Traditional share of voice (SOV) is effectively obsolete, yet many organizations have replaced it with an equally flawed successor: AI share of voice.
Software vendors now claim to measure brand visibility across ChatGPT, Gemini, Claude, Perplexity, and other AI platforms using a single percentage score. The problem is that these metrics rely on a hidden denominator.
Unlike traditional search, where visibility could be measured against a known keyword set, the universe of possible AI prompts is effectively infinite.
Traditional SOV had limitations, but at least its assumptions were transparent. Marketers defined a fixed keyword set, tracked visibility against competitors, and used that list as a stable denominator. Everyone understood the measurement’s boundaries.
That model no longer exists. Search results are dynamic and personalized, and are increasingly being replaced by conversational interfaces. Yet many AI visibility platforms continue to present precise-looking percentages that can’t be audited or validated.
To stop presenting fictional metrics to leadership teams, we must rethink how we define and measure visibility in AI search.
Why traditional SOV metrics now fail
The basic assumptions of search engine optimization and digital brand tracking have been broken by two major shifts: the disappearance of the static results page and the rapid rise of personalized, conversational answers.
Search engines have become highly dynamic, personalized landscapes that change shape continuously based on real-time data.
Between AI-generated summaries, localized results, continuous scrolling, interactive merchant grids, and real-time social feeds, no two users will encounter the same interface, even when entering the exact same query at the exact same moment.
Because the search environment changes constantly, attempting to calculate a precise “share” of that screen has become a mathematical impossibility.
Your customers search everywhere. Make sure your brand shows up.
The SEO toolkit you know, plus the AI visibility data you need.
Start Free Trial
Get started with
The new volatile normality of rankings
Securing the top ranking position in the older marketing model meant capturing a highly predictable percentage of user click-through rates.
In the modern search landscape, however, ranking first organically might place a brand below several sponsored listings, an AI-generated overview, interactive question accordions, and featured discussions from community platforms.
Because search engines now construct layouts dynamically in response to immediate user intent and past search history, rankings fluctuate by the hour.
Measuring share of voice based on static positions is as unproductive as trying to measure the volume of an ocean wave with a wooden ruler.
The modern AI share of voice
When marketing teams realized that traditional rank tracking was losing its utility, software vendors quickly introduced alternative metrics, branded as LLM Visibility or AI share of voice.
These dashboards present highly polished, authoritative percentage scores that suggest a brand’s footprint has been successfully mapped across platforms like ChatGPT, Claude, Gemini, and Perplexity.
These tools fail to deliver on this promise, exposing a fundamental methodology problem that we must address directly.
| Legacy tracking (transparent) | LLM visibility (black box) |
| – Define fixed keyword list (known). – Measure rank on static SERPAuditable denominator. |
– Infinite possible user prompts. – Vendor runs small, arbitrary subset. – Subjective denominator. |
The infinite tail
Legacy SEO tools relied on a user-defined keyword list that served as a transparent denominator, whereas modern conversational engines present an entirely different mathematical reality where the universe of possible user prompts is effectively infinite.
Buyers no longer search for solutions using simple, two-word phrases. Instead, they enter highly specific, conversational queries that describe their exact organizational context, integration needs, and feature requirements.
Because no marketing tool can realistically sample this infinite universe of natural language, software vendors instead select a small, arbitrary subset of static prompts, run them through AI models behind the scenes, and aggregate those limited outputs into a representative global percentage.
This process creates a metric that only measures share of voice within a contrived and artificial environment, presenting a closed sandbox as if it were the open web.
Get the newsletter search marketers rely on.
The issue with black-box metrics
Marketers maintained full visibility into the data they were analyzing with legacy tracking tools, which meant that if a system reported a specific percentage of visibility, the underlying keyword list could be audited and adjusted. Modern LLM visibility tools obscure their denominator within proprietary, vendor-defined systems that are almost certainly incomplete.
This structural flaw became incredibly clear in September 2025, when OpenAI updated to its ChatGPT 5.0 model. Following this release, the platform-wide volume of outbound citations and source links dropped.
For marketing teams relying on LLM tracking dashboards, this model change resulted in a sudden, sharp decline in their reported visibility metrics. The decline had nothing to do with a loss of brand relevance or a failure in marketing strategy. ChatGPT had simply changed how it presented source data to users.
This update demonstrates that modern AI metrics are highly volatile and largely out of your control. While software vendors are genuinely trying to solve an incredibly complex engineering problem, the underlying methodology simply cannot support the high-confidence dashboards they deliver, meaning these metrics should be treated as directional signals rather than hard numbers.
Beyond AI share of voice: 3 metrics that matter more
We must shift our focus from measuring pure search volume to measuring how effectively a brand is integrated into the broader context of digital discussions.
As search queries morph into conversational discovery, a brand’s visibility is no longer defined by the keywords it owns, but by how deeply it is embedded in the conceptual models used by AI.
1. Share of mentions
AI models synthesize relationships between concepts rather than simply indexing pages, meaning a brand must exist within the model’s training data, fine-tuning datasets, or real-time retrieval sources to be surfaced at all.
Share of mentions tracks how frequently your brand name, products, or key executives are naturally included in the responses generated across the broader information ecosystem.
This metric shifts the operational focus from ranking positions to vocabulary inclusion, ensuring that a brand is recognized by the model even when it is not explicitly prompted for a vendor list.
To influence this metric, organizations must focus on securing organic mentions across high-trust forums, developer communities, and authoritative industry publications where AI models actively gather and update their information.
2. Share of recommendations
When buyers use conversational engines to make purchasing decisions, they regularly ask for direct comparisons, shortlists, and product recommendations to simplify their research process.
Share of recommendations measures how often your product or service is explicitly featured when a user asks an AI engine to act as an advisor on a specific business challenge.
This approach shifts our focus from raw traffic acquisition to winning the buyer’s consideration set, which is critical because conversational engines filter out the noise of the web to deliver a highly curated list of options.
If your product positioning is overly generic, the model will struggle to categorize your offering and will default to recommending competitors that have established a much clearer, highly documented use case.
3. Share of narrative
Merely securing a mention in an AI response is insufficient if the context of that mention portrays your brand poorly, as high visibility within a negative framework can quickly become a strategic liability.
Share of narrative measures the qualitative attributes, adjectives, and associations linked to your brand name in conversational outputs, allowing you to understand how your business is being framed.
| Narrative | What it tracks | The core strategic question |
| The “best” narrative | How often you are framed as the premium, gold-standard market leader. | Is the model positioning our brand as the most capable solution available? |
| The “popular” narrative | How often you are cited as the default, widely adopted industry standard. | Is the model identifying our brand as the most commonly used option? |
| The “budget” narrative | How often you are categorized as the cost-effective, value, or entry-level alternative. | Is the model framing our brand primarily as a low-cost, entry-level alternative? |
If an AI engine includes your brand frequently but consistently describes your product as a complex, legacy system, your high share of voice may actually be damaging your sales pipeline.
See the complete picture of your search visibility.
Track, optimize, and win in Google and AI search from one platform.
Start Free Trial
Get started with
Reframing your success metrics
Leadership teams require competitive benchmarks to evaluate market performance, meaning you cannot simply stop reporting on share of voice without offering a viable alternative.
Transitioning your executive reporting smoothly requires a structured, three-step plan.
Reframing the executive narrative involves educating your leadership team on the limitations of modern AI dashboards.
This means explaining the hidden denominator problem and demonstrating why treating these figures as absolute metrics introduces unnecessary risk.
Contributing authors are invited to create content for Search Engine Land and are chosen for their expertise and contribution to the search community. Our contributors work under the oversight of the editorial staff and contributions are checked for quality and relevance to our readers. Search Engine Land is owned by Semrush. Contributor was not asked to make any direct or indirect mentions of Semrush. The opinions they express are their own.
