Routines are custom automations that execute multi-step tasks without manual effort. Unlike basic scripts, Revo Routines combine structured logic with AI intelligence to handle complex, context-aware automations.

Routines were previously called Workflows. The feature is the same — only the name has changed.

Why this matters for email:

Routines can analyze incoming customer feedback, create Jira tickets automatically, and trigger email responses or Slack updates. When a support email arrives, a routine can route it, draft the reply using relevant product context, and log the issue, all before you even see it.

Structure: Trigger → Steps (Nodes) → Action
​
​Example: Meeting ends → Extract action items → Create Jira tickets → Post summary to Slack
​

Navigation:
Click Routines in the left sidebar to see your automation library. From here you can:

You don't have to build automations by hand. Just describe what you want in Revo Chat and Revo creates the routine for you — picking the trigger, adding the right steps, and connecting your tools.

You can always open a chat-built routine in the visual builder below to fine-tune it further.

Your routine begins with a trigger that tells Revo when to run. You can choose from six trigger types:

Manual Trigger
Start the routine only when you manually click "Run this routine." Useful for on-demand reports or testing before scheduling.

Scheduled Trigger
Run the routine on a repeating cadence. See Scheduling a routine below for the full set of options.

Important: Toggle the routine to "Active" or it won't run even with a schedule set.

Event Trigger
Starts automatically when something happens inside Revo. Available event options:

Webhook Trigger
Respond to events from external apps. An outside service sends an HTTP request to your routine's unique webhook URL, triggering the automation instantly (e.g., form submission, support ticket, code push).

Slack Trigger
Start the routine based on Slack events like a message posted in a specific channel, a reaction added, or a command typed.

User Input
Allow users running the routine to provide specific information before execution (e.g., date range for a report, customer name for analysis).

A trigger on its own is coarse. "On meeting created" fires for every meeting, including your internal standup. Trigger filters let you narrow a trigger to the events you actually care about, using the details of the event itself.

Add one or more filter rows under the trigger. Each row compares a field on the event to a value you choose, for example:

Each trigger declares its own filterable fields, so the options you see change depending on which trigger you picked.

Why use a filter instead of a condition step? A filter is evaluated before the routine starts, and it is a direct comparison rather than an AI judgement. That means it is:

Previously the only way to narrow a trigger was to hang a condition step off it, which meant paying for an AI call on every irrelevant event just to be told "no". If you built a routine that way, moving that logic into a trigger filter will make it faster and cheaper.

A scheduled trigger can run on any of these cadences:

You also set the timezone the schedule runs in.

Things worth knowing:

Click the small circle at the bottom of the trigger card to add nodes. Each node represents one step in your automation:

Revo Node
Give AI instructions in plain language (e.g., "Summarize the meeting and extract action items").

Integration Action Node
Run a pre-built action on one of your connected integrations as a direct API call — no AI involved. Pick the integration, pick the action, and fill its parameters from the routine's own data. Because there's no model call, these steps are fast, free, and do exactly what they say. Use this instead of a Revo Node whenever the step is a definite operation rather than a judgement call, like "create this Jira issue" or "add this row".

The available actions are read live from each integration, so connecting a new tool makes its actions available straight away.

Jira Node
Create issues in Jira with auto-filled fields (project, issue type, assignee, labels, custom fields) based on routine data or AI-generated content.

Confluence Node
Create pages in Confluence directly from routine outputs. Useful for auto-generating documentation, meeting recaps, or knowledge base articles.

Send Email Node
Send automated emails to specific recipients with content generated from previous routine steps (summaries, reports, alerts).

Slack Node
Post messages, summaries, or alerts to specific Slack channels based on routine results.

Condition Node
Create branching logic with "if/then" statements (e.g., "If priority is High, notify the CTO in Slack. Otherwise, create a standard ticket").

A condition can run in one of two modes:

Transform Node
Reshape, filter, or restructure data from previous steps before passing it to the next node. Useful for formatting outputs or extracting specific fields.

Create Insight Node
Generate a saved analysis or report in your Revo Insights feed. Perfect for weekly digests, competitive analysis, or strategic summaries.

Update Item Node
Modify existing Revo objects (meetings, issues, insights, user feedback, agent contexts) by updating specific fields with new values.

Analyze Video Node
Process meeting recordings using both visual and audio AI to detect emotions, engagement levels, hesitation, or non-verbal signals beyond the transcript.

Best practice: Use one clear instruction per node. Combining multiple tasks in a single node can produce inconsistent or unpredictable results.

Anywhere you can type a value — a Slack message, a Jira field, an integration action parameter — you can pull in data from the trigger or an earlier step using double braces:

{{input.title}}

Nested values work too, which is what you'll need for most event triggers, since they carry their data one level down:

{{input.meeting.title}}
​{{input.email.sender}}

If you previously found that nested paths rendered as blank and worked around it, you can now use them directly.

For manual/scheduled routines: Click Run this routine to test execution
​For event/webhook triggers: Perform the real action (e.g., create a test meeting or send a webhook payload)

Review the output to ensure each step executed correctly.

If you've added trigger filters, remember your test event has to match them or the routine won't start at all.

Go to the Runs tab to see execution history. Click any run to inspect:

This visibility makes debugging and improving routines straightforward.

Trigger: Meeting created (event trigger), filtered to external meetings only

Steps:

Because the trigger is filtered, your internal standups never start this routine — and never cost you anything.

Email impact: When a customer emails asking "What are the next steps from our call?" Revo's draft references the auto-generated action items and ticket numbers from the routine.

Trigger: Every Friday at 5 PM (scheduled)

Steps:

Email impact: When customers report issues, Revo's email drafts can reference "We've heard similar feedback from 12 customers this week and have prioritized a fix for next sprint."

Trigger: 5 days before scheduled release (scheduled based on cycle end date)

Steps:

Email impact: Stakeholder emails asking "Are we ready to ship?" get drafted responses citing actual blocker count, documentation status, and risk assessment.

Begin with a basic two-step routine (e.g., Meeting ends → Send summary to Slack) before adding complex conditions and branches.

If you can express "should this run at all?" as a comparison on the event, put it in a trigger filter. It's free, instant, and keeps your run history clean of routines that started only to stop again.

Each routine should solve one clear problem. Instead of one massive "do everything" routine, create specialized automations for specific needs.

Check the Runs tab weekly. If a step consistently fails or produces poor output, refine the instruction or add more context sources.

By default, routines are visible to your entire team. You can make a routine private if it's experimental or tied to personal tasks.

How to make private:

Important: Private routines are only editable by you, but the results (where reports go) depend on your output settings. You control whether results go to private channels or shared team spaces.