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

# Coach

> Guide people as they use AI: a welcome, a first-use introduction, and a reminder when patterns need attention.

## What Is Coach?

Coach is Velatir's enablement agent. It does not decide which services are allowed or whether a prompt contains sensitive data. It speaks up at a few moments that matter, and stays quiet the rest of the time.

Create a Coach configuration, set it as your organisation default or assign it to specific workspaces, and tune what it says.

<Note>
  Coach requires **version 1.9.0** or later of the Velatir browser extension. Velatir for Desktop deploys the extension for you; if you manage the extension yourself, make sure that version has rolled out.
</Note>

## What It Does

Each section can be turned on or off independently.

<AccordionGroup>
  <Accordion title="Welcome message" icon="hand-heart">
    Shown the first time someone in the workspace reaches any AI service. Seen once per person.

    Use it for a short, organisation-wide note: treat output as a draft, keep sensitive information out of prompts unless the service is approved for it, and so on.
  </Accordion>

  <Accordion title="Service introductions" icon="book-open">
    A brief intro the first time someone uses a particular service, with optional **acknowledgement**. When acknowledgement is required, the person confirms they have read it before they continue.

    Leave acknowledgement off if you only want the intro to appear; that is the default.
  </Accordion>

  <Accordion title="Repeat reminders" icon="bell">
    A nudge after the same person has been stopped several times in the same category — for example, after Data Protector has blocked similar prompts. You set the threshold (from 3 onwards). This is guidance, not a warning.
  </Accordion>
</AccordionGroup>

You can edit the heading and body of each message. Built-in defaults are a starting point; change them to match how your organisation talks about AI.

## Coach Is Not an Observer or an Enforcer

[Gatekeeper](/agents/gatekeeper) and [Data Protector](/agents/data-protector) review every trace and run as Observer or Enforcer. Coach does not. It has no allow/block verdict on a prompt. It delivers messages in the browser at the moments above.

If you want a finding to also notify a channel, configure that on the security agents, not on Coach. See [Escalations](/traces/escalations).

## Language

Coach messages follow the language you set under **Settings → Agents**. You can keep one language for the organisation, follow each person's browser or OS language, or override per workspace. Edit the wording per language on the Coach configuration; a live preview shows the notification as it appears in the browser. See [Configuring agents](/agents/configuring-agents#localisation).

## AI Literacy

Coach is what produces [AI Literacy](/insights/ai-literacy) insights. The grade reflects how cleanly AI is used and, when you require acknowledgement on service introductions, how often those messages are acknowledged.

Until a Coach is assigned and people start seeing messages, the AI Literacy view stays empty.

***

<CardGroup cols={2}>
  <Card title="AI Literacy" icon="graduation-cap" href="/insights/ai-literacy">
    The grade Coach feeds, and how to read it.
  </Card>

  <Card title="Configuring agents" icon="sliders-horizontal" href="/agents/configuring-agents">
    Named agents, the organisation default, and workspace assignment.
  </Card>
</CardGroup>