Co-Authored by @Harisha Paildar

IBM FileNet P8, finding information is no longer the biggest challenge—understanding the content quickly is.Users often spend significant time opening documents, reading through pages, comparing versions, and searching for specific information.

IBM Content Assistant addresses this challenge by bringing Generative AI capabilities directly to FileNet repositories. Features such as document summarization, semantic search, document comparison, andquestion answering help users interact with enterprise content using natural language.

At the center of these AI capabilities is the GenAI Configuration object. It acts as the bridge between FileNet P8and the AI services used by Content Assistant.

This blog provides an overview of GenAI Configuration, its purpose, key use cases, and how it enables andconfigures in FileNet P8 using the ACCE (Administration Console for Content Platform Engine).

What is GenAI Configuration?

GenAI Configuration is a special FileNet object introduced by the GenAI Extensions Add-on. It acts as thecentral configuration object that connects a FileNet Object Store to IBM Content Assistant and underlying AIservices such as watsonx.ai.

Once configured, users can ask natural language questions about repository content and receive AI-generated responses based on documents stored in FileNet.

Why is GenAI Configuration Needed in ACCE (Administration Console for Content Platform Engine)?

Step 2: Initiate Activation

In the right pane, the Activation tab displays a message indicating that Content Assistant is currently inactive:

"Content Assistant functionality is not active for this object store."

To begin the activation process:

1. 1.  Locate the Activate button below the message.
   2. Click Activate.
   3. The Activation Wizard opens in a new dialog window.
   4.  Follow the wizard steps to configure and activate the GenAI service for the selected object store.

Step 3: Configure Access Code and Service URL

The first page of the Activation Wizard collects the information required to connect Content Assistant to theGenAI service.

Access Code

The Access Code field is mandatory and cannot be left blank.

• • •  Compass (SaaS) customers: Enter the access code provided on the Content Assistant dashboard.
    • Horizon (On-Premises) customers: Enter any non-empty value.

Service URL

Specify the URL of the Content Assistant service.

• • • Default URL: https://filenetai.saas.ibm.com/prod
    •  US SaaS customers: Use the default value.
    • Frankfurt SaaS customers: Replace the default value with the regional service URL.
    • Horizon customers: Enter the URL of the on-premises Content Assistant service.

Proxy Configuration

If your environment requires outbound traffic through a proxy server, check the checkbox for “Use a proxy server”.

• • • When enabled, the next wizard step collects proxy configuration details.
    • When disabled, the wizard proceeds directly to the Vector Indexing configuration.

  Click Next to continue.

Step 4: Configure Proxy Server Settings (Optional)

 This step is displayed only when Use a proxy server is selected.

  Proxy Server Host

 Enter the hostname or IP address of the proxy server. Examples:

• • • proxy.company.com
    • 192.168.1.100

 Proxy Server Port

 Enter the port used by the proxy server. Common valuesinclude:

• • • 8080
    • 3128
    •  8888 

TLS Enabled

 Select TLS Enabled if the proxy server requires secure TLS/SSL communication. Leave it unchecked for standard HTTP proxy connections.

After entering the required information, click Next to proceed.

Step 5: Configure Vector Indexing

The final step of the Activation Wizard configures the initial vector index used by Content Assistant.

The wizard displays the following information:

Activation will create an initial vector index. If you want to use the service-defined default embedding model for this index, leave the field 

below   empty.Otherwise, enter the desired embedding model name.

Embedding Model Name (Optional)

Choose one of the following options:

• • •    Leave the field blank to use the service-defined default embedding model (recommended).
    •    Specify a model name if your deployment requires a particular embedding model.

Example :

• • •  ibm/slate-30m-english-rtrvr
    • ibm/slate-125m-english-rtrvr

Click Finish to complete the activation process. Upon completion:

• • • Content Assistant is activated for the object store.
    • The initial vector index is created.
    • The wizard closes and returns you to the GenAI Configuration interface.

   Once activation is complete, administrators can update additional configuration settings through the GenAI Configuration interface.

Configuration Tabs

Tab 1: General

The General tab contains the core connection and network configuration required for Content Assistant tocommunicate with the GenAI service.

1.  Access Code (Read-Only)

 Displays the access code that was provided during the activation process.

• • • •  Visible for reference purposes only.
      • Cannot be modified after activation except (#RESET-ME-NOW!).

2. Service URL

Specifies the endpoint URL used to connect to the Content Assistant service.

• • • • Editable after activation.
      • Update this value if the service endpoint changes or if you need to connect to a different deployment environment.

3.  Proxy Server Host

Defines the hostname or IP address of the proxy server used for outbound connections.

• • • • Optional setting.
      • Can be added, updated, or removed at any time.

4.  Proxy Server Port Specifies the port used for proxy communication.

• Editable after activation.
• Update this value whenever the proxy configuration changes.

 5. TLS Enabled

Determines whether secure TLS/SSL communication is used when connecting through the configured proxy server.

• • • •  Select the checkbox to enable TLS.
      • Clear the checkbox to use standard HTTP communication.

Updating Configuration

To modify any editable setting:

1. 1. Open the General tab.
   2. Update the desired field values.
   3.  Click Save in the toolbar to apply and persist the changes.

Tab 2: Vector Indexing

The Vector Indexing tab allows administrators to configure vector indexing behavior and manage the lifecycleof the vector index used by Content Assistant for semantic search and AI-powered retrieval.

Prefer Markdown Extract

Determines the format used when extracting document content for vector indexing.

• • •  Enabled: Documents are extracted in Markdown format before indexing.
    • Disabled: The default extraction format is used.

                This setting controls the property.

          Markdown extraction can improve the quality of indexing for documents containing headings, lists, tables,and other structured content.

Batch Processing Settings

These settings control how documents are processed during indexing operations.

Indexing Batch Items

Specifies the maximum number of documents processed in a single indexing batch.

Indexing Batch Characters

 Specifies the maximum number of document characters processed in a single indexing batch.

  After modifying these settings, click Save to persist the changes.

Index Management

The lower section of the tab displays the current vector index status and provides actions for managing the activation and deactivation of GenAI configuration.

Index Status

The Index Status field indicates the current state of the GenAI vector index and its availability for ContentAssistant operations.

• • •  Active – The GenAI configuration and vector index are active and available for semantic search, summarization, and other AI-powered capabilities.
    • Inactive – The GenAI vector index is not active. Semantic search and indexing capabilities areunavailable until a new index is created or the index is reactivated.
    • Resync – The system is synchronizing the GenAI configuration and vector index with the ContentAssistant service. Once synchronization is complete, the index returns to the Active state.

 Click Save in the toolbar to apply and persist the changes.

Tab 3: Summary Generation

The Summary Generation tab allows administrators to customize how Content Assistant generates AI-powered document summaries.

Summary Extraction Format

Administrators can choose whether document content is extracted in Markdown format before being sent for summary generation.

• • • Enabled – Preserves document structure such as headings, lists, and tables, often resulting in more accurate summaries.
    • Disabled – Uses the default extraction format.

Maximum Summary Length

Defines the maximum length of generated summaries.

• • • Lower values produce concise, high-level summaries.
    • Higher values generate more detailed summaries.

 Organisations can adjust this setting based on their business requirements and the desired level of detail.

Document Summary Prompt Template

The prompt template determines how the AI model generates document summaries. Administrators cancustomize the prompt to:

• • •   Control the tone and style of summaries.
    •   Focus on specific sections of a document.
    •   Emphasise key business information.

Click Save in the toolbar to apply and persist the changes.

Tab 4: Search

 The Search tab controls how Content Assistant processes natural language queries and retrieves relevant information.

Prefer Markdown Extract

This option determines how document content is extracted before being provided to theAI model during search and  question-answering   operations.

• • • Enabled – Content is extracted in Markdown format, preserving document structure such as headings, lists, and tables.
    • Disabled – The default extraction format is used.

For documents with rich formatting and structured content, enabling Markdown extraction can improve the quality and accuracy of  generated responses.

Default Language Model

Administrators can specify the language model used for search and question-answering operations.

If no model is specified, the service uses the default model configured for the environment.

Query Prompt Template

This template defines how user questions are interpreted and how retrieved content is used to generate responses.

 Customising the prompt allows organizations to:

• • •  Improve search relevance.
    • Align responses with business terminology.
    • Provide domain-specific guidance.
    • Optimize retrieval and answer generation behaviour.

Click Save in the toolbar to apply and persist the changes.

Tab 5: Document Comparison

 The Document Comparison tab controls how Content Assistant analyzes and compares documents using AI.

Comparison Prompt Template

The comparison template defines how differences between documents are identified and presented.

 Administrators can customise the comparison experience to:

• • • Highlight content changes.
    • Focus on structural differences.
    • Identify policy or contract modifications.
    • Emphasise business-critical updates.

Click Save in the toolbar to apply and persist the changes.

Tab 6: Watsonx Connection

The Watsonx Connection tab allows administrators to configure and manage the connection between Content Assistant and their IBM  watsonx.ai    environment.

This tab provides the settings required to authenticate and communicate with a watsonx.ai deployment, including:

• • •  API Key – Used to securely authenticate with watsonx.ai services.
    • Space or Project ID – Identifies the watsonx.ai space or project that contains the AI models to be used by Content Assistant.
    • Provider URL – Specifies the watsonx.ai service endpoint for your deployment region or private environment.

Click Save in the toolbar to apply and persist the changes.

Tab 7: Properties

The Properties tab provides a generic view of the underlying Content Assistant configurationsettings.

Unlike the specialised tabs, which expose commonly used settings through a guided user interface,the Properties tab allows administrators to view and edit configuration values directly.

Note: The GenAI Configuration capability is supported in CCX 26.0 and later releases. This feature requires the GenAI add-on and IBM Content Assistant to be installed and configured. The configuration screens, available settings, and supported AI capabilities described in this document correspond to the CCX 26.0 release.