Tags group your prompts so you can filter your visibility chart to a subset of them. Without tags you read one account-wide average; with tags you can see that Nike is strong on racing shoes and weak on trail.
That comparison is where most of the value in tagging sits.
Creating and applying tags
Create a tag and apply it to any number of prompts. You can tag during import, or select prompts on the prompts page and tag them in bulk afterwards — tagging existing prompts works fine, so there's no need to get your taxonomy right before you start.
Tags belong to a single brand. If you track more than one, each keeps its own set, so a tag named for one brand's product line never appears in another's.
They attach to the prompt itself rather than to a platform, so adding a new AI platform to a tagged prompt inherits the tags automatically.
Names are matched loosely
Tag names are matched on a simplified form, which means "Racing Shoes", "racing shoes" and "Racing Shoes" all resolve to the same tag rather than creating three near-duplicates.
Useful in practice: you don't have to remember exactly how you capitalised a tag last month, and a bulk import won't fragment your taxonomy over stray capitals.
Colours
Each new tag gets a colour automatically, cycling through a palette so a fresh brand ends up with visually distinct tags without any work. You can set a specific colour on the tags page if you want your groupings to match how you present them elsewhere.
Choosing a taxonomy
The tags worth having are the ones matching a decision you actually make. Three patterns that work:
Product line — racing, trail, everyday. Shows where you're strong by category, which usually maps onto who owns which content.
Funnel stage — research versus ready-to-buy. Being absent from "best running shoes for a first marathon" matters more than being absent from a broad definitional question, and tagging is how you separate the two.
Market — pairs naturally with country-targeted prompts, and answers "are we losing everywhere or just in one region?"
Resist tagging by anything you won't filter on. A tag per prompt is just a second copy of the prompt list.
Removing tags
Deleting a tag removes the grouping and leaves every prompt untouched — still tracked, still accumulating history. Tags are a lens over your prompts rather than a container holding them, so reorganising your taxonomy carries no risk to your data.
Where tags show up
Filtering the visibility chart is the main one. Once several tags have a few weeks of history, comparing them side by side is usually the fastest way to find where your effort is best spent — and the tag with the weakest numbers is a good candidate for reviewing your prompt phrasing.
