Online Help | Desktop App

Configuring integration with AI

Introduced in 2024.2

Updated in 2026

Alloy Navigator comes with a set of built-in AI-powered features that help users work smarter and faster by streamlining tasks, cutting down manual effort, and delivering a better user experience. To enable AI assistance, you need to provide your organization’s API key and configure additional settings in the desktop Settings App, under AI Integration > AI Configuration. This page guides you through setting up AI integration.

INFO: For details on available AI capabilities, see AI capabilities.

IMPORTANT: Alloy Navigator supports language AI models from the most trusted providers and gives you complete control over AI usage and the flexibility to manage associated costs. Explore which AI models you need and what each one is used for in Supported OpenAI models for AI integration.

Turn AI assistance on or off

You can control the availability of AI features in Alloy Navigator using a single toggle: the Enable AI features checkbox on the AI Integration > AI Configuration page. Use this option to enable or disable AI assistance across Alloy Navigator with a single click.

Depending on your configuration, there are three possible scenarios:

  • Initial setup

    By default, AI features are enabled, but no API key is configured. When attempting to use AI features, users with administrative rights will see a prompt to complete the setup (regular users will not be prompted to access the Admin Center).

    To complete the setup, you'll need to configure AI integration as described in Configure AI integration below.

  • AI integration is set up, AI features are enabled

    Once you've provided the API key and kept AI features enabled, both AI-Powered Insights and AI Writing Assistance become active and available to users.

  • AI features are disabled

    If you disable AI features, both AI-Powered Insights and AI Writing Assistance are completely hidden from the interface and unavailable to users—regardless of whether an API key or other AI integration settings are in place.

Configure AI integration

AI integration requires a verified account with your AI provider and a valid secret API key.

Alloy Navigator integrates with AI services from supported vendors, including OpenAI and Azure OpenAI. Alloy Navigator does not operate its own AI service, nor does it transmit any data to these AI services except what users submit themselves. We do not store, analyze, or control the information sent to AI services. All security, privacy, and data protection concerns, including any potential risks of data leakage, should be directed to the respective AI service providers.

To configure AI integration:

  1. In the Settings App, navigate to AI Integration > AI Configuration

  2. Specify your AI provider and enter your API key.

    1. Under AI Provider, specify your provider:

      • OpenAI API: Use this option for direct integration with OpenAI's models, offering flexible and powerful AI capabilities for various text-processing tasks.

      • Azure OpenAI Service: Choose this option if you prefer to use OpenAI's models hosted on Microsoft Azure, providing additional integration with Azure's cloud services and security features. Then use the Endpoint URL field to enter your unique URL provided by Azure to communicate with the API. You can obtain this URL in the Azure portal, it typically looks like this: https://<your-resource-name>.openai.azure.com/

        INFO: For instructions, see Integrations: Microsoft AI Foundry integration.

    2. In the API key field, provide your secure key for authenticating requests to your AI provider’s API.

      • For OpenAI: You can find or generate your API key in your OpenAI account dashboard, under the API Keys section.

      • For Azure OpenAI Service: You can obtain your API key from the Azure OpenAI resource, under the Keys and Endpoint page.

  3. Choose the completion model.

    • For OpenAI: In the Completion model field, keep the default selection, choose an alternative from the list, or enter a model name manually.

    • For Azure OpenAI Service: In the Completion deployment name field, enter the name of the deployment configured for the completion model in your Azure environment.

    INFO: For recommendations, see Completion model (AI response generation model).

  4. Turn Text-to-Speech feature on or off.

    Use the Enable Text-to-Speech checkbox to allow or prevent the AI from converting its responses into natural-sounding audio.

    Then specify the Text-to-Speech model you want to use:

    • For OpenAI API: Keep the default Text-to-Speech (TTS) model. You can also choose an alternative from the list or enter a model name manually.

    • For Azure OpenAI Service:: Enter the name of the deployment configured for the Text-to-Speech model in your Azure environment.

    INFO: For recommendations, see Text-to-Speech model.

  5. Control speech recognition.

    Use the Enable speech recognition checkbox to allow or prevent the AI from transcribing spoken input into text.

    Then specify the transcription model you want to use:

    • For OpenAI API: Keep the default transcription model. You can also enter an alternative model name manually.

    • For Azure OpenAI Service: Enter the name of the deployment configured for the transcription model in your Azure environment.

    INFO: For recommendations, see Transcription model.

    TIP: For speech recognition to function, the site hosting the Web App instance must allow access to audio input devices. Make sure there are no restrictions, such as Permissions-Policy: microphone or browser settings, that block microphone access.

  6. Control the AI's access to internal knowledge.

    Here you can enable or disable the AI's ability to search your organization's content for more relevant, context-aware suggestions.

    • Allow access to internal knowledge: Select the checkbox to allow AI to access your Service Desk content (Incidents and Knowledge Base Articles.

      Then specify the embedding model you want to use:

      • For OpenAI API: Keep the default embedding model. You can also enter an alternative model name manually.

      • For Azure OpenAI Service: Enter the name of the deployment configured for the embedding model in your Azure environment.

      INFO: For recommendations, see Embedding model.

      TIP: When the access is allowed, the system starts a system scheduled task that builds and maintains the vector index database that stores Incidents and Knowledge Base Articles. AI-Powered Insights and the AI Assistant search this content for more relevant responses. In the web-based Admin Center, yYou can view the task progress directly on the AI Configuration page, under Vector indexing progress. For details, see Vector index for AI-powered actions and AI Assistant.

    • Allow access to internal knowledge: Clear the checkbox to prevent access. This will result in less relevant suggestions.

      When cleared, AI-Powered Insights and the AI Assistant will rely solely on the AI's pre-trained knowledge and will not access your internal Service Desk content.

  7. Click Save to apply your changes.

Related Information:

Supported OpenAI models for AI integration

Vector index for AI-powered actions and AI Assistant

Integrations: Microsoft AI Foundry integration

Managing AI Writing Assistance actions