AI Share of Voice: A New Metric for Brand Visibility in the Age of AI

The marketing landscape has welcomed a new metric: AI Share of Voice (AI SOV). While initially met with a sigh of added complexity for marketers, AI SOV offers valuable insights into a brand’s visibility within AI-generated responses compared to its competitors. However, its true utility hinges on a precise understanding of its calculation and interpretation. This comprehensive guide delves into what AI SOV is, how it’s computed, how to measure it effectively, and crucial caveats to avoid misinterpreting the data.

Understanding AI Share of Voice

AI Share of Voice quantifies a brand’s visibility within AI responses across a defined set of prompts and compared against a specific list of competitors. It essentially answers the question: "How often is my brand mentioned by AI when users are asking about topics relevant to my industry?"

The Calculation of AI Share of Voice

A fundamental method for calculating AI SOV involves:

AI SOV = (Responses that mention your brand / Responses that mention any tracked brand) * 100

For example, if three prompts related to email marketing software are analyzed on ChatGPT, and a brand appears in two out of three responses, while two competitors also appear in two responses each, a simple calculation would suggest each brand has a 67% AI SOV (2 mentions / 3 total responses).

However, this basic calculation treats all prompts as equal. In reality, the prominence and search volume of the underlying keyword for a prompt significantly impact its reach. A prompt like "best email marketing software," which is frequently asked, carries more weight than a more niche query.

To address this, advanced tools like Ahrefs’ Brand Radar incorporate an "AI-adjusted volume." This metric refines the calculation by factoring in the monthly Google search volume of the keyword associated with a prompt, multiplied by a platform ratio. This ratio reflects how much traffic each AI platform directs to websites relative to Google organic search, derived from Ahrefs’ traffic data. For instance, if a prompt’s keyword has a 20,000 monthly search volume, and ChatGPT’s platform ratio is 0.115, the AI-adjusted volume for that prompt on ChatGPT becomes 2,300.

When a brand is mentioned or its website is cited in an AI response, the AI-adjusted volume of that prompt contributes to the brand’s "estimated impressions." The AI SOV is then calculated as the brand’s total estimated impressions divided by the total estimated impressions of all responses mentioning any tracked brand.

Using the previous example, if a brand’s mentions stem from prompts with lower AI-adjusted volumes, its AI SOV will decrease. For instance, if a brand’s mentions only come from prompts totaling 3,335 AI-adjusted impressions, its share drops to 31% (impressions from those mentions / 3,335 total impressions). It’s crucial to note that "estimated impressions" represent a modeled estimate of exposure, not a definitive count of individuals who saw an answer.

Platform-Specific Considerations

Different AI platforms have distinct usage patterns and, consequently, unique platform ratios. For example, ChatGPT currently has a ratio of 0.115, Gemini 0.011, and Perplexity 0.01. This means a 20,000-search volume prompt generates 2,300 estimated impressions on ChatGPT but only 220 on Gemini.

AI SOV is calculated per platform, and when combined, the overall score becomes a weighted average, with platforms generating more impressions having a greater influence. A brand might achieve a 28% AI SOV on ChatGPT and 11% on Gemini. If ChatGPT accounts for ten times the impressions, the combined score might be around 26.5%, potentially masking the weaker performance on Gemini. Therefore, it’s essential to analyze each platform individually before examining the aggregated score.

The Importance of Reporting Scope

Every AI SOV calculation is contingent on specific choices: the prompts selected, the platforms monitored, the competitors included, the geographical location, and the time period. Altering any of these parameters will change the resulting percentage. This is why different tools can report varying AI SOV for the same brand; they are measuring different things. Always clearly define and report the scope of your measurement, including the brands, aliases, competitors, market niche, location, platforms, dataset, prompt group, tracking frequency, and date range.

When is AI Share of Voice Useful?

AI SOV is particularly valuable for answering strategic marketing questions, such as:

• Identifying visibility gaps: Understanding where a brand is underrepresented in AI responses.
• Assessing competitive positioning: Gauging a brand’s presence relative to key rivals.
• Evaluating content performance: Determining if content is being leveraged by AI.
• Informing content strategy: Pinpointing topics and sub-topics that require more attention.

For example, if a brand tracks 200 prompts related to "email marketing software" and finds it has 12% AI SOV overall, but a deeper dive reveals strong visibility in broad comparison prompts and weakness in specific "automation" queries, this provides actionable intelligence. The 12% is less meaningful than the insight that "automation" is a content gap requiring targeted efforts.

However, it’s crucial to remember that AI SOV measures visibility, not necessarily positive sentiment, actual recommendations, traffic generation, or conversion rates.

Measuring AI Share of Voice in Brand Radar

Brand Radar offers two primary methods for measuring AI visibility:

1. Broad Index Exploration: This provides immediate insights by searching a pre-collected corpus of AI responses from search-backed prompts. It matches your brand and competitor entities and calculates reports based on existing data. AI SOV in this view represents the brand’s share of estimated impressions within the filtered corpus, relative to tracked competitors.

2. Tracked Prompts: For more focused monitoring of commercially significant questions, users can define and track custom prompt sets. This involves adding one prompt per line, specifying locations, AI assistants, and update frequency.

Steps for Measurement:

• Add Brands and Competitors: Input your brand and its genuine name variations and associated websites. Group related entities for accurate tracking.
• Define Market and Platform Scope: Utilize report filters to specify the desired market niche, AI platforms (ChatGPT, Gemini, Perplexity, Copilot, Google AI Overviews, Google AI Mode), and location.
• Create Saved Reports: Save configurations for consistent future analysis. Mapped reports to Ahrefs projects ensure consistent tracking.
• Build Prompt Sets: For Tracked Prompts, either manually enter prompts or use AI generation tools as a starting point, focusing on commercially relevant questions.
• Understand the Metrics: Analyze mentions, citations, estimated impressions, and AI SOV side-by-side. Explore AI responses, cited domains, and pages for deeper context.
• Maintain Measurement Consistency: Avoid frequent changes to the measurement setup (prompts, competitors, platforms) to ensure like-for-like comparisons. Document any scope changes rigorously.

Interpreting AI Share of Voice Data

The true value of AI SOV lies in its comparative analysis:

• Compare Against Competitors: An 18% SOV is meaningless in isolation. When contrasted with competitors’ scores (e.g., Competitor A at 30%, Competitor B at 25%, Competitor C at 15%), it reveals a clear competitive landscape and highlights areas for improvement.
• Compare Against Yourself: Tracking your brand’s SOV over time, with consistent measurement parameters, reveals trends. A declining trend, when analyzed against consistent parameters, is more indicative of genuine visibility changes than a fluctuating score due to measurement alterations.
• Use Industry Benchmarks Cautiously: General industry benchmarks can be misleading if the scope and methodology differ significantly from your own measurement. Always question the context of any benchmark.

Understanding Fluctuations in AI SOV

A sudden increase in AI SOV from 20% to 27% requires careful investigation. Potential reasons include:

1. Genuine Visibility Change: If the measurement setup remained identical, this indicates a real improvement in brand presence within AI responses, possibly due to new content or competitor decline.
2. Measurement Change: Adding prompts, removing competitors, or altering platforms invalidates direct comparisons. Documenting scope changes is crucial for accurate interpretation.
3. AI System Changes: Updates to AI models, retrieval mechanisms, or inherent variability in response generation can cause fluctuations. Trust trends that persist over multiple reporting periods rather than isolated spikes.

Final Thoughts

As AI search technology matures, AI Share of Voice is poised to become an increasingly valuable metric. While improvements in data, sampling, and demand estimation are expected, the fundamental challenge of measurement scope will remain. The true advantage will not stem from obsessing over minor percentage point differences, but from establishing a trusted measurement setup, monitoring changes over time, and understanding the underlying reasons for those shifts. AI SOV, when used as a tool for understanding evolving brand positioning in AI responses, rather than just another score to chase, becomes a genuinely useful metric for marketers.