Marketers should be cautious about using AI visibility scores to make budget or strategy decisions, according to new industry guidance.
The Interactive Advertising Bureau (IAB), a trade group representing the digital advertising industry, says many current tools provide directional signals rather than reliable business intelligence.
These tools measure how often brands and websites appear across platforms such as ChatGPT, Gemini, and Perplexity. But results can vary depending on the prompts tested, the platforms included, and how frequently each prompt is repeated.
That means a precise-looking “share of voice” score may suggest more certainty than the underlying data supports.
IAB’s framework evaluates four areas: whether a brand appears, how prominently it is presented, how accurately it is portrayed, and whether the response encourages further action.
The guidance says tests involving fewer than 50 queries should be considered exploratory. More reliable measurement requires larger prompt sets, repeated testing, multiple AI platforms, and enough data to distinguish genuine changes from normal variation.
Although the guidance comes from the US-based IAB, the measurement problem is global. Brands and agencies in every market are being asked to evaluate tools that use different methodologies and can produce conflicting results.
The framework is not yet a formal standard. It does not establish a universal minimum number of prompts or repetitions, nor does it define an acceptable level of uncertainty.
For now, marketers should ask AI visibility providers to disclose which prompts they track, how often tests are repeated, which platforms and markets they cover, and whether reported changes fall outside normal variation.
