Measures how often a brand appears in responses generated by a specific AI model for a given search intent.
Marketers or SEO specialists reading analytics dashboards and AI model output data.
01What it is and how it works
When an AI model like GPT‑4 processes a user query, it draws from its training data and any real‑time signals it has access to. Share of Model counts each time the model includes a brand name in its answer, then divides that count by the total number of answers it generated for the same query across a defined period. The result is a ratio that tells you the brand's visibility inside that model’s output stream.
It is the percentage of times a brand shows up in a model's answers.
02What to do about it
1. Pull the latest Share of Model data from your analytics dashboard. 2. Identify queries where your brand’s share is below 10 % but the intent is relevant. 3. Optimize the brand’s structured data (Schema.org) and on‑page content to align with those queries. 4. Run a small prompt‑engineering test: ask the model the same query with a brand‑focused context and see if the share improves. 5. Schedule a weekly review to track changes after any content updates.
03How it is measured or noticed
The metric is calculated by the product’s backend using logs of model completions. Each completion is tagged with the query and any brand mentions detected via string matching or entity extraction. The share is then displayed as a percentage in the “Brand Visibility” panel of the dashboard. You can also export the raw counts as CSV for deeper analysis.
04Common mistakes
- Counting brand mentions in user‑generated content instead of model output.
- Including generic terms that happen to contain the brand name (e.g., "apple" the fruit).
- Averaging shares across unrelated queries, which masks query‑specific gaps.
05Limits
Share of Model only applies to AI‑generated text, not to traditional web search results or image outputs. It does not capture brand sentiment—only presence. The metric can be confused with “share of voice” in paid search, but the two are calculated from different data sources.
06Worked example
"For the query ‘best running shoes’, our dashboard showed a Share of Model of 22 % for Brand X over the past week, while Brand Y had 5 %. After adding structured markup to Brand Y’s product pages, the next week’s share rose to 12 %.*"
Frequently asked questions
How is Share of Model different from Share of Voice?
It depends on the medium being measured. Share of Model looks at how often a brand appears in the text generated by a specific AI model for a given intent, while Share of Voice measures brand presence across traditional media channels such as search results, ads, and social posts.
Should we try to increase our Share of Model for every search intent?
Usually you focus on intents that matter most to your business goals. Boosting Share of Model on low‑value or irrelevant intents can waste resources and may not improve actual conversions.
How does the product calculate Share of Model?
It calculates the metric by analyzing backend logs of model completions. Each time the AI model returns a response, the system checks whether your brand is mentioned and aggregates those counts relative to total responses for the intent.
Does Share of Model include image outputs from the AI?
No, Share of Model only applies to AI‑generated text. Image results are measured by separate visual‑content metrics and are not part of this calculation.
What should I look for if my Share of Model drops suddenly?
You would notice a decline in the percentage of responses that mention your brand in the dashboard. This could indicate changes in the model’s training data, competitor activity, or a shift in user intent relevance, and you should investigate recent content updates or campaign adjustments.
How long after a marketing change will Share of Model reflect the impact?
It typically takes a few days to a week for the effect to appear, depending on how quickly the model incorporates new signals. You can monitor interim trends in the logs while waiting for the full impact to stabilize.
Can third‑party AI models be included in the Share of Model metric?
Currently the metric is limited to the models that are integrated with our product’s backend. Adding third‑party models would require a separate integration and is not supported out of the box.
Asked out loud
spoken, not typedThe same term in the words somebody uses speaking to an assistant rather than typing into a box — written from the situation, which is why each one carries the situation it came from.
Usually it means your Share of Model is low for that intent, which could be due to insufficient relevant content or recent algorithm updates. Check the dashboard for the exact percentage and consider boosting SEO or content signals that align with the launch keywords.
It depends on the current Share of Model value shown in the app. If the percentage is below your target threshold, you may need to adjust your content strategy or run targeted campaigns to improve visibility.
Usually the latest metrics are refreshed every 24 hours, so you should have the most recent Share of Model data by tomorrow morning. If you need a quick snapshot now, you can request a provisional report from the system.