IBM’s research in Artificial Intelligence dates back to 1950s; but it reached a significant milestone when their supercomputer Deep Blue defeated Russian chess grandmaster ‘Garry Kasparov’ in 1990s. Later on, in 2010 IBM came up with Watson - a computer system named after Thomas J. Watson, IBM's founder and first CEO. It was capable of answering questions posed in natural language. In May 2023, IBM revealed Watsonx – a commercial Generative AI and scientific data platform based on cloud; an AI tool which can be tailored for enterprises as per their requirements.
Businesses across the globe had several failed attempts, crude and flawed journeys of mainframe modernization, primarily due to the void of skilled mainframe resources and a foolproof strategy. To clear up the mess, IBM took a step back and smartly came up with a plan that can help customers to easily understand their existing, mission-critical business applications which aren’t well documented in certain cases. In 2023, IBM added another jewel to their crown - IBM Watsonx Code Assistant. It leverages generative AI to accelerate development while maintaining the principles of trust, security and compliance at its core. Now, programmers and IT Operators around the world could speed up application modernization efforts and generate automation to rapidly scale IT environments. With the recent availability of IBM Watsonx Code Assistant for Z, enterprises can now leverage Gen AI as they seek to modernize their mainframe applications and adopt a hybrid cloud strategy. It also provides the flexibility to deploy its gen AI based capabilities for either on-premises or as-a-service architecture.
IBM Watsonx Code Assistant for Z is not like other Generative AI tools which just throw the COBOL statements to the large language models (LLM) and expect them to be converted to JAVA. This new competence enables an end-to-end, AI-assisted modernization experience for z/OS COBOL applications with tooling for each step of the journey. The product follows the modernization lifecycle with an application discovery capability, which maps out a technical understanding of the application and its interactions with other systems. Then, an automated refactoring capability leverages the information captured in application discovery to identify selected elements to re-model the monolithic application into modular COBOL business services. It then leverages generative AI to transform individual COBOL business services into object-oriented Java code. Next, it does is the validation of the outcomes, making the entire transformation process a test-driven development.
IBM Watsonx Code Assistant for Z even works beyond COBOL programs; for other z/OS sub-systems like IMS, CICS and others. IBM has pulled in a huge number of experts and full stack developers writing semantically equivalent COBOL and Java code pairs, who knows the subject on both the sides and that actually makes the product a unique fit in market. These experts are continuously progressing to make their curated LLM models better others with each growing day.
The entire AI assisted translation process is developed to make life easy for the open-source developers and boost their productivity based on natural language processing inputs, without forcing them to become COBOL experts. This helps organizations to reduce their total cost and complexity of mainframe modernization initiatives by using robust and AI driven automation. For example, a leading US-based property and casualty insurance company sought to integrate generative AI capabilities to modernize their legacy application and managed to reduce 80% of time for a developer to understand the application and 30 % time to be able to explain and potentially document code.