vocabulary of the platform. No need to memorize it: come back here when a term is not clear to you. In the end you will see how they all fit together.

flowchart TD
  T["Tenant<br/>tu organización"] --> E["Entorno<br/>e.g. Producción"]
  E --> A["Agentes<br/>las máquinas que ejecutan"]
  E --> S["Sesiones de robot<br/>cuentas de Windows (desatendido)"]
  E --> P["Procesos<br/>paquete + versión, listo para correr"]
  P --> X["Ejecuciones<br/>cada corrida concreta"]
  E --> Tr["Triggers<br/>cuándo se lanza un proceso"]
  E --> C["Credenciales<br/>secretos para acceder a sistemas"]
  PK["Package<br/>designed in Designer"] -->|is assigned to| P

A tenant is your organization within Zoan Automation: a space completely isolated from any other client. It has its own subdomain (tu-empresa.cloud.zoansoftware.com), its users, its data and its configuration. Nothing is shared between tenants. See Architecture › Multi-tenancy.

An environment (environment) groups the resources of a work context: agents, processes, credentials and triggers. The usual thing is to separate Production, Testing and Development, so that an experiment does not affect what is already underway. A resource always lives within an environment. See Environments.

An agent is the component that executes automations on your infrastructure. It is the Zoan Agent program running on a computer or server: it connects to Zoan Cloud, announces that it is available and waits for orders. Each agent belongs to an environment. See What is an agent.

Estados posibles:

EstadoDescription
onlineConnected and available to receive executions
busyRunning a process right now
offlineDisconnected or no recent signal (heartbeat)

The Machine Key is the secret key that an Agent uses to identify itself to Zoan Cloud the first time it connects. It is generated when the agent is created in the portal and displayed only once. Think of it as the “password” for that machine. See Windows Installation.

A robot session is a Windows user account that the bot uses to run unattended execution. It gives the robot a desktop to work on when no one is logged into the machine — necessary to automate the interface (desktop or windowed browser) on a server. It is chosen when launching the execution. See Robot sessions.

A package is the artifact that contains the logic of your automation—what you designed in Designer—packaged and published to the platform with a semantic version number (e.g. 1.2.0). Versioning allows you to improve an automation without breaking the one already in production. See Publish and version.

A playbook is a concrete automation: the sequence of activities that is executed. A package can contain one or more playbooks (one main and other callable ones). “Playbook” is what you see and edit in the Designer; “package” is how it is published and distributed. See Designer Overview.

An activity is the minimum block of an automation: it performs a specific action (open a browser, read an Excel, send an email, decide between two paths). You build a playbook by combining activities. The complete catalog is in Activities.

A process is the configuration that associates a package (at a specific version) with an environment. It is the “executable unit”: what the agent runs when it receives an order. Separating package and process allows you, for example, to have the same package running in Tests with version 1.3.0 and in Production with 1.2.0. See Create a process.

An execution (execution) is a specific run of a process on an agent. It has its own life cycle and generates logs that can be consulted from the portal. If a process runs three times today, that’s three executions. See States and life cycle.

Estados posibles:

EstadoDescription
queuedIn queue, waiting for an agent to be available
runningRunning on the agent
completedFinished successfully
failedEnded with error
cancelledCancelada manualmente

A trigger defines when an execution is launched. It can be manual (you launch it from the portal or via API) or scheduled (with a cron expression, so that it runs only at a time or frequency). See Triggers.

A credential securely stores sensitive information — passwords, tokens, API keys — so your bots can access systems that require login without writing the secret in the playbook. The agent obtains it at runtime and is never exposed in the code or logs. See Credential Management.

A queue is a list of pending work items (for example, 500 invoices to process). It allows work to be distributed and one or more bots to take elements one by one, marking each one as successful or failed. It is the basis for high volume automations and retrying what failed.