AI models recommend entities. When someone asks for “the best gaming laptop”, the answer usually names a specific product line — not just a company. If AI says “get the ASUS ROG Zephyrus” while your brand only appears as “Lenovo also makes gaming laptops”, you lost the recommendation slot even though you were technically mentioned.
Mention keywords let Ayzeo track this difference: they are the product lines, sub-brands, and alternative spellings that count as your brand. The Sub-Brand Portfolio panel then shows which of them AI actually names — and where it only speaks about you generically.
Setting up mention keywords
- Open your project’s Settings and find the Mention Keywords section.
- Add every name that should count as your brand: product lines (ThinkPad, Marketing Hub), sub-brands (Legion), and common alternative spellings of your brand name.
- Save. From the next analysis run on, every AI answer is scanned for these exact terms; the matched term and its surrounding text are stored per answer.
Good keywords are names, not descriptions. “Legion” and “Sales Hub” work; generic words like “pro” or “plus” alone will match unrelated text. Prefer the shortest form customers and journalists actually use.
Keywords also strengthen your core metrics: any keyword match counts as a brand mention, so a review that praises your product line without naming the parent company is still credited to you.
Reading the Sub-Brand Portfolio matrix
Once keywords are configured, the Sub-Brand Portfolio section appears on the AI Visibility Analytics page. Rows are your keywords; each cell shows how often AI named that keyword in the latest answers of the selected range, and across how many prompts. Darker cells mean more mentions. Click any cell to see the prompts behind it, each with the text snippet where the keyword appeared — and from there, the full AI answer.
- Columns — on Pro and Enterprise plans, the matrix splits by your prompt categories (tags), so you can see that a gaming sub-brand dominates gaming prompts but is absent from creator prompts. On other plans there is a single “All prompts” column.
- The “Generic (no keyword named)” row — the most important one. It counts answers that mention your brand without naming any specific product line. A large Generic share in a category means AI knows the company but not the product — and generic mentions rarely win recommendation slots.
Closing the generic-vs-specific gap
If a product line barely shows up while a competitor’s equivalent is named constantly, the fix is entity building — making the product line a recognizable name in the sources AI reads:
- Name the product line consistently on your own pages: titles, comparison pages, spec sheets — not just the parent brand.
- Check the AI Sources & Destinations panel for the review sites and communities AI consults in that category, and make sure the product line is present and correctly described there.
- Track the matrix over the following weeks — keyword mentions move faster than overall visibility because they respond to specific pages being updated.
Frequently Asked Questions
I don’t see the Sub-Brand Portfolio section. Why?
The section only renders when the project has at least one mention keyword configured. Add keywords in the project settings; the matrix fills with data from the next runs (and from any past answers where those terms already appeared).
Does the matrix show past data for a keyword I just added?
Partially. Matching happens when an answer is analyzed, so full coverage starts with the next run. Historical answers can be reprocessed — contact us if you need a backfill for a newly added keyword.
Why are category columns missing for me?
The per-category (tag) breakdown is included on Pro and Enterprise. Other plans see the same keyword analysis against a single “All prompts” column.
Track Your Product Lines
Add your sub-brands as mention keywords and see whether AI recommends them by name.
Open Your Project Settings