Approval Workflows

Trame includes built-in human-in-the-loop controls that let you review and approve workflows before they execute. This system ensures safety and oversight for automated processes, especially during testing and production rollouts.

How Approval Gates Work

Workflow Status Controls

Approval behavior is controlled by workflow status:

Draft: All runs (manual and triggered) pause for approval
Pilot: Triggered runs start but pause for approval before execution
Live: Runs execute immediately without approval
Paused: All trigger events are ignored

Approval Process Flow

Trigger: Workflow is started (manually or via trigger)
Planning: AI generates execution plan with steps and tools
Approval Gate: Execution pauses, approval request is created
Human Review: Approver reviews planned actions
Decision: Approve to continue or reject to skip
Execution: If approved, workflow runs with planned steps

Approval Request Details

What Approvers See

Each approval request includes:

Workflow context: Name, description, and trigger source
Planned steps: Detailed list of actions the AI will take
Required tools: Which integrations will be used
Input data: Trigger payload or manual run parameters
Execution timeline: When the request was created and when it will timeout

Approval Interface

Approvers can:

Review planned actions in detail before execution
Approve to allow the workflow to proceed
Reject with optional feedback for learning
View context about the trigger or manual run source

Safety Timeouts

Approval requests automatically expire if not responded to
Configurable timeout periods prevent indefinite waits
Expired requests are marked as skipped with timeout reason
Notifications can alert approvers of pending requests

Managing Approvals

Who Can Approve

Approval permissions are based on organization roles:

Owners: Can approve any workflow in the organization
Admins: Can approve workflows they have access to manage
Members: Cannot approve workflows (execution only)

Approval Notifications

The system provides multiple notification channels:

In-app notifications: Real-time alerts in the dashboard
Email alerts: Notifications sent to approver email addresses
Slack integration: Optional Slack notifications (if connector enabled)

Approval History

Complete audit trail includes:

Who approved or rejected each request
When decisions were made with timestamps
Rejection reasons and feedback provided
Approval response times for performance tracking

Workflow Status Strategies

Draft Mode

Use for: Initial development and testing

All executions require approval
Safe for experimenting with new workflows
Allows iteration without risk of unintended automation
Perfect for validating workflow behavior

Best practices:

Test thoroughly in Draft before promotion
Use for workflows that touch sensitive systems
Validate all edge cases and error handling
Ensure proper connector permissions

Pilot Mode

Use for: Controlled production rollouts

Automated triggers start the workflow
Human oversight maintains execution control
Builds confidence before full automation
Allows monitoring of real trigger events

Best practices:

Monitor approval patterns and response times
Document common approval decisions
Train team members on expected behaviors
Gradually build confidence in automation

Live Mode

Use for: Full production automation

No approval gates - immediate execution
Maximum automation efficiency
Requires high confidence in workflow reliability
Suitable for proven, stable workflows

Best practices:

Promote only after extensive testing
Monitor execution logs closely after promotion
Maintain rollback procedures
Document incident response procedures

Rejection Handling and Learning

Rejection Feedback

When rejecting a workflow execution, approvers can provide:

Reason for rejection: Why the execution was not appropriate
Improvement suggestions: What should change for future approvals
Context feedback: Additional information about the trigger or situation

Learning Integration

Rejection feedback is used for:

AI improvement: Teaching the system about inappropriate executions
Workflow refinement: Identifying configuration improvements
Pattern recognition: Understanding when approvals are typically rejected
Process optimization: Reducing unnecessary approval requests

Common Rejection Patterns

The system learns to identify:

Inappropriate timing: Workflows triggered at wrong times
Insufficient context: Missing information for proper execution
Permission issues: Attempting actions without proper authorization
Business rule violations: Actions that violate organizational policies

Advanced Approval Features

Conditional Approvals

Configure approval requirements based on:

Trigger source: Different rules for different input channels
Time of day: Require approvals outside business hours
Data sensitivity: Additional oversight for sensitive operations
Execution scope: Extra approval for high-impact actions

Approval Workflows

Chain approvals for complex scenarios:

Multi-level approval: Sequential approval from different roles
Parallel approval: Multiple approvers must all consent
Escalation paths: Automatic escalation for timeout or rejection
Conditional routing: Different approvers based on content or context

Integration with External Systems

Connect approval processes to:

ITSM tools: Service management approval workflows
Business process platforms: Enterprise approval routing
Communication tools: Slack, Teams, or email-based approvals
Audit systems: Compliance and governance platforms

Monitoring and Analytics

Approval Metrics

Track important approval performance indicators:

Response times: How quickly approvals are processed
Approval rates: Percentage of requests approved vs rejected
Timeout rates: How often requests expire without response
Approver distribution: Which team members handle most approvals

Workflow Performance Impact

Understand how approvals affect automation:

Execution delays: Time from trigger to completion
Throughput impact: How approvals affect automation volume
User satisfaction: Response times and process efficiency
Automation confidence: Trends in approval requirements

Best Practices

For Workflow Creators

Start conservative: Begin with Draft status for all new workflows
Test thoroughly: Validate behavior before requiring approvals from others
Document clearly: Provide clear workflow descriptions for approvers
Monitor patterns: Learn from approval/rejection trends

For Approvers

Review carefully: Understand planned actions before approving
Provide feedback: Use rejections as learning opportunities
Respond promptly: Minimize delays in automation processes
Communicate concerns: Share feedback about workflow improvements

For Organizations

Define clear policies: Establish guidelines for approval requirements
Train team members: Ensure approvers understand workflow context
Monitor performance: Track approval metrics and process efficiency
Iterate on processes: Continuously improve approval workflows

Troubleshooting

Approval Requests Not Appearing

Check user role permissions for approval access
Verify notification settings and email delivery
Confirm workflow is in Draft or Pilot status
Review workflow trigger configuration

Slow Approval Response Times

Analyze approver availability and workload
Consider expanding approver groups
Review timeout settings for appropriateness
Implement escalation procedures for critical workflows

Frequent Rejections

Review workflow configuration and trigger conditions
Provide additional context or training to approvers
Consider refining trigger criteria or business rules
Evaluate if workflows should remain in Pilot longer