Quickstart

A first run takes about ten minutes of your attention and a few minutes of ours. Everything below happens inside the app; nothing needs to be installed on your site.

1. Create a target A target is one brand: a name, its website URL, and the markets you want to be measured in. The URL is not decoration — it is what tells the analysis model which company you are when another business shares your name.
2. Build a prompt library Prompts are the questions your customers ask an assistant. Generate a set with AI, import your own, or pull them out of Discover. Aim for a mix across the four intents.
3. Select engines and markets Pick the assistants that matter for your audience, and the countries you sell into. The app shows the exact number of probes the selection will consume before you commit.
4. Run the probe Each prompt is sent to each engine in each selected market, and every answer is graded independently. The run continues in the background — you can leave the page.
5. Read the panel Final Score, AI Score, Mention Rate and Sentiment at the top; competitors, cited sources, individual probes and the region comparison in the tabs below.
6. Run a Site Audit Visibility has a floor set by your own site: if AI crawlers can't reach or read it, no amount of prompting will fix the score. The audit tells you which of those is true.

1. Create a target

Dashboard → Create Target. You need three things:

Field	What it is	Why it matters
Target name	Your brand as customers say it	The string the analysis model looks for in every answer
Website URL	Your homepage	Disambiguates same-named companies, and is the page the Site Audit fetches
Regions	The markets to monitor	The first country is the primary market; changeable later in Settings

Brand information is optional but worth filling in: a sentence or two on what you sell and who for gives the analysis and recommendation models real context instead of guesses.

New targets start in the US with a single market. Adding markets multiplies how many probes a run consumes — see Plans & usage.

2. Build a prompt library

Prompts tab. Three ways to fill it, and they combine:

• Generate with AI — choose how many prompts to produce per intent. The generator is told your brand and industry, and writes questions the way a shopper would.
• Batch import — paste your own, one per line. Duplicates are detected, and each imported prompt is classified into an intent automatically.
• Discover — website analysis, Google Trends and Reddit threads, each turned into a probe-ready prompt you can add with one click.

A useful first library is roughly 60% discovery prompts (category and scenario, which never name your brand) and 40% branded prompts (comparison and trust, which do). Discovery measures whether AI brings you up on its own; branded measures how you’re judged once you’re in the conversation. See Prompt intents.

3. Select engines and markets

Select the prompts to run, then Review & Run Probe. The dialog asks for engines and markets and shows the total operation count before you start.

The count is not prompts × engines × regions. Engines that accept a country parameter run once per market; engines that don’t run once in the primary market only:

probes = prompts × Σ_engine (region-aware ? regions : 1) Claude and Doubao are region-agnostic, so they cost one probe no matter how many markets you monitor.

4. Run the probe

Runs execute in the background. A run is one probe group: all the probes launched together, one row in your score history, one set of stats.

While it runs, each probe follows the same path — query the engine, capture the answer and its citations, grade the answer against the rubric for that prompt’s intent, extract competitors and sources. Failed probes are retried up to three times; anything still unresolved is flagged and excluded from the averages rather than counted as a zero.

5. Read the panel

Visibility tab. Four cards at the top:

Card	Reads as	The short version
Final Score	0–100	0.8 × average AI Score + 0.2 × SEO Score
AI Score	0–100	Mean of every probe’s AI Score in the current scope
Mention Rate	%	Share of discovery conversations that mention you
Sentiment	0–100	Tone, averaged over answers that actually mention you

Below them: score history, AI-generated interpretation and prioritised actions, and the Overview / Competitors / Sources / Probes / Region Comparison tabs. The full tour is in Reading the visibility panel.

Read one probe end to end

Open the Probes tab and click any row. You get the exact prompt, the full answer, the score breakdown, the competitors named in that answer and every source it cited. One probe explains more about a score than any average does.

6. Run a Site Audit

Site Audit tab. Two halves, run together:

• Structural checks — robots.txt and AI crawler access, sitemap, meta tags, headings, structured data, and whether your content is readable without JavaScript. GPTBot, ClaudeBot and PerplexityBot don’t execute JavaScript; a client-rendered page can look empty to them.
• Lighthouse — Performance, Accessibility, Best Practices and SEO from PageSpeed Insights. The mean of those four is the SEO Score that contributes 20% of your Final Score.

Both feed one ranked Priority fixes list.

Where to go next

Core concepts The vocabulary the rest of the docs assume. Scoring methodology How one answer becomes a number, and how numbers become a panel. Prompts guide Building a library that measures something worth knowing.