AI Visibility
AI Share of Voice
Enter your domain and one competitor. We run the same buyer prompts against the AI engines and show which of you gets named more often, prompt by prompt.
What share of voice means when the answer is a paragraph
In classic search, share of voice is a count of positions. Ten links exist and you hold two. An AI answer has no positions. It names three or four sources in prose and everything else on the topic is absent, so the measure that matters is how often you are in the answer at all, and who is there when you are not.
The comparison above builds a prompt set from the category you enter, sends each prompt to the connected engines once, and reads both domains out of the same answer. Neither side gets a different question, a different day, or a different engine mix, which is what makes the two percentages comparable.
The number to act on is not your score. It is any prompt where the rival is cited and you are not, because that row points at a specific page you can read and a specific claim you can beat.
Reading the prompt by prompt table
| Row | What it tells you | What to do |
|---|---|---|
| Both cited | The category has room for more than one answer on that question. | Hold it. Keep the page current and add the detail the answer left out, so you stay the fuller source. |
| Only you cited | You own the question today, usually because one page of yours states something nobody else states plainly. | Find that page, work out which passage got lifted, and reuse its shape on the prompts you lose. |
| Only the rival cited | They have a page, a mention, or a review the engines trust for that question. | Open the answer, follow it to the source, and match that claim on a page of your own with more specifics. |
| Neither cited | A third party owns the question, usually a roundup post or a forum thread. | The cheapest gap to close. You are competing with a generic listicle rather than a competitor's best work. |
The prompts are generated from the category you type, so the comparison is only as sharp as that input. A word like software produces vague prompts nobody asks. A phrase like AI SEO tool for marketing agencies produces prompts a buyer would recognise.
Share of voice questions
My score is zero and my competitor is at sixty. Where do I start?
Read the answers before you touch your site. In most categories the cited sources are review sites, roundups, and forum threads rather than the rival's own pages, which means the gap is coverage rather than copy. Getting named accurately in the three or four sources the engines keep quoting moves the number faster than another rewrite of your homepage.
Can I compare against more than one competitor?
One per run here. Run it again with a different rival to build the picture, keeping the category wording identical so the prompt set stays the same between runs. Tracking a full competitive set continuously is what the paid Competitor Agent does.
Does a competitor's ad budget affect their share of voice?
Not directly. AI answers are assembled from organic sources and there is no paid placement inside them. Advertising still shows up indirectly, because a brand more people search for gets written about more, and those articles are what the engines read.
How often should I re-run the comparison?
Monthly, or after you publish something aimed at a prompt you lost. Two runs a day apart mostly measure engine noise. On a five prompt set, one answer flipping is a twenty point swing, so treat anything smaller than a full prompt as unchanged.
Previews the Competitor Agent
This compares two sites on a handful of prompts. Rankfastt watches up to ten rivals across every tracked prompt and flags exactly what their page has that yours does not.