When WatsonX Code Assistant for Z was introduced by IBM, it primarily targeted developers and software architects involved in modernizing mainframe applications. It is designed to help developers understand legacy code, refactor and optimize it, transform it to Java, and validate the generated code. As the first GenAI-driven tool from IBM specifically aimed at mainframe users, it generated a lot of buzz. Some people expected a magic wand that could automatically transform 30+-year-old software into state-of-the-art applications. However, while WCA4Z is a great code assistant, human expertise is still very much needed in the loop.
The situation for system programmers is somewhat different. Mainframes remain critical systems for many organizations, and the main challenge is maintaining smooth operations and improving infrastructure amidst an aging population of system programmers and a dwindling pool of skills in most mainframe shops. I believe a smart assistant driven by GenAI can greatly help here. Recognizing this opportunity, IBM introduced watsonx Assistant for Z (WA4Z), leveraging advanced AI technologies to simplify and optimize interactions with the IBM Z platform. It offers two main features:
- Answers to queries about the mainframe platform. IBM built an LLM model to learn from IBM documentation, including manuals, reference books, and Redbooks. Users submit their queries using prompts similar to other LLM prompts. The new version allows users to ingest their own documentation into WA4Z and I think this is a killing feature. More information below!
- Automation of both complex and routine tasks. This watsonx assistant is built on the IBM watsonx Orchestrate platform and incorporates a customized large language model (LLM) with a domain-specific retrieval-augmented generation (RAG) framework.
For most experienced mainframers, the benefits are immediately apparent. Some of these benefits can be realized almost instantly, but WA4Z must be trained and customized to achieve full functionality. The good news is that customization can be done gradually; you can start using it immediately, and the system will only get better with time.
Ultimately, just imagine how your data center will perform after you teach WA4Z about your mainframe environment and customize automation for typical activities.
- Enhanced Productivity: By providing quick and accurate answers, the assistant helps users complete tasks more efficiently, reducing the time spent searching for information. WA4Z provides accurate answers, step-by-step guidance, and useful links during your interactions, acting like a knowledgeable guide. Make sure to learn prompt engineering tips and tricks to get the best results!
- Intelligent Automation: The assistant can recommend and execute relevant automated skills, simplifying the execution of tasks and improving overall efficiency. Additionally, WA4Z allows you to import existing automation tools like Ansible, JCL, and REXX into a skill catalog, making it easy to create advanced workflows.
- Knowledge Codification: IBM Z experts can codify their knowledge into the assistant, making it accessible to all users and enhancing collaboration.
- Operational Insights: The assistant provides operational insights with system context, helping users make informed decisions.
- Improved Onboarding: Young mainframer users can quickly access the IBM Z knowledge base and get answers to their questions, accelerating their learning curve and boosting confidence.
- Customization: Users can personalize the assistant based on their role, team, experience, and access level. This ensures that the assistant meets the specific needs of different users.
Bring Your Own Documentation
I believe the “Bring Your Own Documentation” feature of WA4Z is a crucial aspect of the tool, allowing users to upload (ingest) their own documentation directly into the system. This feature is particularly important because it enables the tool to tailor its assistance based on the specific context and intricacies of your mainframe environment. By ingesting technical documentation about your system, WA4Z can provide more accurate and relevant insights, recommendations, and automation tailored to the unique needs of each organization. This customization ensures that the tool is not only more effective but also aligns closely with the organization’s operational requirements and security protocols, ultimately enhancing productivity and maintaining data integrity.
Imagine a scenario where a young system programmer needs to create a new database table in Db2. A standard answer would provide the necessary syntax and commands. However, WA4Z goes a step further by offering guidance on specifics related to the Db2 and z/OS environment like naming conventions, allocation, parameters as well as best practices, and general advice on where, how, and when to create a new table. It also considers important aspects such as security, backups, and data governance policies. It’s just like having an experienced colleague by your side!
Flexible deployment
In the first version of WA4Z, IBM initially supported only on-prem deployment, which required significant hardware resources to run the tool. On-prem deployment makes a lot of sense, especially if you plan to upload (ingest) your own documentation, which obviously includes sensitive and confidential information.
With the new versions, IBM has significantly optimized the whole package, and it now supports more budget-friendly GPUs (like L40s). Additionally, a cloud version is now supported, providing a dedicated instance for each customer to prevent any possible security or privacy breaches.
Future enhancements
The technology around GenAI is currently a very hot topic in the IT industry, and we can expect rapid advancements over the coming months and years. IBM has proven to keep pace and WCA4Z/WA4Z are gaining momentum in the mainframe community. In forthcoming releases, I expect enhanced support for Linux, additional operational skills, and multi-language support. Stay tuned!
Typical users of IBM watsonx Assistant for Z within a mainframe team include:
- System Programmers: They are responsible for installing, configuring, and maintaining the mainframe operating system and related software. The assistant helps them automate routine tasks and troubleshoot issues more efficiently
- Operators: These users monitor and manage the day-to-day operations of the mainframe. The assistant provides them with quick answers to operational queries and helps automate repetitive tasks
- New and Early-Tenure Professionals: The assistant is particularly beneficial for new hires and early-career professionals, helping them quickly learn and navigate the mainframe environment
- Mainframe Experts: While experts can benefit from the assistant by offloading routine queries, they can also contribute their knowledge to the assistant, making it more robust for other users
Watsonx Assistant for Z changelog
Since the original introduction, IBM has significantly improved the product. In the past six months, they have announced several new releases and versions that bring substantial new value to customers:
Watsonx Assistant for Z V2.1 (December 2024)
- Get Operational Insights on IBM Z: Users can now inquire about the health of their LPARs, identify top CPU consumers, explore IBM Db2® buffer pool issues, or review system alerts—all by interacting with their personalized AI assistant. Currently, it supports OMEGAMON performance metrics for LPARs, Db2 instances, and situation events.
- Low-code Skills Studio: Skills Studio in IBM watsonx Assistant for Z simplifies the creation of powerful automations for builders and system programmers using a low-code approach. Users can leverage prebuilt skills or import automations from tools such as Ansible, REXX, and JCL.
- Lower Hardware Requirements: By minimizing its footprint, the solution supports more budget-friendly GPUs (like L40s) and significantly reduces infrastructure requirements, including CPU and memory. This leads to a lower total cost of ownership, making it an even more cost-effective option for clients.
- Flexible Deployment: Due to the lower hardware requirements, WXA4Z can now be deployed both in the cloud and on-premises.
- Contextual Search: Maintain context for subsequent questions.
- Expanded Content in RAG: In addition to IBM Z documentation, RAG now includes 35 additional Z-related Redbooks.
Watsonx Assistant for Z V2.0 (September 2024)
In its second major version, watsonx assistant for z introduces significant improvements:
- Bring Your Own Documentation: Now you can integrate your own documentation into the system. This feature adds a personal touch and allows access to indexed content from your data sources using RAG.
- Expanded Content: IBM has enriched the RAG with content from over 220 IBM Z software products, providing more comprehensive answers to questions about the mainframe platform.
- Additional Pre-built Skills: New pre-built skills for common z/OS tasks are available, enabling builders to integrate these automations into an assistant to speed up time to value. These skills now include z/OS, Db2, IMS, and CICS, along with the capability to import APIs to trigger JCL, REXX, and Ansible as skills.
- Simplified Architecture: The architecture has been streamlined with embedded OpenSearch and the elimination of the prerequisite for watsonx Discovery.
Watsonx Assistant for Z V1.0 (June 2024)
IBM introduced WXA4Z with a set of initial features:
AI Assistant Capabilities:
- Answer Generation: Provides responses based on IBM Z-specific content to address users’ questions. Skill Recommendations: Suggest relevant automated skills that can be utilized during conversations.
Builder Capabilities:
- Automation Import: Allows the import of existing automation into a skill catalog and the creation of custom automation workflows.
- Assistant Customization: Enables the building of an assistant with specific actions and settings, which can then be shared with users to facilitate learning and action-taking.
Benefits of Integration:
- Accelerated Onboarding: Helps new professionals quickly become productive with the support of a side-by-side AI assistant.
- Enhanced Productivity and Confidence: Improves productivity, autonomy, and confidence across all skill levels by providing a central hub for finding answers and executing both common and complex tasks.
These updates ensure that IBM watsonx Assistant for Z continues to evolve, offering enhanced capabilities and benefits to its users.
If you’re curious about the latest updates to IBM watsonx Code Assistant for Z (WCA4Z), check our blog for more details.
