> ## Documentation Index
> Fetch the complete documentation index at: https://docs.fullreach.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Get started

> From signup to the first results: the setup wizard step by step, what happens in the first hour, and what each day's collection does.

FullReach AI asks the AI platforms this brand's questions every day and keeps every answer. This guide goes from signup to the first results.

## Sign up

Open [fullreach.ai/start](https://fullreach.ai/start) and enter your company address, or continue with a Google or Microsoft work account. The trial needs no card. [Account and billing](/guides/account-and-billing) says more about signing in and the trial.

## The setup wizard

The wizard asks for what the daily collection needs. Each step proposes something, and you correct it.

<Steps>
  <Step title="Who are you tracking?">
    This brand, or brands you manage for clients. It decides only which website the wizard reads next. An agency adds each client as its own project later.
  </Step>

  <Step title="Add brand details">
    The website, the name and the location. The wizard reads the website to draft the next step.
  </Step>

  <Step title="Describe this brand">
    What it does, its industry, how it presents itself, what it sells, whether it sells online, and the other names it goes by. The prompts are written from this.
  </Step>

  <Step title="Personas">
    Who asks the questions, and each persona's share of the prompts. A persona can carry a sentence that goes in front of each of its prompts. See [Personas](/reference/personas).
  </Step>

  <Step title="Review competitors">
    Up to 5 competitors, found on the web. Keep, remove or add them. Each competitor needs its website. See [Competitors](/reference/competitors).
  </Step>

  <Step title="Choose a market and a language">
    The country the answers are collected in, or Worldwide, and the language the buyers ask in. More markets come later, with Copy to market. See [Markets and languages](/reference/markets-and-languages).
  </Step>

  <Step title="Review topics">
    Up to 10 subjects the buyers ask about. See [Topics](/reference/topics).
  </Step>

  <Step title="Review prompts">
    The questions, written for each topic and persona: up to 50, and up to 8 a topic. Edit, remove or add them. Each one runs every day on every platform. A question that asks for providers measures more than a question that asks how to do something.
  </Step>
</Steps>

After the prompts, the wizard asks a few of them live and shows a first look at the answers. That preview goes through the platforms' APIs, so its answers can differ from the daily collection's. Then you pick a plan for the trial, and the project starts.

## The first results

The first collection starts as soon as the project exists. The dashboard says "Your first results are on their way", with the prompts, platforms and markets it collects. It usually takes under an hour, and a mail tells you when the results are ready.

A few answers can still arrive later that day. Trials also get a recap mail after their first 3 days. See [Mail](/guides/mail).

The [Crawlability](/reference/crawlability) page works before the first results, because it reads the site, not the answers.

## Each day

Every day, in the early hours UTC, the collection runs for each project:

1. **Ask.** Each prompt goes to each platform through the platform's own app, in its market, as a logged-out visitor in a fresh session. See [How answers are collected](/reference/collection). A day is a UTC date.
2. **Ask again.** An answer that failed is asked again. Some tries go at once, and the rest later the same day.
3. **Analyse.** The analysis finds the brands in each answer and labels each mention. See [How answers are analysed](/reference/analysis).
4. **Read the sites.** The crawler audit reads each monitored site's robots.txt and homepage. Google Analytics and the crawler logs are read, when they are connected.
5. **Build the worklist.** The [Actions](/reference/actions) are built again from the stored answers, and their pages are read.

Every figure then counts the new day. The newest day can still change while its failed answers are asked again. [Perception](/reference/perception) asks its own questions once a week.

## Where to start reading

1. **[Overview](/reference/overview):** the four headline figures and the strongest competitors.
2. **[Prompts](/reference/prompts):** which questions name this brand, and which name nobody at all.
3. **[Actions](/reference/actions):** where to act, ranked, with the evidence.
4. **[Crawlability](/reference/crawlability):** anything on the site that blocks the AI crawlers.