Content Generation
Generate, translate, and improve content at scale with AI.
8 min read
Content Generation
Generate content for dozens or hundreds of entities in a single operation. Define a prompt template once, feed it your data, and let AI produce complete, structured content — blog posts, glossary definitions, comparison pages, product descriptions, and anything else your entity types define.
How It Works
Define your prompt template
Every entity type with content generation enabled has a prompt template. This template uses {{field.fieldName}} placeholders to inject entity data into the AI prompt. The template tells the AI exactly what to produce and in what format.
Write a comprehensive glossary definition for the term "{{field.term}}".
Requirements:
- Write a clear, concise definition (2-3 sentences)
- Follow with an expanded explanation (2-3 paragraphs)
- Include practical examples where relevant
- Use simple language accessible to beginners
- Naturally incorporate the term and related keywords for SEO
Return JSON with fields: definition, meta_title, meta_description,
category, related_terms
Select entities to generate
Navigate to your entity type table and select the entities you want to generate content for. You can select all, choose specific rows, or filter to only entities with incomplete fields.
AI processes each entity
WISEROWS creates a background job and processes entities in batches. The system uses your project's knowledge base (brand voice, personas, guidelines) to inform the generation. Each entity's field data is injected into the prompt template, so every generated piece is unique and contextual.
Review generated content
Generated items land in a review queue with status "pending." You preview each item, edit if needed, then approve or reject. This human-in-the-loop step ensures quality before content enters your main dataset.
Approve and publish
Approved content becomes a real entity in your system. From there, it follows your normal publishing workflow — through status stages (Draft, In Review, Approved, Published) and optionally out to your CMS via integration connectors.
Generation Modes
Each entity type's content generation uses one of four modes that controls how the AI approaches the task:
Mode | Description | Best For |
|---|---|---|
| data | Structured data extraction and formatting | Product specs, technical metadata, data normalization |
| mixed | Combination of structured data and prose | Landing pages, case studies with metrics + narrative |
| content | Long-form prose content generation | Blog posts, glossary definitions, guides |
| creative | Creative writing with more freedom | Marketing copy, social posts, taglines |
content for most SEO pages. Use creative only when you want the AI to take liberties with tone and structure.Prompt Template Syntax
Prompt templates support field placeholders that are resolved per-entity before the AI processes them:
| Syntax | Description | Example |
|---|---|---|
{{field.fieldName}} | Injects the value of a specific field from the entity | {{field.title}}, {{field.product_a}} |
Writing Effective Prompts
A good generation prompt has four parts:
- Task — What to write (e.g., "Write a detailed comparison between...")
- Requirements — Specific constraints (word count, structure, tone)
- SEO guidance — Keywords to target, search intent to match
- Output format — Which JSON fields to return
Here is a real prompt template from the Comparison Page bucket:
Write a detailed comparison between "{{field.product_a}}" and "{{field.product_b}}".
Requirements:
- Start with a brief overview of both products
- Compare across key dimensions: features, pricing, ease of use,
support, integrations
- Use a structured format with clear sections
- Provide an honest, balanced verdict
- Include a recommendation for different use cases
- Optimize for "{{field.product_a}} vs {{field.product_b}}" search queries
Return JSON with fields: title, comparison, meta_title,
meta_description, verdict, winner
Bulk Generation
Generate content for large numbers of entities efficiently:
Navigate to the entity type
Open the entity type you want to generate content for (e.g., Glossary Terms, Comparison Pages).
Select entities
Use the table's checkbox selection. You can select all visible entities, use filters to narrow down (e.g., only entities missing a meta_description), or pick specific rows.
Click Generate
Click the Generate button in the toolbar. Choose whether to generate all fields or only specific empty fields.
Monitor the job
A background job is created. Track progress in the Jobs panel — it shows how many items have been processed, how many succeeded, and any errors.
Review results
Once the job completes, go to the review page to approve or reject each generated item. Approved items become entities; rejected items are discarded.
Batch Sizes
Each bucket template defines an optimal batch size based on the content complexity:
Content Type | Default Batch Size | Typical Generation Time |
|---|---|---|
| Glossary Terms | 10 | ~30 seconds per batch |
| Blog Posts | 5 | ~60 seconds per batch |
| Comparison Pages | 5 | ~60 seconds per batch |
| Guides | 3 | ~90 seconds per batch |
| Case Studies | 3 | ~90 seconds per batch |
| Connection Pages | 5 | ~45 seconds per batch |
Knowledge-Aware Generation
Content generation automatically incorporates your project's knowledge base. This means the AI does not write in a vacuum — it has access to your brand context.
What the Knowledge Base Provides
Knowledge Type | How It Influences Generation |
|---|---|
| Brand voice | Consistent tone, vocabulary, and style across all generated content |
| Buyer personas | Content tailored to your target audience's pain points and language |
| Writing guidelines | Rules about formatting, word choice, prohibited phrases |
| Domain knowledge | Industry-specific facts, product details, competitive positioning |
| SEO guidelines | Keyword density targets, heading structure rules, internal linking preferences |
Enrichment vs. Creation
Content generation supports two modes of operation:
- Creation — Generates entirely new entities from a title or seed data. The AI fills in all fields from scratch.
- Enrichment — Takes existing entities with partial data and fills in the missing fields. The AI merges generated data with what already exists, preserving your manual edits.
Enrichment is useful when you have a spreadsheet of products with titles and prices, but need the AI to write descriptions, meta tags, and category assignments.
Auto-Approve vs. Review Queue
When an entity type has the status workflow disabled, generated content is auto-approved and immediately becomes a real entity. When the status workflow is enabled (the default for all bucket templates), content goes through the review queue.