## Opening AI Settings and Understanding What You Can Control
In Atloria, AI settings are managed at the organization level, so the choices you make here apply across your shared workspace rather than only inside one project. Start by opening the admin area, then go to your organization settings and look for the AI settings section. On this screen, you should expect to see separate areas for **Providers**, **Prompt Behavior**, **Usage Visibility**, and **Request History**. These sections work together: the provider section controls which AI service Atloria uses, prompt behavior controls how Atloria frames AI requests, usage visibility controls who can see AI activity details, and request history shows past AI activity in a log-style view.

[SCREENSHOT: AI settings page showing Providers, Prompt Behavior, Usage Visibility, and Request History sections]

Different team members usually focus on different parts of this page. A **Project Administrator** typically handles the provider connection because that area includes the credentials and default service choices Atloria needs before AI features can work. A **Documentation Manager** usually reviews prompt behavior because those settings influence how consistently Atloria drafts or assists with documentation content. A **Support Team Lead** often checks usage visibility and request history so they can monitor activity, review past requests, and investigate reported issues.

Before changing anything, take a moment to identify the controls on the page. Look for:
- A provider selector or provider card
- Form fields for connection details
- Prompt-related settings such as default instructions or behavior options
- Visibility switches or access controls for usage and history
- A request history table or log with rows of past activity

Because these settings are organization-wide, even small changes can affect multiple teams. If you need help understanding how this admin area fits into the broader workspace, see [Using the Admin Workspace](doc:using-the-admin-workspace).

## Preparing Your Organization for AI Configuration
Before you begin setup, make sure you can actually edit organization settings in Atloria. If you can open the admin workspace but cannot change fields or save updates in the AI settings area, you may only have view access. You need organization administration access to configure providers, adjust prompt behavior, and control who can see usage and request history.

It also helps to gather all provider details before opening the form. The provider setup area may ask for several values, and entering them in one session reduces the chance of saving incomplete settings. Prepare the exact details your team plans to use, including:

| Item to prepare | Why it matters |
|---|---|
| Provider name | Needed to choose the correct option in the provider selector |
| API key or credential | Required to connect Atloria to the selected AI service |
| Default model | Determines which model Atloria uses for requests |
| Endpoint value | Needed if the provider form includes a custom connection address |
| Region value | Needed if the provider setup includes location-specific routing |
| Organization identifier | Needed if the provider requires organization-level account mapping |

You should also decide in advance who can view AI usage and request history. These screens may contain sensitive information about how Atloria is being used, and request history may include prompt and response details depending on what appears in your organization’s settings pages. Many teams give full access to administrators, monitoring access to support leads, and narrower review access to documentation managers.

Finally, review any internal rules your organization has for AI prompts, generated content, and request logging. Because these settings affect all projects in Atloria, it is better to align on prompt handling and visibility before you click **Save**. For related admin access guidance, see [Managing User Access and Administrative Permissions](doc:managing-user-access-and-administrative-permissions).

## Setting Up Your AI Provider
1. Open the **AI settings** page in your organization’s admin area and go to the **Providers** section.

2. Find the provider selector and choose the AI provider your organization wants Atloria to use. If more than one provider is available, make sure you select the one your team approved for organization-wide use.

3. Complete every required field shown in the setup form. Depending on what Atloria displays for your provider, this may include:
   - **API key**
   - **Model**
   - **Endpoint**
   - **Region**
   - **Organization identifier**

4. Enter each value exactly as provided by your organization or provider account owner. Pay close attention to copied credentials. A single incorrect character can prevent Atloria from connecting successfully.

5. Click **Save** and wait for the page to finish updating. After submission, look for any confirmation message, validation warning, or status indicator on the provider section. If Atloria shows a connected, active, or valid status, your provider details were accepted. If it shows an error, return to the fields and correct the highlighted values before saving again.

[SCREENSHOT: Providers section with provider selector and connection fields]

You can return to this same section later to update an existing setup. Common reasons to edit the provider configuration include:
- Replacing an expired or rotated API key
- Changing the default model used by Atloria
- Updating endpoint or region values
- Switching the organization to a different provider

When you update an existing provider, save the changes and check the status again right away. That helps you catch connection problems before users start seeing failed AI actions elsewhere in Atloria. If your organization uses AI heavily across documentation workflows, it is worth coordinating provider changes during a low-activity period so teams are not interrupted.

## Controlling Prompt Behavior for Organization-Wide AI Responses
1. From the **AI settings** page, open the **Prompt Behavior** section.

2. Review the options Atloria shows for how prompts are structured or sent. Depending on your organization’s setup, this area may include default instructions, behavior switches, or other controls that shape how Atloria asks for AI responses.

3. Read each option carefully before changing it. Prompt behavior settings affect how consistently Atloria responds across projects, documentation work, and support-related tasks. A change made here can influence AI-generated drafts, suggested wording, and other responses seen by multiple teams.

4. Adjust the settings based on the kind of output your organization needs:
   - For more consistent, controlled responses, use stricter instructions and any available behavior limits.
   - For broader drafting help or more flexible writing support, allow more open-ended behavior where the page provides that option.

5. Click **Save** and note any confirmation message shown on the page.

[SCREENSHOT: Prompt Behavior section with default instructions and behavior options]

This section is especially important for documentation teams. If your writers want Atloria to produce content in a more uniform style across projects, organization-level prompt behavior is the place to standardize that. It also matters for support workflows. A support team may prefer tighter, more predictable responses so answers stay aligned with approved documentation and internal guidance.

A useful way to think about these settings is by balancing consistency and flexibility:
- **Stricter settings** are better when your team wants repeatable wording, controlled tone, or narrower response patterns.
- **Broader settings** are better when your team wants help brainstorming, drafting, or exploring different wording options.

Because this area affects everyone using AI features in Atloria, review prompt changes with the people who rely on AI most often, especially documentation managers and support leads. If you want to compare this with project-level content workflows, see [Configuring AI Settings for Your Organization](doc:configuring-ai-settings-for-your-organization) and [Creating and Editing Documentation Pages](doc:creating-and-editing-documentation-pages).

## Managing Who Can See AI Usage and Request History
1. Open the **AI settings** page and go to the **Usage Visibility** area.

2. Review the available visibility controls. These settings determine whether organization members can open AI usage pages and see consumption details.

3. Set access based on role. A practical pattern is:
   - **Administrators**: full access to usage and request history
   - **Support Team Leads**: access to monitor usage and inspect request history
   - **Documentation Managers**: limited review access where appropriate

4. Next, open the **Request History** visibility controls and decide who should be allowed to inspect past AI requests. If Atloria displays prompts or responses in the history view, be careful not to give broad access unless your organization is comfortable with that level of visibility.

5. Save your changes, then reopen the settings page to confirm the selected visibility options remain in place.

The **Usage** area and the **Request History** area serve different purposes. In Atloria, usage usually helps you understand overall activity, while request history helps you inspect individual events. Depending on what your organization’s pages display, the usage area may show:
- Request counts
- Activity totals
- Provider-related usage details
- Time-based summaries

The request history area may show:
- Timestamps
- Provider references
- Request outcomes
- Prompt text
- Response text
- Status entries

[SCREENSHOT: Usage visibility controls and request history access settings]

Use the narrowest visibility that still supports your team’s work. Administrators usually need the broadest access because they manage organization-wide settings. Support leads often need monitoring access to investigate issues. Documentation managers may only need enough visibility to review generated output patterns rather than every request detail. For broader admin guidance, see [Reviewing Security and Audit Controls](doc:reviewing-security-and-audit-controls).

## Reviewing Request History and Usage Details
1. Open the **Request History** section or page in Atloria’s AI settings area.

2. Read the log one row at a time. Each row is typically used to represent one AI request. Focus first on the **timestamp** so you know when the request happened, then check the **provider** reference and the **outcome** or status shown for that entry.

3. If Atloria provides filters, sorting controls, or a date range selector, use them to narrow the list. This is especially helpful when you are reviewing a large number of requests or investigating a specific issue reported by a user.

4. Switch to the **Usage** area when you want a higher-level view. Usage details are useful for spotting increases in activity, confirming that teams are actually using AI after rollout, or noticing unexpected consumption patterns tied to a provider or time period.

[SCREENSHOT: Request history table with timestamp, provider, and outcome columns]

When you review request history, look for patterns rather than isolated entries. For example:
- Several failed requests in the same time window may point to a provider setup problem.
- A sudden increase in activity after a rollout may confirm adoption by support or documentation teams.
- Repeated requests around the same topic may show where users need better source documentation.

Support teams can use request history to match a reported issue with the time it occurred. If a user says an AI action failed or returned an unexpected result, filter the log by date and review entries around that time. Documentation managers can use the same history to review how generated content aligns with expected style or prompt behavior.

Usage details are better for trend monitoring, while request history is better for investigating specific events. Used together, they give you both the big picture and the row-by-row detail needed to understand how AI features are being used in Atloria. For more on monitoring activity, see [Monitoring AI Usage and Request History](doc:monitoring-ai-usage-and-request-history).

## Verifying Your AI Configuration
1. Return to the **Providers** section and confirm the saved provider shows a healthy status, confirmation message, or other successful connection result. If Atloria offers a test action or if your team can run a first live AI action from another screen, use that to confirm the provider is active.

2. Check the **Prompt Behavior** section and verify that your latest settings are still saved. Then compare AI output before and after the change if you have a recent example available. This is the easiest way to confirm that organization-level prompt updates are affecting responses as expected.

3. Review **Usage Visibility** and **Request History** access by signing in with the intended role, or by asking someone with that role to open the same AI settings pages. Confirm that administrators can see what they need, support leads can monitor activity if intended, and documentation managers only see the level of detail you planned.

4. Open **Request History** and make sure new activity appears after a test or live request. If the history is empty, confirm that AI requests have actually been made since the provider was configured.

Common issues to check:
- **Invalid API credentials**: Reopen the provider form and re-enter the API key exactly.
- **Missing provider fields**: Make sure every required field in the provider setup form is completed.
- **Empty request history**: Confirm that at least one AI request has been sent after setup.
- **Users cannot see usage data**: Recheck the visibility settings and confirm the affected role has the intended access.

[SCREENSHOT: Saved provider status and request history after a successful test]

Verification is worth doing immediately after setup and again after any provider change, model update, or visibility adjustment. That way, you can catch problems before they affect documentation teams, support workflows, or project users across Atloria.

## Overview
AI settings in Atloria give your organization one place to control how AI features are connected, guided, and monitored. The most important areas on the page are **Providers**, **Prompt Behavior**, **Usage Visibility**, and **Request History**. Together, these settings determine which AI service Atloria uses, how requests are shaped, who can review activity, and how past requests are displayed for troubleshooting or oversight.

A typical setup flow looks like this:
- Open the organization-level **AI settings** page
- Choose and save your provider in **Providers**
- Review organization-wide instructions in **Prompt Behavior**
- Decide who can see activity in **Usage Visibility**
- Confirm request logging and review access in **Request History**
- Run a test and verify the results

These settings are best managed with clear ownership:
- **Project Administrators** usually handle provider setup and updates
- **Documentation Managers** usually review prompt behavior for consistency
- **Support Team Leads** usually monitor usage and request history

Because the settings apply across the organization, changes should be made carefully and verified right away. A provider update can affect whether AI actions work at all. A prompt behavior change can alter the tone or consistency of generated content. A visibility change can either help teams investigate issues or expose more detail than intended.

If you are continuing your AI setup work, the next most useful guides are:
- [Setting Up and Maintaining AI Providers](doc:setting-up-and-maintaining-ai-providers)
- [Monitoring AI Usage and Request History](doc:monitoring-ai-usage-and-request-history)
- [Managing User Access and Administrative Permissions](doc:managing-user-access-and-administrative-permissions)

If you are reviewing AI use from a documentation workflow perspective, see [Creating and Managing AI Support Agents](doc:creating-and-managing-ai-support-agents) and [Using Analytics to Prioritize Documentation Improvements](doc:using-analytics-to-prioritize-documentation-improvements).