> For the complete documentation index, see [llms.txt](https://docs.userlane.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.userlane.com/assistance/the-assistant.md).

# The Assistant

The in-app AI helper: what it answers, how it uses connected knowledge, and where it stops.

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><picture><source srcset="https://259567114-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FF6bxg8i3T6U4vZ3k7PHi%2Fuploads%2FmfyjQYtXffHIhbZR7pp9%2Fclarity__white.png?alt=media&#x26;token=dd4bd7c7-3230-4349-952a-c401750f13ea" media="(prefers-color-scheme: dark)"><img src="https://259567114-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FF6bxg8i3T6U4vZ3k7PHi%2Fuploads%2F8kVcuofmP1EIlrXUsLqx%2Fclarity.png?alt=media&#x26;token=90935e1a-3f44-4b01-90c8-f9e406c13a8e" alt="" data-size="line"></picture></td><td><h4>Connections</h4></td><td><a href="/assistance/the-assistant/connections.md">Connections</a></td></tr></tbody></table>

{% columns %}
{% column %}
{% content-ref url="/pages/jaGpQWjETS3dwv2yneLk" %}
[Configure the AI Assistant](/assistance/the-assistant/configure-the-ai-assistant.md)
{% endcontent-ref %}
{% endcolumn %}

{% column %}
{% content-ref url="/pages/JgC043lTBvbd2lZFOXX3" %}
[Assistant analytics](/assistance/the-assistant/assistant-analytics.md)
{% endcontent-ref %}
{% endcolumn %}
{% endcolumns %}

The AI Assistant[^1] lets your users ask a question in their own words and get a written answer, drawn from the Userlane content you have published, without leaving the application they are working in. Instead of searching through a list of guides or articles, a user describes what they need, and the Assistant replies, points them to the next step, and connects them to support when it cannot help. It sits in the Assistant panel alongside guides and announcements, reachable through the sidebar tabs.

<div align="left" data-with-frame="true"><figure><img src="https://259567114-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FF6bxg8i3T6U4vZ3k7PHi%2Fuploads%2FSwJGf7eGTNCvSopUfGZM%2FAI%20assistant.png?alt=media&#x26;token=c6afffcf-7b4c-41ce-ada3-011eff62d80f" alt="The AI Assistant in its initial state showing the greeting, starter questions, and three sidebar tabs" width="375"><figcaption><p>The AI Assistant before a conversation starts.</p></figcaption></figure></div>

## How a conversation works

A user types a question the way they would ask a colleague. The Assistant looks across the content published for that application, writes a short answer from what it finds, and cites the source it drew from. Three things can happen next.

**An answer with sources.** The Assistant returns a written response and shows where it came from: a source title and a quoted passage. The user can open the source directly.

**A next step.** Where it helps, the Assistant offers a next step the user can take without leaving their task. It can start a guide, open an announcement, or open a linked document.

**A suggested follow-up.** After answering, the Assistant suggests a next question based on what was asked. The user can tap it or type something different.

The Assistant keeps the thread of the conversation, so a user can refine or build on the previous answer. A user can also start a fresh chat at any time and pick up a recent one later.

When the Assistant cannot answer from your content, it points the user to the [support contact](/assistance/the-assistant/connections/support-email-routing.md) you have set, so nobody is left at a dead end. After every answer, thumbs up and thumbs down buttons let the user give feedback.

## Where answers come from

The Assistant is grounded in the content you have published and does not invent answers about your application. It draws from several sources, and every answer cites which one it used.

**Userlane content** is the first source: guides, announcements, and documents published for the application.

**Connected knowledge bases** come next. Through knowledge connectors, the Assistant answers from documentation the organization already keeps in systems such as SharePoint, Confluence, KnowledgeOwl, or Zendesk. Connecting existing documentation means the Assistant can answer without new content being written.

**A connected AI service** is an additional source, if the organization has linked one. The connection is retrieval-only and verifies each user through Microsoft Entra ID, so the service returns content that user is allowed to see, and the Assistant cites it like any other source. See [Connect your own AI service](/assistance/the-assistant/connections/connect-your-own-ai-service.md).

**General knowledge** is the Assistant's own model. It is on by default, so the Assistant can answer questions that your content does not cover, and those answers are labeled as coming from general knowledge. You can switch this off for applications that should only ever answer from your own content.

Sources appear with a title and a quoted snippet, so the reader can judge the answer before acting on it.

<div align="left" data-with-frame="true"><figure><img src="https://259567114-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FF6bxg8i3T6U4vZ3k7PHi%2Fuploads%2FWZWyGlOYLmEpzQulbwdx%2Fassistant4.png?alt=media&#x26;token=78badbdc-14fd-43e9-8c57-784470736aa5" alt="A full Assistant response showing sources, a guide card, and feedback buttons" width="375"><figcaption><p>An answer with its sources, a guide to start, and feedback buttons.</p></figcaption></figure></div>

<div align="left" data-with-frame="true"><figure><img src="https://259567114-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FF6bxg8i3T6U4vZ3k7PHi%2Fuploads%2FE0BmgYSR865P2bpp2hbX%2FAIAssistant2.png?alt=media&#x26;token=1b1715a5-d4b8-4971-91f5-9f2714ce8969" alt="Close-up of a source citation with title and quoted snippet" width="375"><figcaption><p>A source citation with its title and quoted snippet.</p></figcaption></figure></div>

## Access and language

The Assistant only draws on content the user already has access to. For Userlane content and for a connected AI service, it does not surface anything a user could not otherwise see. For knowledge connectors, access follows the account or registered application the connection uses, so what that account can read is what the Assistant can answer from. [Knowledge connectors](/assistance/the-assistant/connections/knowledge-connectors.md) explains how to scope it.

Users can ask in their own language, and the Assistant replies in the same language, based on their Userlane language settings.

Answers are generated by AI and should be checked for accuracy.

{% hint style="warning" %}
**The Assistant answers, shows, and routes. It does not take actions inside the application.** Guardrails[^2] keep the Assistant off topics you restrict, and the general knowledge setting controls whether answers can come from outside the organization's own content.
{% endhint %}

## Configuring the Assistant

As an administrator, you control how the Assistant behaves for your users: whether it is on and who sees it, its name and welcome message, its tone, the topics it will not answer on, whether it can use general knowledge, the starter questions it offers, and the support contact it hands off to. These settings live under **Agentic Assistance** in the portal and apply per application. [Configure the AI Assistant](/assistance/the-assistant/configure-the-ai-assistant.md) covers each one.

## Relevant articles

{% content-ref url="/pages/jaGpQWjETS3dwv2yneLk" %}
[Configure the AI Assistant](/assistance/the-assistant/configure-the-ai-assistant.md)
{% endcontent-ref %}

{% content-ref url="/pages/QG9FzFKcxvGs0ur1QJBQ" %}
[Connections](/assistance/the-assistant/connections.md)
{% endcontent-ref %}

{% content-ref url="/pages/JgC043lTBvbd2lZFOXX3" %}
[Assistant analytics](/assistance/the-assistant/assistant-analytics.md)
{% endcontent-ref %}

{% content-ref url="/pages/14ViydxzHwBkmZMnQj6h" %}
[AI features: what data is processed and where](/security-and-privacy/ai-features-what-data-is-processed-and-where.md)
{% endcontent-ref %}

## FAQ

<details>

<summary>What is the difference between the AI Assistant and a guide?</summary>

The AI Assistant answers questions conversationally. A guide walks someone through a process step by step on the application screen. When a question maps to a process, the Assistant hands off to the relevant guide.

</details>

<details>

<summary>Does the AI Assistant remember previous questions?</summary>

Yes. The Assistant keeps the thread of a conversation, so a user can build on the previous answer. A user can start a new chat at any time and return to a recent one later.

</details>

<details>

<summary>Can the AI Assistant answer from the organization's own documentation?</summary>

Yes. Through knowledge connectors, the Assistant answers from content in SharePoint, Confluence, KnowledgeOwl, or Zendesk. Every answer cites the source it drew from.

</details>

<details>

<summary>What happens when the AI Assistant cannot answer?</summary>

The Assistant points the user to the support contact set for the application, so the user is not left at a dead end.

</details>

<details>

<summary>Can the organization control what the AI Assistant talks about?</summary>

Yes. An administrator can tell the Assistant which topics it should not answer on, and it declines those questions and points the user to support. The general knowledge setting controls whether the Assistant can answer from outside the organization's own content.

</details>

<details>

<summary>Does the AI Assistant use general knowledge?</summary>

By default, yes, when the organization's content does not cover a question. Those answers are labeled as coming from general knowledge. An administrator can switch this off for any application.

</details>

[^1]: The AI Assistant is the conversational feature inside the Userlane Assistant panel that answers questions from published and connected content, and hands off to guides or support.

[^2]: Guardrails are topics an administrator tells the Assistant not to answer on. On those topics the Assistant declines and points the user to support.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.userlane.com/assistance/the-assistant.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.