# Gen AI

The Gen AI application enables organizations to use Generative AI features within Infraon. After installation, admins can configure providers such as Gemini, OpenAI, Azure, and Infraon, add API keys, set token limits, and select models for different tasks. The system uses the chosen provider for each AI action, monitors token usage, switches keys when needed, and falls back to Infraon AI if limits are reached. All AI activity and key usage are captured in audit logs for tracking.

## How to Install

### **Installation Steps**

**Step 1:** Go to the Marketplace tab on the left navigation panel to begin the integration. Click on the **Gen AI**card from the available integrations to open the plugin details page.

![](https://content.gitbook.com/content/E4mkwSP8a1BSD9BFNFav/blobs/DaIkP4pr9GAQFPtOd3ue/Unknown%20image)

**Step 2:** On the plugin page, click the **Install** button to add the Gen AI integration to your Infraon environment.

Once installed, the configuration page will become available for setup.

![](https://content.gitbook.com/content/E4mkwSP8a1BSD9BFNFav/blobs/QId0Furw3SsAx8S3DxR6/Unknown%20image)

After installation, the available providers to add are:

* OpenAI
* Azure
* Infraon
* Gemini

**Step 3:** To configure an AI provider, click **Add New** and enter the required details for the API key.

<figure><img src="https://content.gitbook.com/content/E4mkwSP8a1BSD9BFNFav/blobs/0Hvt0GNM7CJwWiGVIMe5/Unknown%20image" alt=""><figcaption></figcaption></figure>

Use the table below to understand each field and the type of information to provide.

![](https://content.gitbook.com/content/E4mkwSP8a1BSD9BFNFav/blobs/BTeY0lI8IqJmHlRh27p6/Unknown%20image)

| Label                     | Actions                      | Description / Example                                                                                                                  |
| ------------------------- | ---------------------------- | -------------------------------------------------------------------------------------------------------------------------------------- |
| **Key Name \***           | Enter a text value           | <p>A name to identify this API key within Infraon.<br><br><strong>Example:</strong> openai-key-001</p>                                 |
| **API Key \***            | Enter the provider’s API key | <p>Paste the API key generated from the selected provider’s dashboard.<br><br><strong>Example:</strong> sk-proj-7c6eUKP04ylIyYT...</p> |
| **Monthly Token Limit\*** | Enter a numeric value        | <p>Defines the maximum number of tokens allowed for this key per month.<br><br><strong>Example:</strong> 2000000</p>                   |
| **Chat Model\***          | Select from the dropdown     | <p>The chat or language model to be used for text-based tasks.<br><strong>Example:</strong> gpt-4o-mini</p>                            |
| **Embedding Model\***     | Select from the dropdown     | <p>The embedding model used for vector generation or semantic tasks.<br><strong>Example:</strong> text-embedding-3-large</p>           |

**Step 4:** After adding API keys, select the corresponding app from the Available Apps list to view the installed keys. Click **Provider Settings** to configure how the selected AI provider is used within Infraon.

<figure><img src="https://content.gitbook.com/content/E4mkwSP8a1BSD9BFNFav/blobs/KDaxvilm46ijXNa5dEEr/Unknown%20image" alt=""><figcaption></figcaption></figure>

You can assign models to specific AI features or apply a single configuration to all AI features. Use the table below to understand each setting.

![](https://content.gitbook.com/content/E4mkwSP8a1BSD9BFNFav/blobs/xiCaG62qhPVQSdZDoFUI/Unknown%20image)

<table><thead><tr><th width="190.60003662109375">Label</th><th width="163.5999755859375">Action</th><th>Description / Example</th></tr></thead><tbody><tr><td><strong>Use for All AI Features</strong></td><td>Toggle On/Off</td><td>When enabled, the selected model is applied to all AI features.<br><br>When disabled, you can map specific models to individual features.</td></tr><tr><td><strong>Features*</strong></td><td>Select one or more features.</td><td>Choose the Infraon AI features that should use this provider’s configuration.<br><br><strong>Example:</strong> Select from Text Enhancement, KB Conversion, Service Classification, and Bots.</td></tr><tr><td><strong>Model*</strong></td><td>Select from the dropdown.</td><td>Pick the model to use for the selected features. This depends on the provider and the available model list.<br><strong>Example:</strong> gpt-4, gpt-4o-mini, etc.</td></tr></tbody></table>

Click **Save** to apply the provider configuration.

**Step 5:** Click the **Settings** icon to view and update how the installed API keys are applied across Gen AI features.

![](https://content.gitbook.com/content/E4mkwSP8a1BSD9BFNFav/blobs/7foMfblvQaAdcFyTAUbU/Unknown%20image)

<table><thead><tr><th width="188.20001220703125">Label</th><th width="191.800048828125">Action</th><th>Description / Example</th></tr></thead><tbody><tr><td><strong>Find Similar Incidents</strong></td><td>Toggle ON/OFF and select an API key</td><td>Uses AI to identify and suggest similar incident tickets based on description and context.<strong>Example:</strong> When raising an incident, the system suggests: “Similar incident found: Network outage reported yesterday.”</td></tr><tr><td><strong>Agentic AI Assistant</strong></td><td>Toggle ON/OFF and select an API key</td><td>Enables the AI-powered assistant to help with responses, summaries, and task suggestions during ticket handling.<strong>Example:</strong> While updating a ticket, the assistant suggests a draft response or a summary.</td></tr><tr><td><strong>SLA Breach Prediction</strong></td><td>Toggle ON/OFF and select an API key</td><td>Predicts potential SLA breaches before they occur by analyzing ticket data and patterns.<strong>Example:</strong> System highlights: “This incident is likely to breach SLA in 2 hours.”</td></tr><tr><td><strong>Incidents to KB</strong></td><td>Toggle ON/OFF and select an API key</td><td>Converts resolved incidents into Knowledge Base articles using AI-generated summaries and solutions.<strong>Example:</strong> After closing a ticket, AI generates a KB article draft with issue, cause, and resolution.</td></tr></tbody></table>

Changes applied here determine which provider and key will be used for each AI task within Infraon.