Automations

Automations are how your organization's procedures become executable. The centerpiece is the playbook; around it sits a set of resources that give playbooks and AI agents the structure, data, and knowledge they need to act correctly.

Everything in this page is configured under Admin › Automations.

Playbooks

A playbook expresses a collection of procedures — the steps, the branching, the actions — in a form the platform can run. Playbooks are what make AI participation safe and consistent: an AI agent working a conversation or a task isn't improvising; it's following the procedure your organization wrote down.

Playbooks are managed in Admin › Automations › Playbooks.

The supporting resources

Data Types define the structured shapes your automations work with — the fields and forms of the domain data your procedures read and write.

Datasets hold structured reference data — lookup tables, program rosters — available to playbook steps and agents at execution time.

Knowledge Sets and Knowledge Bases hold the organizational knowledge that grounds AI responses — so answers come from your approved content, not from a model's general training.

Custom Actions define specific, bounded operations an AI can perform — a named capability with defined inputs and outputs, rather than open-ended behavior.

Assets and Asset Sets manage the files and resources (documents, images) your automations reference and deliver.

Analyzers evaluate conversations and work as they happen. Configure them in Admin › Automations › Analyzers; their results appear in Insights › Analyzer Dashboard, alongside the AI Eval Dashboard for evaluating AI quality.

Watching the machine

The Admin › AI Management section provides operational visibility into your AI usage:

• Features — which AI capabilities are enabled for your organization
• Usage — consumption over time, so cost and volume are never a surprise