What's new in IBM Content Assistant
Read about the features and capabilities of IBM® Content Assistant.
6 May 2025
The following are the new features of the IBM Content Assistant 25.2 release.
- Updated version of the Content Assistant package available
- The Content Assistant package
(
fncm-genai-version-n-YYYYMMDDTTTT.zip
) for this release includes these updates:- Updates to the
genai-ce.jar
file for Content Platform Engine environment.Learn more ...
- Updates to the
genaiPlugin.jar
file for Content Navigator environment.Learn more ...
- Updates to the
- Support for IBM Granite 3.3 large language model
- From this release, IBM Content Assistant
supports IBM Granite 3.3 as the default large language model.
Learn more ...
- Better formatted results for query and summary responses
- This release includes some changes to the default query and summary prompt templates. The
changes to the default query prompt template facilitate better formatted results in response to your
questions to the assistant. The changes that are made to the default summary prompt template adhere
to the Granite 3 prompt formatting guidelines.
Learn more ...
- Usage metrics for your Content Assistant instance
- Use the IBM Content Assistant
Dashboard to find the usage report for your Content Assistant instance. For the selected
Content Assistant instance, the
report contains usage statistics like the number of active users, number of document summaries that
were generated, the number of questions asked and the indexed data size.
Learn more ...
- Improved access management for your IBM Content Assistant subscription
- Create an account on the IBM SaaS console to manage the Content Assistant subscription and instances
for your subscription. You can also create users and user groups and manage user access to the
resources included in your account.
Learn more ...
3 April 2025
The following are the new features of the IBM Content Assistant 25.1 release.
- Support for queries in token-dense languages
- From this release, you can ask questions or generate summaries about one or multiple documents that contain content in one of the token-dense languages like Arabic, Armenian, Bengali, Burmese, Hebrew, Hindi, Persian, Russian, Urdu, or Korean.