Guide
How to track your AI visibility
AI visibility measures how often, and how favorably, AI answer engines mention your brand. Because these answers vary by prompt and model, you cannot eyeball it. This guide walks through building a repeatable measurement system: a prompt set, citation and sentiment logging, share-of-voice math, and tracking over time.
Last updated June 2026
- 1
Define a prompt set
List the questions your buyers actually ask AI engines, not the keywords you wish you ranked for. Include category queries ("best CRM for startups"), comparison queries ("X vs Y"), problem queries ("how to track AI citations"), and a few branded ones. Aim for 20 to 50 prompts that cover your funnel. Keep them stable so results stay comparable month to month, and store them in a sheet you can rerun.
- 2
Query each AI engine
Run every prompt through the engines your audience uses: ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews. Run each prompt a few times, since answers are non-deterministic and vary between sessions and accounts. Use a clean or logged-out profile so personalization does not skew results, and capture the full response text plus any linked sources for the next step.
- 3
Log citations, sources, and sentiment
For each answer, record three things: whether your brand was mentioned or cited at all, which source the engine pulled from (your site, a review roundup, a competitor, Reddit), and the sentiment of the mention (positive, neutral, negative). The source column matters most, it tells you which pages to strengthen and which third-party listicles you need to influence to earn more citations.
- 4
Compute share of voice vs competitors
Tally how many prompts mention each brand, then divide your mentions by total brand mentions across the set to get your share of voice. Do the same for direct competitors so you have a benchmark, not just an absolute number. Break it down by engine and by prompt category, since you may dominate AI Overviews while being invisible in Perplexity, which points you to where the gaps are.
- 5
Track the trend over time
A single snapshot is noise; the signal is the direction. Rerun the same prompt set on a fixed cadence, weekly or monthly, and chart citation rate, share of voice, and sentiment over time. Annotate the chart with what you shipped, new comparison pages, schema markup, a Reddit answer, so you can tie content changes to movement in the numbers and double down on what works.
- 6
Automate with a scanner
Manual querying gets old fast and is hard to keep consistent. An AI-visibility scanner runs your prompt set across engines on a schedule and logs citations, sources, sentiment, and share of voice automatically. Orphica includes a built-in AI-visibility scanner alongside competitor and mention tracking, so the logging and trend charts are handled for you. Whatever tool you pick, keep your prompt set fixed so the time series stays valid.
- 7
Iterate on the gaps
Turn findings into action. Where you are uncited, identify the source the engine trusts and earn a place there: improve the page it should cite, add structured data, or get listed in the roundups it quotes. Where sentiment is negative or wrong, publish the correcting facts on a crawlable page. Then rerun the scan to confirm the fix moved, closing the loop between measurement and AEO work.
Tips
- • Keep your prompt set frozen once you start. Adding or rewording prompts breaks comparability, so park new ideas in a separate "v2" set instead of editing the live one.
- • Track the source, not just the mention. Knowing an engine cites a third-party listicle instead of your site tells you exactly where to invest to earn the next citation.
- • Segment by engine. ChatGPT, Perplexity, and Gemini draw from different indexes, so a single blended score can hide that you are strong in one and absent from another.
Frequently asked questions
How do I track my AI visibility?+
Build a fixed set of 20 to 50 buyer questions, run them through ChatGPT, Perplexity, Gemini, and AI Overviews, and log whether each answer cites you, which source it used, and the sentiment. Convert mentions into a share-of-voice score against competitors, then rerun on a weekly or monthly cadence to track the trend. A dedicated AI-visibility scanner automates the querying and logging.
How is AI visibility different from SEO rankings?+
SEO measures your position in a ranked list of links for a query. AI visibility measures whether an answer engine mentions or cites you inside a synthesized response, where there is no ranked list and often only one or two sources are quoted. The metrics differ too: instead of position and clicks, you track citation rate, share of voice across answers, and sentiment of each mention.
How often should I run an AI visibility scan?+
Monthly is enough for most brands to see meaningful movement without chasing day-to-day noise from non-deterministic answers. Run weekly if you are actively shipping AEO content and want a tighter feedback loop, or after a major launch or PR moment. The key is a fixed cadence and an unchanged prompt set so each scan is comparable to the last.
Why do AI engines give different answers each time I check?+
Answer engines are non-deterministic and also personalize by account, location, and recent history, so the same prompt can return different brands and sources between runs. That is why you run each prompt several times from a clean profile and measure rates over many prompts rather than trusting any single response. Averaging across repeats and a fixed prompt set is what makes the trend reliable.
Related
See how you show up in AI search
Orphica is an AI-native CRM + Marketing OS with a built-in AI-visibility scanner — it tracks where your brand is cited across ChatGPT, Perplexity and AI Overviews.
Start free