What Is a Job?
A job is a tracked enrichment operation that includes:- Products: Which products are being enriched
- Attributes: Which attributes are being generated (description, color, meta tags, etc.)
- Status: Current state (pending, running, completed, failed)
- Results: AI-generated content for review
- Metadata: Who started the job, when, and configuration details
Jobs run asynchronously in the background—you can start a large enrichment and continue working elsewhere. Cernel keeps processing even if you close your browser.
Job Lifecycle
Jobs progress through several states:1
Pending
Job is queued and waiting to start. If you have multiple concurrent jobs, they queue in order.
2
Running
AI is actively generating content. You’ll see real-time progress as products complete.
3
Completed
All products processed successfully. Content is ready for your review and approval.
4
Partial Success
Some products succeeded, some failed. You can review successful items and retry failures without re-processing everything.
5
Failed
Job encountered critical errors and couldn’t complete. Error messages explain what went wrong.
Monitoring Jobs
Dashboard Overview
The Dashboard shows all jobs in an activity feed. Each job card displays:- Progress Bar: Visual indicator of completion (e.g., “47/100 products enriched”)
- Status Badge: Color-coded state (Running, Completed, Failed)
- Attributes: Which attributes are being generated
- Timestamp: When the job started
- Initiator: Team member who started the job
Job Detail Page
Click any job card to open the detailed job view.Job Detail View
The job detail page provides granular visibility into every product and attribute being enriched.Job Header
At the top, you’ll see:- Job ID: Unique identifier (useful for support requests)
- Overall Status: Completed, Running, or Failed
- Progress: “78 of 100 products enriched (78%)”
- Started: Timestamp and initiator
- Configuration: Attributes being generated, sites targeted
Job Table
The main table shows every product in the job with individual status:| Column | Description |
|---|---|
| Product | Product name and thumbnail |
| Status | ✅ Success, ❌ Failed, ⏳ In Progress |
| Attributes | Which attributes were requested |
| Results | Quick preview of generated content |
| Actions | Review, Retry, or View Details |
Filtering Job Results
Use filters to focus on specific results:- Show All: Every product in the job
- Successes Only: Products that enriched successfully
- Failures Only: Products that encountered errors
- Pending Review: Items you haven’t reviewed yet
- Accepted: Items you’ve approved for application
Bulk Actions
Select multiple products to perform batch operations:Review & Apply
Approve multiple items at once
Bulk Rerun
Retry failed items
Export Results
Download enrichment data as CSV
Reviewing AI Content
The review drawer is where you evaluate AI-generated content before applying it to your store.Review Drawer Layout
- Side-by-Side View
- Attribute Actions
Compare original vs. AI-generated content for each attribute:Left Column (Original):
- Your existing product data
- What’s currently on your store
- AI-created content
- AI reasoning explaining the choices
AI Reasoning
Below each generated attribute, you’ll see AI reasoning—a plain-English explanation of why the AI made its choices. Example Reasoning:“Color identified as ‘Navy Blue’ based on product title ‘Men’s Navy Cotton Shirt’ and dominant color in product image. Material ‘Cotton’ extracted from title and reinforced by collection ‘Natural Fiber Clothing’.”This transparency helps you:
- Understand AI logic
- Spot incorrect assumptions
- Identify data quality issues in source products
AI reasoning is especially valuable when training your team to spot patterns in AI behavior and improve prompt configurations.
Applying Reviewed Content
After reviewing AI-generated content, you can apply changes back to your store.1
Review Products
Go through each product in the job and accept/edit/reject attributes
2
Select Items
Use checkboxes to select products with accepted changes
3
Click 'Apply Selected'
This pushes approved content back to your connected e-commerce platform
4
Confirm Sync
Cernel syncs with your store and shows success confirmation
Handling Failures
Jobs can fail partially (some products) or entirely (whole job). Understanding failure types helps you resolve issues quickly.Common Failure Types
Missing Primary Collection
Missing Primary Collection
Error: “Product must have primary collection before enrichment”Cause: The AI needs product context from the primary collection to generate appropriate content.Solution: Assign a primary collection to the product, then retry the enrichment.
Insufficient Data
Insufficient Data
Error: “Not enough product data to generate attribute”Cause: Product is missing key information needed for enrichment (e.g., no title, no description, no images).Solution: Add more data to the product in your e-commerce platform, sync, and retry.
Validation Failure
Validation Failure
Error: “Generated content failed validation (character limit exceeded)”Cause: AI generated content that violates configured constraints (too long, wrong format, etc.).Solution: Adjust prompt constraints or retry (sometimes the AI produces different output on retry).
API Rate Limit
API Rate Limit
Error: “Rate limit exceeded for AI provider”Cause: Too many concurrent requests to the AI service.Solution: Cernel automatically retries these. If it persists, contact support.
Platform Sync Error
Platform Sync Error
Error: “Failed to sync product back to Shopify”Cause: Connection issue with your e-commerce platform or invalid data.Solution: Check platform connection in Settings → Sites, ensure product still exists in store, and retry.
Retrying Failed Items
To retry failures:1
Open Job Details
Navigate to the job with failures
2
Filter to Failures
Use the “Failures Only” filter
3
Review Errors
Read error messages to understand what went wrong
4
Fix Issues
Address the root cause (e.g., assign primary collection, add missing data)
5
Select Failed Items
Use checkboxes to select which products to retry
6
Click 'Retry Failed'
Cernel creates a new job for just the failed products—no need to re-enrich successful ones
Retry operations are intelligent—Cernel only re-processes the specific attributes that failed, not all attributes for those products.
Job Performance
Job Speed
Enrichment speed depends on:- Number of products: More products = longer jobs
- Number of attributes: Each attribute requires AI processing
- Attribute complexity: HTML descriptions take longer than single-select attributes
- Concurrent jobs: Multiple simultaneous jobs share processing capacity
- Single-select attributes (color, size): ~2-5 seconds per product
- Short descriptions: ~5-10 seconds per product
- Full descriptions with SEO content: ~15-30 seconds per product
Concurrent Jobs
Cernel supports multiple concurrent enrichment jobs. However, each organization has a concurrent job limit based on plan tier. If you start more jobs than your limit allows, additional jobs queue and wait for running jobs to complete.Best Practices
Start Small, Scale Up
Start Small, Scale Up
For new attribute types or prompt configurations, enrich 5-10 products first. Review quality, refine prompts, then scale to larger batches.
Monitor Jobs Daily
Monitor Jobs Daily
Check the Dashboard daily to catch jobs that need review or have failures. Don’t let jobs pile up waiting for attention.
Use Job IDs for Support
Use Job IDs for Support
If you need help from Cernel support, always include the Job ID from the URL or job header. This helps us quickly diagnose issues.
Review Reasoning, Not Just Content
Review Reasoning, Not Just Content
The AI reasoning often reveals data quality issues in your product catalog (missing info, inconsistent naming, etc.) that you can fix to improve future enrichments.
Export Results for Analysis
Export Results for Analysis
Use the export feature to download enrichment results and analyze patterns—which attributes succeed most, common failure reasons, etc.
Job History & Auditing
All job activity is permanently logged for auditing and analysis:- View in Dashboard: Filter by date range to see historical jobs
- Product History Tab: See all jobs that affected a specific product
- Attribute History: Track how an attribute changed across multiple enrichments
- Export Logs: Download job history for compliance or reporting
Job logs are retained for the lifetime of your organization account and can be accessed anytime.
What’s Next?
Review Workflow Guide
Step-by-step guide to reviewing AI content
Troubleshooting
Solve common job issues
Product Enrichment
Learn about enrichment types
Bulk Operations
Enrich at scale
Next: Explore Product Enrichment to understand the types of content Cernel can generate.