AI Workflow Jobs

Overview

AI Workflow Jobs represent individual executions of your AI Workflows. Each job applies a workflow's steps to generate insights, content, or analysis based on your defined variables and prompts. Jobs provide a detailed record of AI processes, allowing you to track outputs, monitor performance, and save results for future reference.

AI Workflow Jobs can be created on-demand for immediate execution or scheduled to run at specified intervals, giving you flexibility in how you deploy your AI processes.

Viewing Your AI Workflow Jobs

The AI Workflow Jobs list provides a comprehensive overview of all your workflow executions, presented in a searchable, sortable table.

AI Workflow Job Information

For each AI Workflow Job, you can see:

• Name - The descriptive title of your workflow job
• Status - Visual indicator showing the current state:
  • New - Created but not yet started
  • Enqueued - Waiting to begin processing
  • Running - Currently executing steps
  • Completed - Successfully finished all steps
  • Failed - Encountered an error during execution
  • Canceled - Manually stopped by user
• Last Updated - Relative time since the job was modified
• Last Completed - When the job last finished successfully
• Statistics - Key metrics displayed as badge counts:
  • Steps - Number of prompt steps in the associated workflow
  • Sites - Number of sites associated with this job

AI Workflow Job List Actions

• Search - Filter jobs by typing in the search field at the top of the list
• Workflow Filter - View only jobs associated with a specific workflow
• Multi-select - Click the checkbox next to jobs to select multiple items
• Delete - Remove selected jobs (with confirmation dialog)
• Duplicate - Create a new job based on an existing one
• View Details - Click anywhere on a job row to open its complete execution details

AI Workflow Job Details

The job details page provides comprehensive information about a specific workflow execution.

Job Header

• Name - Editable title that identifies this specific job execution
• Status Badge - Visual indicator showing the current execution state with textual updates when a job is running
• Action Buttons:
  • Start - Begin execution (for new or previously completed jobs)
  • Cancel - Stop a running job

Job Options

The Job Options section displays fundamental configuration settings:

• Workflow - The parent workflow this job is based on (clickable link to workflow)
• Status - Current execution state with detailed status information
• AI Model - The language model processing all steps. The AI Model is defined in the workflow.
• Workflow Steps - Number of steps in the associated workflow
• AI Credits - Estimated usage metrics:
  • Total credits required (estimate only)
  • Input token count
  • Output token count
• Notification - Option to receive alerts when job completes
• Sites - Associated websites this job applies to

Variables

The Variables section shows values used during this specific job execution:

• Workflow Variables - Default values inherited from the parent workflow
• Job Overrides - Custom values specific to this job execution
• Usage Indicators - Visual cues showing where variables are referenced

Variables can be edited before running a job to customize the execution without changing the parent workflow. There is a visual status indicator that show variables that have been overridden in the job.

Output

The Output section displays the final result of your workflow job execution:

• Content Display - Formatted presentation of the combined output.
• Status Indicators:
  • For new/canceled jobs - Prompt to start execution
  • For running jobs - Progress indicator
  • For failed jobs - Error notification
• Action Buttons:
  • Save to File - Download the output as a text file
  • Start Chat - Begin a conversation with the AI using this output as context
  • Download PDF - Export the output in PDF format
• Continuation Suggestions - AI-generated prompts for further exploration of the results

Step Outputs

The Step Outputs section provides detailed results from each individual step in the workflow:

• Sequential Display - Ordered list of all step results
• Step Information:
  • Step number and name
  • Full output content
  • AI model used (if different from workflow default)
  • Processing timestamps
• Action Buttons:
  • Run From Step - Restart execution from a specific point
  • Save to File - Download individual step output
  • Model Info - Details about the AI model used for this step

Benchmark Output

The Benchmark Output section shows performance testing results if configured:

• Content Display - Formatted presentation of benchmark analysis
• Dashboard Visualization - Interactive charts for default benchmarks
• Action Buttons:
  • Run Benchmark - Execute benchmarking independently
  • Save to File - Download benchmark results

Running AI Workflow Jobs

You can execute workflow jobs in several ways:

Starting a New Job

1. From the Workflows list, click the "Create Job" button next to your desired workflow
2. Provide a descriptive name for the job
3. Customize variables if needed
4. Click "Start AI Workflow Job Now" to begin execution
5. Monitor progress and view results on the job details page

Rerunning an Existing Job

For jobs that have completed, failed, or been canceled:

1. Navigate to the job details page
2. Click "Start AI Workflow Job Now" to re-execute with the same settings
3. Confirm the action in the dialog if prompted

Running from a Specific Step

To continue a job from a particular point:

1. Navigate to the job details page
2. In the Step Outputs section, find the step you want to start from
3. Click the "Run From Step" button
4. Confirm the action in the dialog if prompted

This feature is useful for debugging or when earlier steps don't need to be repeated.

Canceling a Running Job

If you need to stop execution in progress:

1. Navigate to the job details page
2. Click "Cancel AI Workflow Job"
3. The job status will change to "Canceled"

Working with Job Results

Viewing and Analyzing Results

• The Output section shows the final combined result
• Step Outputs provide detailed intermediate results
• Benchmark Output offers performance metrics

Saving and Sharing Results

Results can be exported in several ways:

• Save to File - Download as plain text
• Save and Start Chat - Export and continue working with the output in a chat interface
• Download PDF - Create a formatted document

Continuing the Conversation

After job completion, you can:

• Use the provided conversation suggestions to explore the output further
• Start a new chat with the output as context
• Create a new job based on the results

Managing AI Workflow Jobs

Deleting Jobs

To remove unwanted or outdated jobs:

1. From the jobs list, select one or more jobs using the checkboxes
2. Click the delete icon
3. Confirm the deletion in the dialog
4. Alternatively, delete individual jobs using the trash icon in each row

Note: Deleting a job removes all associated outputs and cannot be undone.

Duplicating Jobs

To create a new job based on an existing configuration:

1. From the jobs list, locate the job you want to duplicate
2. Click the duplicate icon
3. Provide a new name
4. Optionally modify variables or settings
5. Start the new job or save it for later execution

Best Practices

Job Naming

• Use clear, specific names that identify the purpose and iteration
• Consider including date information for recurring jobs
• If testing variables, note key parameters in the name

Monitoring Long-Running Jobs

• Jobs automatically update their status as they progress
• For complex workflows, check the step outputs to track progress
• Set the notification option to be alerted on completion

Troubleshooting Failed Jobs

• Check the step outputs to identify where the failure occurred
• Review variable values for potential issues
• Try running from a specific step after resolving problems
• Check if credit limits have been reached

Managing Resources

• Delete old or unnecessary jobs to keep your workspace organized
• Be mindful of credit usage for large or complex workflows
• Use benchmarking to optimize workflows before running full jobs

Site Association

• Associate jobs with specific sites to organize content by website
• Site association enables filtering and contextual awareness in outputs
• Multiple sites can be selected for jobs that apply across properties

Advanced Features

Variable Overrides

• Customize the behavior of a workflow without changing the original
• Test different scenarios by running jobs with varied inputs
• Save successful configurations by duplicating effective jobs

Step-by-Step Execution

• Run jobs from specific steps to save time during development
• Use partial runs to debug problematic sections of a workflow
• Build complex workflows incrementally by testing sections independently

Benchmarking

• Evaluate the performance of your workflow with objective metrics
• Compare different approaches using benchmark data
• Use insights to optimize for efficiency, quality, or cost

Job Status Monitoring

• Real-time updates as jobs progress through steps
• Automatic polling adjusts frequency based on activity
• Status indicators provide at-a-glance understanding of job state