1. What the Tags Explorer shows
The Tags Explorer lets you analyze your AI visibility broken down by the topic tags assigned to your prompts.
Instead of looking at aggregate numbers, or having to click through all tag filters in your report one by one, you can see exactly in which categories your companies performs best.
2. Why knowing which topics drive AI visibility matters for your GEO strategy
Reports give you an aggregated overview of your AI visibility. That’s useful, but it can hide important nuances. Performance is not uniform across all topics. You might be highly visible in one category, while being completely absent in another that is just as relevant to your business.
Breaking your data down by categories or tags helps you uncover these differences. It shows where you are already strong and where you are underrepresented. AI models also build associations on a topical level. They don’t just recognize your brand globally, but connect it to specific themes, use cases, and problem spaces. This means you can be strongly associated with one topic, while not being considered at all in another closely related area.
This directly shapes your GEO strategy: double down on the areas where you already perform well, and focus your efforts on closing the gaps in categories where you should be visible but currently are not.
3. Measuring performance across categories
Every prompt in your report can be assigned one or more tags (for example: by product line, industry vertical, persona, competitor, or intent type).
You can open the Tags tab from the Explore section.
The Tags Explorer aggregates performance across those tags and surfaces four views:
Tag Visibility - How many responses and brand mentions each tag generated, plus your mention share.
Tag Citation Share - How many responses generated citations, and what share pointed to your domain.
Tag Combination Explorer (Mentions) - When two tags appear together in a prompt, does your mention rate go up or down versus each tag alone? This is the lift calculation for tag pairs.
Tag Combination Explorer (Citations) - Same logic, applied to citation rate.
All views are filter-aware and respond to your active model, language, country, and date filters.
4. How to add and manage tags on your prompts
Tags can be added during prompt creation or at any point afterwards.
Tags are assigned at the prompt level, not the response level. This means a tag describes the category of question being asked - and applies to every response that prompt has ever generated, including past ones.
If you add or change a tag on a prompt today, the updated tag will immediately apply to all historical responses for that prompt. Your tag-level analytics always reflect your current tagging, across the full response history.
Adding tags during prompt creation: When setting up a new prompt, you can assign one or more tags directly in the creation flow. For a full walkthrough of prompt setup, see the Prompt Overview article.
Adding or editing tags on existing prompts:
Go to the Prompts section of your report.
Open the prompt you want to tag.
Add or update tags in the prompt detail view.
The Tags Explorer will immediately reflect the updated tags across all historical responses for that prompt.
A few things to keep in mind when building your tag structure:
Tags work best when they reflect meaningful segments you actually want to compare - product lines, buyer personas, competitor names, use case categories, or intent types are all good starting points.
A prompt can carry multiple tags, which means it will appear in multiple tag rows and combination pairs in the explorer.
If the Tags Explorer shows empty states, it means your prompts do not yet have tags assigned. Start by tagging your highest-volume prompts first.
5. Explore Tag combinations.
Open the Tag Combination Explorer and sort by lift to find pairs of tags where combining two topics produces a disproportionately higher or lower mention rate than either tag alone. Use the search bar in the combination explorer to investigate specific tag pairings you care about.
AI prompts are naturally long tail. Users combine multiple intents, contexts, and constraints into a single query. By analyzing combinations of tags, you get a much more realistic view of how you perform in these niche scenarios. This helps you understand not just broad category performance, but how well you show up in the specific, high-intent situations that actually drive decisions.
6. How to turn tag insights into content and GEO actions
If a tag has high mention share but low citation share: AI models know your brand in this category but aren't citing your website. This means your content exists in training data but may lack the authority signals to get referenced in live search responses. Add more citable, source-worthy content on this topic.
If a tag has low mention share despite being core to your product: AI models don't associate you with this topic yet. Create dedicated content - comparison pages, guides, use case studies - that explicitly positions you in this space.
If a tag combination shows high lift: This is a signal that AI models respond strongly to prompts that combine these two topics. Build content that bridges them - and make sure your prompt dataset includes more of these pairs.
Add additional prompts for a category: If you want to prioritize a specific category, consider adding more prompts that cover this category.