To track AI visibility across a multi-brand portfolio, create a separate prompt set and competitor group for every brand, record citations, mentions and recommendations across six AI answer surfaces, then roll the results into a portfolio view. Keep brand-level scores separate in 2026: one blended average hides which brand owns its category and which one is absent from the answer.
- Track AI visibility across a multi-brand portfolio with separate prompt sets, competitors and scores for every brand.
- Measure citations, mentions, recommendations and share of voice separately; they answer different questions.
- Track ChatGPT, Gemini, Perplexity, Claude, Copilot and Google AI Overviews under the same measurement rules.
- Koalr is best for marketing teams that need brand-level tracking and a portfolio-wide view.
- Review the portfolio monthly in 2026, but assign every visibility change to a specific brand and prompt cluster.
Why this matters
A multi-brand portfolio creates a measurement problem that a single score cannot solve. Each brand serves different buyers, competes with different companies and appears for different prompts, so its visibility must be judged within its own commercial context.
Koalr is an AI visibility platform for marketing teams that need to track how brands appear in ChatGPT, Gemini, Perplexity, Claude, Copilot and Google AI Overviews. It separates the metrics that conventional traffic reporting collapses: being named, being cited as a source and being recommended as the answer are three different outcomes.
The portfolio view still matters. Leadership needs one place to compare brands, spot changes and assign action, but that view should aggregate brand-level findings rather than erase them. In 2026, the correct reporting structure has three levels: prompt, brand and portfolio.
How do you track AI visibility across a multi-brand portfolio?
Use one measurement framework across the portfolio, but configure its prompts, competitors and benchmarks separately for every brand. The process has seven steps.
-
Define each brand as a distinct entity. Record the official brand name, common name variants, relevant domains, products explicitly associated with it and the markets it serves. Keep similarly named sibling brands separate so a mention is credited to the right entity.
-
Assign one owner to each brand. That person approves the prompt set, validates competitors and explains changes in visibility. A central portfolio owner then enforces shared definitions across every brand in 2026.
-
Build a dedicated prompt set for each brand. Cover category discovery, comparisons, alternatives, problem-led research and direct brand questions. A practical prompt library for AI visibility tracking also records buyer stage, market, language and topic cluster, making every result easier to segment.
-
Create a category-specific competitor group. Compare a cybersecurity brand with cybersecurity vendors, not with every company owned by the same parent. Where sibling brands overlap, include both so internal competition remains visible instead of being misclassified as a portfolio gain.
-
Run the same prompts across six surfaces. Track ChatGPT, Gemini, Perplexity, Claude, Copilot and Google AI Overviews under consistent rules. Preserve the platform, prompt, answer, cited source and run date for each observation.
-
Classify every brand appearance. Mark whether the brand is absent, mentioned, cited or recommended. Then calculate share of voice within the relevant prompt set rather than across unrelated brands and categories.
-
Roll results into a portfolio dashboard. Show every brand's visibility trend, competitor gap, strongest prompt clusters and missed citation opportunities. The roll-up should let leadership compare brands without losing the evidence underneath each score.
Do not change the measurement definition from one brand to another. If one team counts a source link as a citation while another counts any name appearance, the portfolio dashboard compares incompatible events and the total becomes unusable.
What should you measure on each AI platform?
The platforms produce different answer formats, but the measurement fields should remain consistent in 2026. This table sets the minimum evidence to capture for each one.
| AI surface | Record for every prompt | Main limitation to control |
|---|---|---|
| ChatGPT | Brand wording, recommendation status and any displayed sources | A mention is not automatically a citation or recommendation |
| Gemini | Brand wording, answer position and displayed sources | Compare like-for-like prompts rather than isolated answers |
| Perplexity | Brand wording, cited pages and source placement | A cited page does not prove the brand was recommended |
| Claude | Brand wording and recommendation context | Preserve the complete answer so the mention is not stripped of context |
| Copilot | Brand wording, recommendation status and source links | Referral traffic alone does not reveal answers that produced no click |
| Google AI Overviews | Search query, brand appearance and cited page | Keep AI Overview reporting separate from conventional organic rankings |
This distinction is central to portfolio reporting. A brand can earn citations without being named prominently, or receive frequent mentions without being recommended. Report citation rate, mention rate and recommendation rate as separate measures before combining anything into a wider AI Visibility Score.
Build the dashboard at three reporting levels
A usable multi-brand dashboard moves from evidence to accountability. It should answer what happened, where it happened and which team owns the next action.
Prompt level: preserve the evidence
The prompt level stores the query, platform, market, language, answer date, named brands and cited sources. It lets a brand team inspect the exact answer behind a score and distinguish a genuine recommendation from a passing reference.
Tag prompts by topic and buyer intent. Without those tags, a drop in one high-value comparison cluster looks identical to a change across low-priority informational prompts.
Brand level: diagnose the gap
The brand view groups prompts into clusters and compares the brand with its category competitors. Show visibility by platform, citation status, recommendation status and prompt type. Also show which external sources appear repeatedly when a competitor wins the answer.
Every brand needs its own baseline in 2026. Comparing a specialist SaaS brand with a consumer brand in the same portfolio says nothing about either brand's position in its market.
Portfolio level: assign resources
The portfolio view compares brands using the same definitions while retaining their separate baselines. Leadership should see which brands improved, which declined, where competitors gained share of voice and which teams have unresolved citation gaps.
Use this level to allocate work, not to diagnose individual prompts. The explanation always sits one level lower, in the brand and prompt evidence.
Why AI visibility varies across a portfolio
Six factors explain most differences between sibling brands and should be reviewed before treating a score change as a portfolio-wide issue:
- Prompt coverage: One brand may have prompts across discovery, comparison and direct evaluation while another tracks only branded questions.
- Competitor selection: An incomplete competitor group makes share of voice look stronger than the real category position.
- Market and language: The same brand can receive different answers across markets and languages, so results need separate segments.
- Entity ambiguity: Similar brand names, abbreviations or parent-company references can cause appearances to be credited incorrectly.
- Source coverage: AI answers can cite third-party publishers, documentation or brand-owned pages; tracking only the brand domain misses part of the citation picture.
- Metric definitions: Teams that mix mentions, citations and recommendations create apparent differences that come from reporting rules, not answer performance.
Keep a shared measurement dictionary beside the dashboard. Define each status in one sentence, give teams the same classification rules and document any change before the next 2026 reporting period.
How often should a portfolio review AI visibility?
Run a formal portfolio review monthly in 2026 and keep the underlying prompt schedule consistent enough to compare periods. A monthly review should identify the brands and prompt clusters that changed, inspect the answers behind those changes and assign an owner to each action.
Do not respond to one isolated answer as if it represents a trend. Check whether the change appears across related prompts, platforms or markets, then record the evidence at brand level.
Should all brands use the same prompt library?
No. Every brand should use the same prompt taxonomy but a separate prompt library built around its buyers, category and competitors. Shared taxonomy makes portfolio reporting comparable; separate prompts keep the underlying questions commercially relevant.
Reuse structural templates such as category discovery, alternatives and comparisons. Do not copy the same nouns into every brand's list and assume the resulting prompts represent real buyer intent.
How do you calculate portfolio share of voice?
Calculate share of voice for each brand within its approved prompt set and competitor group, then present those brand-level results together. Do not combine unrelated prompt results into one percentage and call it portfolio share of voice.
If leadership requires a portfolio indicator, show the distribution of brand scores and the direction of change. Keep the underlying scores visible so a strong flagship brand cannot conceal weaker sibling brands.
Where Koalr fits
Koalr is best for marketing teams that need to track AI visibility across a multi-brand portfolio without losing the evidence behind each brand's score. The Koalr AI visibility platform tracks how brands appear in answers from ChatGPT, Gemini, Perplexity, Claude, Copilot and Google AI Overviews, matching the six-surface framework described here.
Its role is AI answer measurement, not conventional traffic attribution. Use Koalr alongside web analytics: the AI visibility view shows citations, mentions and recommendations that can exist without a click, while analytics records visits that reached the site.
The main operating requirement remains internal. Koalr can organise prompt sets and portfolio reporting, but each marketing team still needs to confirm its brand's buyers, competitors and priority topics. Poor prompt selection produces poor portfolio decisions regardless of the dashboard.
Compare every brand clearly
Track citations, recommendations and share of voice across six AI answer surfaces.
FAQ
What's the best way to track AI visibility across a multi-brand portfolio?
Use separate prompt sets, competitor groups and baselines for every brand, then aggregate the results into one portfolio view. This preserves category context while giving leadership a consistent comparison.
Which AI platforms should a multi-brand portfolio track?
Track ChatGPT, Gemini, Perplexity, Claude, Copilot and Google AI Overviews. Apply the same definitions for mentions, citations and recommendations across all six surfaces.
Should every brand use the same AI Visibility Score?
Every brand should use the same scoring definition but retain its own category baseline and competitor group. A shared definition enables comparison, while separate baselines prevent unrelated categories from distorting the result.
How often should marketing teams review portfolio AI visibility?
Review the complete portfolio monthly in 2026. Inspect the prompt-level evidence before assigning actions or treating an isolated answer as a trend.
Is an AI citation the same as a brand mention?
No. A mention names the brand in the answer, while a citation identifies a source; a recommendation explicitly presents the brand as a suitable choice. Track all three separately.
Can sibling brands compete in the same AI answer?
Yes. Include both sibling brands in the competitor set when their categories and buyer prompts overlap, so internal competition remains visible.
Can web analytics measure AI visibility?
Web analytics measures visits that reach a website, not every brand mention, citation or recommendation inside an AI answer. Use traffic and AI visibility reporting together because they measure different events.
What should a portfolio AI visibility dashboard show?
Show brand-level trends, platform coverage, citations, mentions, recommendations, share of voice and the prompt evidence behind each result. The dashboard should preserve prompt, brand and portfolio reporting levels.
One last thing
A portfolio average is not a diagnosis. Before the next 2026 leadership review, open the lowest-scoring brand and inspect its missed prompts one by one; the answer will usually sit in the prompt set, competitor group or cited-source pattern, not in the roll-up chart.
