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

# Managing Settings

Configure global AI settings in the Umbraco backoffice.

AI Settings provide a central place to configure system-wide defaults, including which profiles are used when no specific profile is specified.

## Accessing Settings

1. Navigate to the **AI** section in the main navigation.
2. Click **Settings** in the tree.

![The AI Settings page with Default Chat, Classifier, and Embedding profile pickers](https://583970477-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FCHKT9bQhbR3swAdBvil4%2Fuploads%2Fgit-blob-0997e615588d276c1a96a702cb6449a31d98a8b8%2Fbackoffice-ai-settings-page.png?alt=media)

## Available Settings

### Default Chat Profile

The profile used for chat operations when no profile is explicitly specified.

| Field                | Description                         |
| -------------------- | ----------------------------------- |
| Default Chat Profile | Select from available chat profiles |

When set, this profile is used by:

* `IAIChatService.GetChatResponseAsync()` without a profile ID
* Prompts without an associated profile
* Agents without an associated profile

### Classifier Chat Profile

An optional profile for internal classification tasks such as agent routing. The classifier profile lets you use a cheaper or faster model for basic classification decisions, reducing costs.

| Field                   | Description                         |
| ----------------------- | ----------------------------------- |
| Classifier Chat Profile | Select from available chat profiles |

When set, this profile is used by:

* Agent routing in "Auto" mode, in both Contextual Copilot and Copilot Workspace (selecting the best agent for a user prompt)

If not set, the default chat profile is used instead.

When the experimental Decision capability is enabled and a Default Decision Profile is set, Auto mode tries Decision first. The classifier chat profile is then used only as a fallback.

### Default Embedding Profile

The profile used for embedding operations when no profile is explicitly specified.

| Field                     | Description                              |
| ------------------------- | ---------------------------------------- |
| Default Embedding Profile | Select from available embedding profiles |

When set, this profile is used by:

* `IAIEmbeddingService.GenerateEmbeddingsAsync()` without a profile ID

### Default Speech-to-Text Profile

The profile used for speech-to-text operations when no profile is explicitly specified.

| Field                          | Description                                   |
| ------------------------------ | --------------------------------------------- |
| Default Speech-to-Text Profile | Select from available speech-to-text profiles |

When set, this profile is used by:

* `IAISpeechToTextService.TranscribeAsync()` without a profile ID
* The voice input button in Contextual Copilot and Copilot Workspace

### Default Image Generation Profile

The profile used for image generation when no profile is explicitly specified.

{% hint style="warning" %}
Image Generation is experimental. This setting is only shown when the `Umbraco:AI:Experimental:ImageGeneration` feature flag is enabled. A value saved earlier is kept while the setting is hidden. See [Enabling image generation](/ai-in-umbraco/using-the-api/image-generation.md#enabling-image-generation).
{% endhint %}

| Field                            | Description                                     |
| -------------------------------- | ----------------------------------------------- |
| Default Image Generation Profile | Select from available Image Generation profiles |

When set, this profile is used by:

* `IAIImageGenerationService.GenerateImagesAsync()` without a profile
* The Image Generation Management API and `UaiImageGenerationController` without a `profileIdOrAlias`

### Default Decision Profile

The profile used for decisions when no profile is explicitly specified.

{% hint style="warning" %}
Decision is experimental. This setting is only shown when the `Umbraco:AI:Experimental:Decision` feature flag is enabled. A value saved earlier is kept while the setting is hidden. See [Enabling Decision](/ai-in-umbraco/using-the-api/decision.md#enabling-decision).
{% endhint %}

| Field                    | Description                             |
| ------------------------ | --------------------------------------- |
| Default Decision Profile | Select from available Decision profiles |

When set, this profile is used by:

* `IAIDecisionService.AskAsync()` without a profile
* The Decision Management API and `UaiDecisionController` without a `profileIdOrAlias`
* The Ask Yes/No, Ask Pick-One, Ask Score, and Ask Questions Umbraco Automate actions without a profile
* Agent routing in "Auto" mode, which tries Decision before the classifier chat profile. See [Decision-Based Routing](/ai-in-umbraco/add-ons/agent-copilot/copilot.md#decision-based-routing)

### AI Disclosure Notice

Controls the notice that tells users a response is AI-generated. The notice reads "Responses are AI-generated and may be inaccurate." and appears in:

* The chat in Contextual Copilot and Copilot Workspace, after the first message in a conversation.
* The prompt preview, once a prompt response has loaded.

| Option      | Behavior                                                                                |
| ----------- | --------------------------------------------------------------------------------------- |
| Always      | The notice is always shown. This is the default.                                        |
| Dismissible | The notice shows a **Dismiss** button. Once dismissed, it stays hidden in that browser. |
| Off         | The notice is never shown.                                                              |

With **Dismissible**, each location is dismissed separately. Hiding the notice in the chat does not hide it in the prompt preview. The dismissal is stored in a cookie in the user's browser for one year, so it applies per browser rather than per user account.

{% hint style="warning" %}
You may be legally required to tell people when content is AI-generated, for example under the EU AI Act. Check which rules apply to your organization before choosing **Dismissible** or **Off**.
{% endhint %}

The setting is found in the **Transparency** group on the Settings page.

## Configuring Settings

1. Navigate to the **AI** section > **Settings**.
2. Select the desired profiles via the pickers, and choose an AI Disclosure Notice option.
3. Click **Save**.

## Settings Precedence

Settings can be specified in multiple places. The order of precedence is:

1. **Explicit parameter** - Profile ID passed in code (always takes priority)
2. **Database settings** - Configured in the backoffice (this screen)

{% hint style="info" %}
For advanced scenarios like CI/CD or infrastructure-as-code, defaults can also be configured in `appsettings.json`. However, database settings take precedence. See [Settings Concept](/ai-in-umbraco/concepts/settings.md) for details.
{% endhint %}

## When Defaults Are Used

| Scenario                                                                                                   | Default Used  |
| ---------------------------------------------------------------------------------------------------------- | ------------- |
| `_chatService.GetChatResponseAsync(chat => chat.WithAlias("my-feature"), messages)`                        | Yes           |
| `_chatService.GetChatResponseAsync(chat => chat.WithAlias("my-feature").WithProfile(profileId), messages)` | No (explicit) |
| Prompt without ProfileId                                                                                   | Yes           |
| Prompt with ProfileId                                                                                      | No (explicit) |
| Agent without ProfileId                                                                                    | Yes           |
| Agent with ProfileId                                                                                       | No (explicit) |

## Clearing Defaults

To remove a default:

1. Open the picker.
2. Select the empty option.
3. Save.

When no default is configured:

* Operations without an explicit profile will fail with an error.

## Related

* [Settings Concept](/ai-in-umbraco/concepts/settings.md) - Understanding settings
* [Profiles](/ai-in-umbraco/concepts/profiles.md) - Profile configuration


---

# 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.umbraco.com/ai-in-umbraco/backoffice/managing-settings.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.