IBM has previewed its upcoming watsonx Code Assistant for Enterprise Java Applications at its annual Think conference.
The generative AI-based code assistant is aimed at accelerating Java application lifecycle with capabilities such as code generation, code explanation, and test generation, underpinned by IBM’s open source Granite family of large language models (LLMs).
Some of the capabilities of the new code assistant, targeted at improving developer productivity, includes navigating complex code structures with the help of generative AI to summarize an application’s key functions, services, and dependencies.
The assistant can also describe the changes needed to upgrade, modernize or enhance an application, with a detailed assessment of complexity and required development effort, the company said.
Additionally, the code assistant can be used to implement code and configuration changes while also documenting them, it said.
Enterprises can also use the assistant to import existing unit tests easily and use generative AI to create new tests that help maintain critical application functions.
Last August, IBM expanded the capabilities of its code assistant to include COBOL code translation into Java in order to help IBM Z systems customers modernize their applications.
IBM has a separate watsonx Code Assistant for generating content for its Red Hat Ansible Automation Platform. This is expected to enable developers to write Ansible Playbooks with AI-generated recommendations, the company said.