Apify 的 AI Brand Share of Voice 如何比較 ChatGPT、Perplexity 和 Gemini 對自家品牌與 3 個競爭對手的提及率（$0.80）

Disclosure: I built the Apify Actor in this post. Written with AI assistance.

In classic SEO you track rankings. In AI search the question is different: when a buyer asks an assistant "what are the best project management tools?", which brands does it name, in what order, and how often is yours one of them? That is your AI share of voice.

AI Brand Share of Voice on Apify runs that experiment for a whole category and returns a scoreboard. You type a category; it writes the buyer questions, asks Perplexity and Gemini (and ChatGPT), reads the brand names out of every answer, merges name variants, and ranks every brand. No brand list and no API key needed.

Your brand vs 3 competitors in 5 minutes

Open the Actor page. The form is prefilled with "project management software" and Asana, Trello and Monday.com.
Category: type yours as a noun phrase, for example "CRM software" or "running shoes".
Brands to always include: your brand plus three competitors. A brand nobody mentions stays in the table at 0%, which is often the finding.
Market (country): leave us or set gb, de, au and so on. Number of prompts: set 10. Samples per prompt and engine: keep the default of 2.
AI engines: keep all three (Perplexity, Gemini, ChatGPT). Click Start.

Want to see real output first? Tick Free sample run: 3 prompts, 1 sample, Perplexity only, no prompt fee.

How to read the table

You get one row per brand, best first:

mentionRate: the share of all analysed answers that mention the brand (0 to 1).
avgPosition: the average place in the answers that name it (1 means named first).
engineSplit: mention rate per engine, so you can see if you win on Gemini but vanish on Perplexity.
citationShare: the share of cited source URLs that sit on the brand's own website.

A final summary row gives the top brand and the five most-cited domains. The key-value store record OUTPUT holds the full report, and TOP_SOURCES the sources table alone. That table lists independent publishers and review sites next to the brands' own sites, so you can see where a mention can be earned.

What it costs

The price is $0.08 per prompt benchmarked. Samples and engines do not change it. Ten prompts = $0.80. The default run of 8 prompts is $0.64; a thorough 30-prompt run is $2.40. A prompt that no engine answers is not charged.

One setting to know: on the included key, ChatGPT answers 1 sample on the first 15 prompts of a run (3 prompts on a free Apify plan). Add your own OpenAI key and it runs on every prompt and sample.

Make it a habit

With 8 prompts and 2 samples you get a reliable first picture of who dominates the category. For a steadier ranking, raise the prompts to 20 to 30. Then put it on an Apify schedule, monthly for example, and compare the brand rows from run to run to see whether your mentionRate is rising.

Benchmark your category for $0.80, and send the scoreboard to whoever asked "are we showing up in AI?"