Use | AI

Begin with configuration

To access AI features, you must configure an AI provider.

EclecticIQ Intelligence Center offers powerful features that require AI to work:

• The AI Assistant that streamlines the analyst workflow.

• Generation of reports from graphs and search results.

• Extraction of Entities from ingested and uploaded text documents.

• Assistance on text editing.

There are also a number of other functionalities that are greatly enchances with AI capabilities:

• Intelligence Compass requirements scanning.

• Natural Language Processing for search.

Data privacy disclaimer

When you use AI features, you send some of your data to the provider configured for that feature so that it can formulate its response.

See the data privacy disclaimer to learn more about how your data is processed.

User permissions

• Configuration of AI features requires a role with modify configurations permission.

• Usage of AI features requires a role with at least read configurations permissions.

• Other permissions may be required for related tasks.
  For example, modify entities is required to create and edit Report entities, where you can use AI features to generate text.

Permission documentation

See our permission documentation for more information on user permissions.

Credit usage

To access AI features, you must configure an AI provider. This configuration will allow EclecticIQ Intelligence Center to place API calls with the chosen AI provider. Placing these calls requires you to have credits for the chosen provider and model’s API service.

The amount of credits required depends on the length of the prompt sent for the action you want to execute and the way the prompt gets broken up (or “tokenized”) by the model it get sent to. See your chosen AI providers’ documentation to learn more about their tokenization process.

Configure

The following sections describe how to configure AI integrations for use in EclecticIQ Intelligence Center. Familiarize yourself with our AI disclaimer, especially the user responsibility section section, before you configure the integrations. Make sure any users who will use the AI integrations do the same.

Persistent configuration

Once you configure AI providers on your Intelligence Center instance, all users that have a role with read configuration permission can access the AI features.

Set up AI providers

Supported providers & models

Provider	Model(s)
OpenAI	ChatGPT-4o
Perplexity.ai	Sonar-pro
Ollama	Various*.

To set up an AI provider:

1. From the left sidebar, go to Settings > AI configuration > AI providers and API keys.

2. On the card of the API provider you would like to set up, select Add API key.

3. Under API Key, enter your key for the chosen provider.

4. Select Save and exit.

Choose model per feature

After setting an API key for your AI provider, you can choose the AI model to use for each of the features.

To choose AI models per feature:

1. From the left sidebar, go to Settings > AI configuration > Features AI model setup.

2. For each feature, select a model from the Select type drop-down menu Not every model is selectable for every feature, because some features require specific models to work.

Ollama integration

Ollama integration allow you to leverage self-hosted LLMs or the open-source LLMs that EclecticIQ can cloud-host for you.

To integrate with an LLM you are hosting through Ollama, you need to have a URL through which the model can be reached and an API key that provides access to the model.

If you’ve contracted EclecticIQ to cloud-host an open-source LLM for you, the URL and API key will have been provided.
If you’re interested in having EclecticIQ cloud-host an open-source LLM, please contact your CSM.