Sontairo AI

AI agents & workflow automation

Sontairo AI is a workspace for AI chat, agents, connected tools, persistent memory, and workflow automation. Teams can give an agent a defined role, use it for one-off work, and build recurring workflows with explicit steps and a reviewable history.

Sontairo · Updated

Who it is for

Teams coordinating recurring work across business tools, especially when the same instructions, project context, and records need to carry from one task to the next.

What you can do with Sontairo AI

Agents with a defined job

Give each agent a name, instructions, and a model configuration. Connect the tools it needs, then use it in chat or send it a task. Instructions and connections make the scope of each agent visible.

Context that carries forward

Persistent agent memory retains useful context across conversations. Projects organize related work so a customer workflow and an internal operations workflow can have their own context.

Workflows you can inspect

Connect triggers, logic, AI, and explicit actions in a workflow. Check setup before activation and publish a version when it is ready. Later edits remain drafts until you publish them.

In practice

Example: a weekly operations briefing

Start with a scheduled trigger, gather the inputs using configured components, and ask an AI step to summarize the results in a fixed format. Connect a notification step to deliver the briefing. Inspect a failed component in run history instead of guessing which part of a long prompt went wrong.

How to get started

  1. Choose one task with a clear input and expected result.
  2. Create an agent and connect the tools needed for that task.
  3. For recurring work, build the workflow, resolve Check setup issues, and activate it.
  4. Review component results in run history and publish changes when you refine the workflow.

Before you start

Interactive agents and unattended workflow AI have different capabilities. An unattended AI step cannot ask follow-up questions or discover and call tools. Configure its inputs and output format in advance, and use explicit connector, webhook, or notification steps for actions.

Questions about Sontairo AI

How is Sontairo AI different from a single chat conversation?

A conversation is one way to use Sontairo. The workspace also stores agents, tool connections, project context, and memory, and supports versioned workflows with triggers and explicit actions.

Can I review an automation after it runs?

Yes. Run history records component status, duration, inputs, and outputs so you can inspect the step that completed or failed.

Can scheduled AI decide which tools to call?

Unattended AI components do not discover or call tools. Define actions as separate workflow components and resolve required setup before activation.

Setup documentation

Explore the rest of Sontairo