IBM Linux on System Z with enhanced Watsonx Assistant AI capabilities. Businesses gain smarter automation, improved efficiency, and optimized hybrid cloud performance.
The primary goal of IBM Watsonx Assistant for Z is to transform the way IBM Z users interact with and control the mainframe. Watsonx Assistant for Z continues to receive interesting updates from IBM to help businesses streamline and simplify their mainframe operations, making them easier to use and less complicated while lowering the learning curve.
IBM is announce Watsonx Assistant for Z V2.2’s release. Here are four salient features among the many product upgrades and improvements:
Clients will have general access to these additional features as of March 26, 2025. Let’s examine the specifics.
Added IBM Linux on System Z support as a supported platform
Watsonx Assistant for Z is now available for clients to build and run on Linux operating systems installed on IBM Z (Linux on IBM Z), expanding the range of platforms available for deployment beyond x86 hardware. Customers may now safely benefit from mainframe features and leverage enterprise-scale AI capabilities on the most reliable, secure, and accessible IBM Z platform.
This release will have a hybrid approach of IBM Linux on System Z support, where LLMs run on x86 hardware to provide the best possible performance.
Expanded pre-built automations, providing real-time system insights
Through AI dialogue, IBM Watsonx Assistant for Z gives users instant access to information about the state of the IBM Z environment. With the addition of more operational abilities, it has greatly increased the coverage since December 2024 debut, when first supported important OMEGAMON performance measures.
This release streamlines level 1 operations, allowing you to ask the assistant to check the health of IMS, MQ, CICS subsystems, Networks, Storage, and JVMs in OMEGAMON, as well as the status of your system and critical applications in IBM Z System Automation, without having to switch between your tools.
Asking the assistant to “Show all CICS regions,” for instance, will provide a tabular list of CICS regions together with important performance indicators like CPU utilisation and transaction rates. You can delve deeper into specifics, including individual CICS transactions, by using skill flows. To find out more, you may also ask the assistant, “Check the status of applications?” that are under IBM Z System Automation’s control.
Watsonx Assistant for Z also provides insights with better data visualisations and formatting, which facilitates information interpretation and action.
More control over the LLMs with improved quality
The most recent Granite 3-8b instruct model is now utilised by the improved IBM Watsonx Assistant for Z. Because of its larger training dataset (12 trillion tokens compared to 2.5 trillion in the previous edition), the new model a specialised variant of the pre-trained IBM Granite Base offers stronger awareness and understanding, which leads to better replies.
By providing the LLM with unique prompts, experts (builders) can now customise the AI assistant’s responses for particular user personas. For example, prompts can be designed to require more thorough and verbose responses if the assistant is for early tenure IBM Z customers. In a similar vein, answers can be succinct, straightforward, and directive if they are intended for seasoned experts. This gives you more flexibility over how the Assistant reacts according to your persona, domain, experience, or important use case.
IBM has increased coverage of Redbooks and updated and adjusted domain-specific Retrieval-Augmented Generation (IBM Z RAG) to incorporate the most recent IBM Z material. Additionally, you can customise an assistant to concentrate on particular items that are crucial to your business requirements.
With the updated models, IBM Z RAG enhancements, and the capability to add personalised prompt instructions, IBM Watsonx Assistant for Z provides more accurate and higher-quality responses.
Pre-configured assistants to get started quicky
In order to automate the formation of a new AI Assistant during installation, it now offer a pre-configured assistant. You may use best practices to install your first AI assistant in a matter of minutes by using the pre-configured assistant. Make use of the pre-configured assistant and adjust it as necessary to satisfy your business needs.
On the horizon: Agentic AI to simplify and transform the IBM Z experience
It want to provide an agentic AI framework with IBM Watsonx Assistant for Z soon, which will further streamline the mainframe user experience and open the door to the next round of productivity improvements.
In order to enable AI agents to behave with greater domain knowledge and automation to drive tool calls, IBM is moving away from a general-purpose generative AI model. AI agents work independently, utilising real-time system data to deliver deeper automation and even more precise responses. To manage your environment more effectively, agentic AI combines the accuracy of reliable business logic with the flexibility of large language models (LLMs).
