Planet Mainframe on LinkedIn
Planet Mainframe on LinkedIn
Planet Mainframe
Vertali Mainframe Security

AI-Powered WLM Cuts Batch Queue Time by Up to 70%

Sep 1, 2026

Artificial Intelligence, Cheryl Watson's Tuning Letter, z/OS

AI-Powered-Workload-Manager

Mark Wilson

Mark Wilson is a globally recognized thought leader and international speaker in mainframe security and technology, as well as a passionate advocate for all things Z. He serves as the Chief Editor and Producer of Cheryl Watson's Tuning Letter and is the Technical Director at Vertali. He has more than 40 years’ experience across numerous industries and diverse mainframe environments. Mark is also Region Manager for GS-UK and has been awarded IBM Champion status for the last five years. For more information email: info@vertali.com

What happens when Workload Manager can anticipate batch demand instead of simply responding to it?

The latest issue of Cheryl Watson’s Tuning Letter leads off with findings from testing IBM AI-powered Workload Manager (WLM). The test used an AI model trained on 30 days of SMF 99.2 data. Results showed:

  • Up to a 70% reduction in average job queue time for scheduled batch workloads
  • Up to a 48% reduction in peak queued requests.

These results headline the newest edition of theTuning Letter, available exclusively to subscribers.

How AI Changes Batch Management

The piece, written by Yan Wang and Anastasiia Didkovska, examines how the capability works, how to train and activate it, and how to evaluate its potential benefits. It also covers simulation mode, which allows teams to review AI predictions before enabling them to influence WLM decisions.

Compared with traditional WLM management, AI-powered WLM:

  • Reduced average job queue time from approximately 83 seconds to 25 seconds

  • Lowered average peak queued requests from approximately 80 jobs to 42

  • Started additional initiators roughly two minutes before scheduled demand arrived

The complete article provides technical details, performance metrics, and implementation guidance that subscribers need to assess the capability in their own environments.

What Is Cheryl Watson’s Tuning Letter?

For decades, Cheryl Watson’s Tuning Letter has helped mainframe professionals improve z/OS performance, reliability, and efficiency.

Each issue turns complex system behavior into practical guidance grounded in technical expertise, testing, and real-world experience. Subscribers gain detailed analysis they can apply to tuning decisions, system changes, and emerging IBM Z capabilities.

More in the Latest Issue

Along with the AI-powered WLM analysis, the September 2026 edition includes:

  • The anatomy of an abend—and what a system dump can reveal
  • Db2 Class 3 wait times and service task suspensions
  • Lessons from a z17 migration and JES2 checkpoint reconfiguration
  • A practical implementation guide for SMF Record Signing
  • A closer look at the z17 single-frame and rack-mount systems
  • The latest New Function APARs, conferences, and events

Subscribers can access the complete issue through the Tuning Letter website.

Not yet a subscriber? Check out subscription options for the latest issue and peruse the entire Tuning Letter archive.

Share Your Thoughts

We hope subscribers find the latest issue useful, practical, and thought-provoking – and that the curious nonsubscribers will consider signing up. We always welcome reader feedback.
All the best,

Mark
Mark Wilson

 

 

Editor, Cheryl Watson’s Tuning Letter

0 Comments

Submit a Comment

Your email address will not be published. Required fields are marked *

Sign up to receive the latest mainframe information

This field is for validation purposes and should be left unchanged.

Recently Published

Read More

The Enterprise Was Always About Data

The Enterprise Was Always About Data

by

AI Simply Reminded Us.For years, enterprise technology conversations have centered on infrastructure. We debated where applications should run. Then we spent the better part of a decade deciding which workloads belonged on premises, which belonged in the cloud, and...

Agentic AI on the Mainframe

Agentic AI on the Mainframe

by

It’s AI month at Planet Mainframe, and for Round Two of our Trivia Challenge, we’re turning to agentic AI. On the mainframe, this emerging technology is moving beyond answering questions and generating code. Agents can interpret goals, reason through multi-step...