Asset Label Check Agent · Use case
How can we use an AI agent to list which campaign assets were made with AI and whether each is labelled on our own infrastructure?
An asset label check agent takes one campaign and lists its images, audio and video from your asset library: for each, whether the library records that it was made or changed with AI, whether a label is recorded, and whether usage rights are on record and in date. It lists first the assets for which the library says nothing. The brand lead decides labels and clears rights. The agent makes and changes no asset, inspects no file to guess how it was made, writes no label, and does not say which assets the law asks to be labelled.
The campaign has forty assets.
Nobody has listed which were made with AI.
- ✗ 01A campaign draws images, audio and video from agencies, stock libraries, colleagues and tools, and how each was made is known to whoever made it and seldom written down in one place.
- ✗ 02In some cases the law asks that people are told when content is AI-made, and a company cannot label what it has not listed.
Where it stops for a person
Always. The brand lead decides labels and clears rights.Always. It lists what is on record, it does not detect and it does not decide: an asset the library does not mark is shown as "not recorded", never as "not made with AI". It does not say which assets the law asks to be labelled, or what a label should say: that is for the brand lead and your counsel. Rights are shown as the record gives them, and a person clears each one. The agents in this pack make text only: none generates an image, audio or video.
It lists what the library records.
The brand lead decides the labels.
Step · 01
Your asset library records, for each asset, how it was made, any label and the usage rights, in fields you choose.
Step · 02
The brand lead or the campaign owner names the campaign.
Step · 03
The agent reads the campaign's assets and their records through your own APIs, read only.
Step · 04
It lists each asset: made or changed with AI as the library records it, label recorded or not, rights on record and their end date. The assets with an empty field come first.
Step · 05
The brand lead decides the label for each, clears rights, and has the missing records completed.
Reads from · connectors
Your own APIs
The assets attached to the one campaign named and the fields your asset library holds for each, read through your own APIs with the operations that change something left out. It reads the record, not the picture: it does not analyse an image, a voice or a face.
Acts through · MCP servers
Nothing. It writes a list. It has no tool that edits an asset or its record, adds a label or a marker to a file, or publishes anything.
What you set up
- Add fields to your asset library for how an asset was made, its label and its usage rights, and decide who fills them in.
- Import the description of the API that holds campaigns and assets, and keep only the operations that read.
- Build the agent in Katonic Studio and test it on an invented campaign, with some records left empty.
- Publish it to the brand team first.
Where it runs
The agent, its knowledge store and the model run on infrastructure you own or control. You choose the model and serve it yourself, so questions, documents and answers are not sent to a public AI service. Asset records and the list stay inside the company.
How you get it today
This page is labelled Use case: the agent is built in Katonic Studio. You describe the job, connect what is named above, add test questions and release the agent once it passes them. When the Catalog holds a tested template for this job, this page will carry the Template label and show how to take it.
Does it detect whether an image was made with AI?
No. It lists what your asset library records. An asset with no record is shown as not recorded.
Does it add the label?
No. It has no tool that changes an asset or its record. The brand lead decides the label and a person applies it.
Does it say which assets must be labelled?
No. It lists what was made or changed with AI and what is labelled. What the law asks in each case is for the brand lead and your counsel.