AI search uses a generated response to help answer a question. The response may include explanations, recommendations, and links to sources. For a business, the relevant questions are whether the answer mentions it, represents it accurately, and helps a potential customer make a decision.
The cited 2025 usage and browsing studies document substantial use of AI assistants and changes in search-result clicking. They measure different populations and activities. Neither study alone establishes the share of your customers who use AI search.The NBER working paper "How People Use ChatGPT", written with OpenAI's own usage data, documents adoption from launch through July 2025.Pew Research Center measured actual click behavior from panelists' browsing data, not survey answers.
Referral visits are one part of the measurement. Crawl-to-referral data describes requests relative to attributed visits; it does not establish that clicks will disappear or that every answer exposure has commercial value.Cloudflare Radar's crawl-to-refer ratios, measured across sites on Cloudflare's network. The ratios move week to week; the imbalance does not.
What an AI answer is made of
An answer depends on learned model parameters and the context supplied to the request. That context can include the prompt, conversation history, uploaded documents, and retrieved sources. Model Memory and Live Retrieval compares answers with and without external retrieval.
Publishing current information can make it available for retrieval. It does not directly edit a deployed model’s parameters. Check the available sources and test conditions before deciding which work could address an error or omission.
The outcomes that count
The following outcomes are useful to distinguish:
- Mention: your brand's name appears anywhere in the answer.
- Recommendation: the answer presents your brand as a pick, not a passing reference.
- Citation: the answer links or credits one of your pages as a source.
- Prominence: where and how noticeably the brand appears. Record the scoring rule rather than assuming a fixed commercial value for each position.
- Accuracy: whether what the answer says about you is true and current. A model can recommend you while quoting two-year-old pricing.
AI search visibility describes the observed outcome. AI search optimization describes the work intended to improve it. This guide includes SEO and paid placement in that broad scope, while measuring paid and organic results separately.
Measuring repeated answers
A generated list can vary across repeated requests. Record position when it is useful, but do not treat one answer as a permanent ranking. Repeat prompts and report the frequency of each outcome.
For example, measure the percentage of completed answers that recommend your business for a defined prompt set. State the product, date, settings, and sample size. Measuring AI Search Visibility defines the metrics.
Which businesses appear most often
A repeated-answer sample can reveal whether a few brands account for most of the recommendations. That concentration describes the selected questions and conditions. It is not a measure of company market share.
Start by counting which brands appear across completed answers. Report the prompt group, product, period, and denominator. Compare customer requirements separately before making an all-category average.
Where to go from here
Continue with AI Search Products and Their Sources, AI Search Terminology, or SEO and AI Search Optimization.