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

# WeCareRemote AI assistant overview

> The WeCareRemote AI assistant helps refugees and NGO staff through intelligent conversation, document retrieval, and multi-language support on demand.

The WeCareRemote AI assistant is your primary point of contact on the platform. Whether you are a refugee looking for services and support or an NGO admin managing resources, the assistant answers your questions, retrieves relevant documents, and guides you through the platform — all through a natural conversation interface powered by leading AI models.

The assistant is built to be approachable: you can write in your own words, in your own language, and at your own pace. It is not a search box — it remembers what you have asked before, can follow up on incomplete information, and surfaces relevant resources from your NGO's connected knowledge base when they exist.

## When to use the AI assistant

Reach for the AI assistant when you need:

* **Quick information** about services, processes, eligibility, or paperwork.
* **Help getting started** on the platform — for example, where to find a particular blog post or how to join a meeting.
* **Drafting help** for an email, application, or message you want to send.
* **A first answer** before deciding whether you need to talk to a human caseworker.

For sensitive matters or when you need a real conversation, the assistant can hand off to an NGO staff member via a video meeting.

## Key capabilities

<CardGroup cols={2}>
  <Card title="Multi-model AI" icon="brain">
    Choose from a range of leading AI providers — including OpenAI, Anthropic, Google, Groq, and more. You can configure a default model or switch per request to match your needs.
  </Card>

  <Card title="Conversation memory" icon="clock-rotate-left">
    The assistant retains the full context of your conversation within a session, so you never have to repeat yourself. Follow-up questions are understood naturally.
  </Card>

  <Card title="Document retrieval" icon="file-magnifying-glass">
    Ask questions about resources, policies, and support documents. The assistant searches connected knowledge bases and surfaces relevant information directly in the conversation.
  </Card>

  <Card title="Role-aware responses" icon="users">
    The assistant adapts its responses based on whether you are a refugee user or an NGO admin, surfacing the most relevant information and actions for your role.
  </Card>
</CardGroup>

## How to start a conversation

<Steps>
  <Step title="Log in to WeCareRemote">
    Sign in to your account at [wcr.is/home](https://wcr.is/home). Your role is automatically detected at login.
  </Step>

  <Step title="Open the AI assistant">
    Navigate to the **AI Assistant** section from the main menu. The chat interface loads immediately — no additional setup is required.
  </Step>

  <Step title="Type your message">
    Enter your question or request in the text field and press **Send**. The assistant responds within seconds, drawing on your conversation history and any connected documents.
  </Step>

  <Step title="Continue the conversation">
    Ask follow-up questions, request clarifications, or change topics. The assistant keeps track of everything said earlier in the session.
  </Step>
</Steps>

<Note>
  Your conversation history is saved after each session. When you return to the AI assistant, previous exchanges are available, so the assistant can provide context-aware responses based on your past interactions.
</Note>

## Tips for better results

* **Be specific.** "What housing support is available in Berlin for a family of four?" gets a more useful answer than "I need housing."
* **Share relevant context.** Mention your role, location, language, or NGO when it might affect the answer.
* **Ask follow-ups.** The assistant remembers the conversation — you do not need to repeat earlier details.
* **Ask in your own language.** Most supported models handle multilingual conversation well; you do not need to switch to English.
* **Tell the assistant when it is wrong.** If a response is off, say so. The assistant adjusts based on your feedback within the conversation.

## Limitations to be aware of

The AI assistant is a powerful tool, but it has limits:

* **It is not a lawyer, doctor, or caseworker.** Treat its responses as a starting point, not as final advice on legal, medical, or administrative matters.
* **It can make mistakes.** Always verify critical information — deadlines, eligibility rules, contact details — with an NGO staff member or an official source.
* **It only knows what is connected.** The assistant draws on its training data plus any knowledge base your NGO has connected. It does not have live access to external systems unless explicitly integrated.
* **Conversation history is stored.** Anything you write is retained per your organization's retention policy — see [Memory and history](/ai-assistant/memory-and-history).

## Next steps

<CardGroup cols={2}>
  <Card title="Configure the AI assistant" icon="gear" href="/ai-assistant/configuration">
    Set your preferred model, connect API keys, and configure authentication.
  </Card>

  <Card title="Supported AI models" icon="list" href="/ai-assistant/supported-models">
    Browse the full list of supported providers and models available on WeCareRemote.
  </Card>
</CardGroup>