📋 Available on: Growth (up to 3), Enterprise (up to 10+), Agency, and Custom plans. Not available on Startup or Free plans.
AI Filters (formerly Magic Fields) are AI-powered classifications that automatically organize your user-generated content from Instagram, TikTok, and YouTube. Each new post is analyzed and categorized based on rules configured for your workspace, so content can be filtered, reported on, and surfaced without manual tagging.
What AI Filters Can Do
AI Filters analyze visuals, captions, audio transcripts, and mentions in each post. Each piece of content is automatically categorized according to the criteria configured for your workspace.
Common classifications include:
Which products are featured
Sentiment (positive, negative, or neutral)
Whether competitors are mentioned
Suitability for paid ads
Specific themes or campaign elements
When new UGC arrives in your workspace, AI Filters run automatically. No manual action is required after setup.
AI Filters vs Super Search
Archive offers two AI tools for finding and organizing content. They serve different purposes:
Use AI Filters when you need:
Ongoing automatic organization of your content
Consistent tracking over time (e.g. product mentions, sentiment trends)
Structured data for reporting and analytics
Filters you'll reuse repeatedly
Use Super Search when you need:
One-time searches for specific content
Visual matching (find content similar to your top performers)
Text detection in images (logos, on-screen promotional text)
Immediate discovery without configuration
AI Filters automatically classify all content as it arrives. Super Search is on-demand discovery.
Setting Up AI Filters
AI Filters are configured by the Customer Success team during onboarding or when a new classification type is needed. Setup follows these steps:
Discovery call — discussion of what should be tracked and how content should be organized (product categories, campaign themes, brand priorities).
Configuration — the team builds the AI Filter, defines categories, and sets up classification rules.
Testing and deployment — the field is tested and deployed to your workspace. AI Filters can be applied retroactively to existing content on request — this is not automatic.
Ongoing use — after deployment, AI Filters run automatically on all new content. Values can be filtered and searched like any other tag in Archive.
Locked Option Lists
For AI Filters that use a fixed list of options — such as an official product catalog, approved campaign codes, or a strict set of labels — Archive can lock the field to predefined options only.
When locked:
Only the values configured during setup are available.
Manually entered values that don't match the list are rejected, instead of being added to the list.
The list stays strictly aligned with the source you defined (e.g. your product catalog).
When the field is not locked (the default), values typed manually in the dashboard that don't match an existing option are added to the list automatically. This is useful for fields where new categories are expected to appear over time, but it can cause unintended values to accumulate in the list.
To enable or change this setting on an existing AI Filter, contact your Customer Success Manager.
Common Uses
Product tracking — tag which products appear in each post to measure engagement and earned media value by product line.
Sentiment analysis — automatically score content as positive, negative, or neutral.
Ad quality scoring — identify UGC posts suitable for paid campaigns.
Campaign classification — tag content related to specific campaigns or themes for performance measurement.
Brand safety — flag content that may not align with brand guidelines or compliance requirements.
Viewing and Filtering AI Filter Results
AI Filter values appear in the UGC detail view — open any post to see its classifications in the side panel. The content library can also be filtered by AI Filter values to surface specific subsets.
You can also filter Reports directly by AI Filters — no longer requires routing content through a Collection or Campaign first. In Reports, open the Filters panel and pick any of your filterable AI Filters. The filter combines with every other Reports filter (Campaign, Collection, Post Type, Followers, etc.) so you can cross-cut performance by any combination. See Filtering Your Reports in Archive for the available operators and behavior notes.
AI Filters are also available as read-only filters in the Campaign Content tab. Open Add Filter inside a campaign's Content tab and pick any of your filterable AI Filters (for example, Sentiment, Brand Violations, or Links) to narrow the campaign's content. The filters mirror your Social Listening AI Filters — if a workspace has none configured, no AI Filter options appear in the dropdown.
A few things to know:
Reports only includes post-level AI Filters. Creator-level AI Filters (those classifying creators rather than UGC) remain available in Media Deck only.
Updates sync in seconds to about a minute. Re-tagging a UGC in Media Deck takes a moment to reflect in Reports filters.
Use thumbs-up / thumbs-down on classifications. Each AI Filter result in the UGC detail view has feedback buttons — flagging correct or incorrect results helps improve accuracy over time.
Credits
Each AI Filter run consumes credits from your workspace's monthly allowance:
Standard AI Filter: 1 credit per post, per field.
Pro Model AI Filter: 2 credits per post, per field.
Reprocessing: reprocessing a field counts as a new run and charges credits again.
Monthly credit allowances by plan:
Growth: 70,000 credits/month
Enterprise: 150,000 credits/month
Additional Credit Packs are available as add-ons. Contact your Customer Success Manager for details.
Getting Started
To set up AI Filters, contact your Customer Success Manager. Setup starts with a discovery conversation about what should be tracked and how the field should behave.