IBM Corp. said today it’s building on the recent launch of generative artificial intelligence coding capabilities for mainframe developers by launching a suite of new offerings that will make it easier to run AI on their legacy data center hardware.

The updates to IBM Z are designed to help customers improve business outcomes by implementing business AI across multiple industries and use cases, the company said. They include a pair of new developer toolkits that support the most widely used AI frameworks and the Python programming language, enhancements to Machine Learning for z/OS, and performance improvements within the IBM z/OS operating system that’s used by mainframes.

According to IBM, the updates will enable clients to enhance the accuracy of insurance policy recommendations, increase the reliability and timeliness of anti-money laundering controls and reduce risk exposure for financial services providers.

The venerable mainframe computer, which can trace its history way back to the 1950s, is typically used by certain enterprises to host mission-critical business applications. They’re appreciated for their ability to process vast amounts of data rapidly, and have use cases in data analytics, enterprise resource planning and large transaction processing.

One of the reasons they continue to play a role is their high degree of availability, which makes them ideal for applications where downtime would be expensive and must be avoided. They’re also praised for their reliability and serviceability.

Although many organizations have shifted to cloud-based computing infrastructure, mainframes are still widely used. IBM claims two-thirds of the Fortune 100 and 45 of the world’s top 50 banks continue to rely on mainframes, along with eight of the world’s top 10 insurers, seven of the top 10 global retailers and eight of the top 10 telecommunications providers.

Given the enduring popularity of mainframes, IBM believes there is a big opportunity to enhance the critical workloads they power with the latest AI capabilities and that is what it’s doing with today’s announcements.

With general availability set for the fourth quarter, the AI Toolkit for IBM Z and LinuxOne will provide support for industry standard open-source AI frameworks such as IBM Z Accelerated for TensorFlow, IBM Z Accelerated for TensorFlow Serving and IBM Z Accelerated for Snap ML, enabling mainframe users to start implementing trusted AI on IBM Z. Meanwhile, the Python AI Toolkit for the IBM z/OS mainframe operating system is available now and will provide developers with access to a library of open-source Python software to support AI workloads that adhere to IBM’s Security and Privacy by Design practices, the company said.

IBM said the idea with the two toolkits is that developers will be able to connect mainframe data and applications to the most widely used AI frameworks and packages.

“These tools include frameworks and libraries that are optimized and supported for IBM Z and they are built to allow developers to start implementing trustworthy AI capabilities on z/OS,” said Elpida Tzortatos, an IBM Fellow and chief technology officer of z/OS and AI on IBM Z and LinuxONE.

Also available in the coming months is an update to Machine Learning for IBM z/OS. This is IBM’s full lifecycle machine learning platform for building, deploying, managing and putting AI models into operation on z/OS. The update will enable users to access IBM’s watsonx.ai, which is a studio of integrated tools for building machine learning and generative AI models.

In addition, IBM’s Cloud Pak for Data, which is a modular set of integrated software components for data analysis, organization and management, is also going to be made available on Z platforms, enabling users to automate the process of building AI models, the company said. The useful thing here is trust, as companies will be able to upload their own data, select a problem type and specify constraints on the models they’re developing.

Finally, IBM said, it’s launching an AI-infused update to IBM z/OS itself.Set to arrive on Sept. 29, the IBM z/OS 3.1 update will feature new AI System Services that can help the operating systems to learn and predict how to optimize information technology processes, simplify management, improve performance and reduce the skill requirements for implementing AI.

IBM customer La Mobilière Suisse Holding SA, the oldest private insurance firm in Switzerland, said it has relied extensively on IBM’s Z mainframes for years to process high volumes of transactions in the most secure way. Recently, it partnered with IBM to create new AI models that enabled it to uncover patterns in its data and improve its predictive capabilities.

“We worked together with IBM to apply the AI capabilities on their trusted IBM z16 systems to process insurance and offer recommendations faster and more accurately,” said La Mobilière’s IT architect Thomas Baumann. “By unlocking hidden data patterns with NLP-based AI functions in near-real time while ensuring privacy and security, we saw 94% accuracy in prediction results.”

Today’s updates are just the latest as IBM strives to enhance its mainframe services with AI. In August, the company launched watsonx Code Assistant for Z as part of an AI-powered mainframe application modernization push. The service uses generative AI to translate legacy mainframe apps written in the old COBOL programming language into Java, making them more accessible to developers.

IBM watsonx Code Assistant for Z is promising because COBOL developers are hard to find these days, with many of those who are skilled in the language having retired from the workforce.

Image: IBM