Use Cases

Here are some practical ways people are using AI with Directus to speed up their content workflows.

Content Creation

Import from Google Docs

Instead of copying and pasting blog posts from Google Docs (and losing all your formatting), just tell the AI:

"Import this blog post from Google Docs into my articles collection"

The AI will create a properly structured entry with the right field types, categories, and author relationships.

Update Content in Batches

Need to publish 20 draft posts at once? Or add tags to a bunch of articles?

"Set all blog posts tagged 'product updates' to published status"

"Add alt text to all product photos that don't have any"

Much faster than clicking through each item individually.

Content from External Sources

Got a press release or competitor analysis you want to turn into structured content?

"Turn this press release into a news item in my press collection, and create related company and person entries"

The AI understands your schema and creates everything with proper relationships.

Asset Management

Fix Your Image Library

We've all been there - hundreds of images named IMG_2847.jpg with no descriptions. The AI can analyze and organize them:

"Go through my product photos and give them proper names, descriptions, and organize them into folders"

Before: IMG_2847.jpgAfter: red-leather-handbag-gold-hardware.jpg with proper alt text and organized into /products/handbags/

Bulk Metadata Updates

"Add alt text to all images that don't have any"

"Create thumbnails and organize all uploaded PDFs from this month"

Schema Changes

Add Fields and Populate Them

When you realize you need a new field across existing content:

"Add a 'featured_image' field to my blog posts and populate it with the first image from each post's content"

"Create a 'reading_time' field and calculate it for all existing articles"

The AI handles both the schema change and the data migration.

Set Up Relationships

Complex relationships are annoying to set up manually:

"Create a tagging system for my products - I need a tags collection and a many-to-many relationship"

"Set up categories for my blog with a parent-child hierarchy"

Real Workflow Examples

E-commerce Product Import

The old way: Download CSV, manually map fields, fix formatting issues, upload images separately, create relationships one by one.

With AI:

"Import these 50 products from this CSV, create the product entries, upload and link the images from the provided URLs, and set up categories"

Event Management

The old way: Create event, manually add each speaker, create session entries, link everything together.

With AI:

"Create an event called 'Tech Conference 2024' with these 10 speakers and their sessions from this agenda"

Content Migration

The old way: Export from old CMS, massage data, import piece by piece, recreate relationships.

With AI:

"Migrate all these blog posts from WordPress, keeping the categories and author relationships intact"

Automation with Flows

You can trigger Directus Flows with AI:

"Run the content audit flow on all blog posts from last month"

"Trigger the email sequence for all users who signed up this week"

This is great for one-off maintenance tasks or testing automation.

Tips for Success

Start simple: Try basic operations like publishing posts or updating metadata before complex schema changes.

Use staging: Test workflows on development data first, especially for bulk operations.

Review everything: The AI is good but not perfect. Always check important changes.

Permissions matter: The AI can only do what your user account can do, which is actually a good safety feature.

Keep backups: Before big bulk operations, make sure you have recent backups.

Common Gotchas

File uploads: The AI can't upload files directly from your computer, but it can work with URLs or files already in Directus.

Complex validation: If you have strict validation rules, the AI might create entries that don't pass validation.

Rate limits: For huge bulk operations, you might hit API rate limits.

Relationships: The AI needs existing entries to create relationships - it can't link to entries that don't exist yet.

The key is treating AI as a smart assistant, not a magic wand. It's really good at the tedious, repetitive stuff that normally takes forever to do manually.