Academy · AI Workforce

Content Tagging Agent: how content gets matched to products

The agent that turns raw tracked content into attributable content — a visual model matching what's actually shown and said against your Product Center catalog.

Social Media MarketerSocial Commerce Specialist3 min readUpdated Aug 21, 2026

Requires first

This is the agent that makes attribution possible at all — without it, content can be tracked and scored, but never actually connected to a specific product or creator for reporting purposes.

How this connects across Socialscale

Credits
Product Center
Content Tagging Agent
Revenue, Cost & ROAS
app.socialscale.ai/settings/ai-workforce

Content Tagging Agent

Reviews content and assigns the correct products and creator to each creative based on spoken mentions and context.

Credit Cost: 2 / match · Matches: 144 / 325

What it matches, and how

This runs on a visual model, not a keyword search — it looks at what's actually shown on screen and listens to what's said, then checks that against Product Center: specifically the product's own image and description. A match isn't just "the creator said the product name" — it's the agent recognizing the physical product in frame and confirming that against the reference image, the same way a person would compare a photo to the real thing. A match assigns the content to both a product and the creator who posted it.

This is why an unsynced — or thin — product is a real problem for this agent, not just a formality: it can only recognize what it has a reference image and description for. See Sync Your Product Catalog for why complete product data (not just a name) matters here specifically.

Pro tip

A product synced with a low-quality or generic stock photo will match less reliably than one with the actual product image used in creator content. If a specific product consistently under-matches, check the image on its Product/Detail page before assuming creators aren't mentioning it (see Reading a Product/Detail page).

When it isn't confident: Review

Not every match is clean. Content the agent isn't confident about doesn't get force-tagged — it's routed to Review in Content Library instead, for a human to confirm or correct (see Content Library: folders, review, and AI scores).

Pro tip

A large, growing Review queue is usually a Product Center problem, not a Content Tagging Agent problem — it typically means creators are mentioning products that aren't in your catalog yet. Fix the catalog gap before assuming the agent needs adjusting.

The two places this agent's work shows up

  • Products (n) on each content item in Content Library — how many catalog products it matched (see Reading a Content/Detail page for exactly what a match looks like once you open one)
  • Revenue, Cost, and ROAS at the content and product level — tagging is the prerequisite step; without it, Storefront Revenue and Content Revenue can't be attributed to anything specific (see Understanding Revenue, Cost, and ROAS)

Watch out

Ad Revenue and Ad Cost can still show up on untagged content, since those come from the Ad Agent matching to your ad accounts directly — but Storefront and Content Revenue specifically depend on this agent's product tag. A gap in one but not the other is a signal, not a bug.

Scope: all content, or mentions only

By default, tagging runs on everything a creator has tracked. If you'd rather limit what gets imported in the first place to only content that mentions your brand, that's a separate, earlier control — see Import Creator Content Only When They Mention You. Tagging still runs the same way on whatever clears that filter.

Workflow playbook

Track commerce attribution: from creator campaign to sale

Launch a creator campaign, manage the collaboration, and trace every sale back to the exact piece of content that drove it.

Social Commerce SpecialistMarketing Manager3 min