IBM has showcased a range of innovations in generative AI at its annual Think conference, demonstrating how these technologies are set to fundamentally transform automation across various business functions. According to IBM Research, these advancements aim to simplify and enhance efficiency in work processes.
Reimagining How the World Codes
At Think 2024, IBM introduced groundbreaking tools designed to streamline software development. Among these is the open-sourced Granite code model, which offers capabilities such as bug fixing, code translation, and detailed explanations. By making these tools accessible, IBM aims to support both legacy and modern systems, ensuring long-term sustainability for developers.
Additionally, IBM announced the watsonx Code Assistant for Enterprise Java Applications. This new assistant can summarize existing Java code, make recommendations, execute upgrades, and generate unit tests. According to IBM Research, the automated test generator framework can increase testing coverage by up to 50%, making it a valuable asset for developers.
Diagnosing and Addressing IT Issues
IBM is also focusing on automating IT operations to enhance system reliability. Site reliability engineers (SREs) are often overwhelmed by the complexity of modern IT systems. IBM’s updates to its IT automation portfolio, including intelligent remediation for Instana and GPU optimization for Turbonomic, aim to alleviate these challenges. The new generative AI technologies can summarize IT issues, identify probable causes, and recommend actionable solutions.
The trace-based reinforcement learning technique applied for probable cause analysis has shown significant improvements, increasing the true positive rate by a factor of 1.6 and reducing false positives by a factor of 200 compared to traditional tools.
The Future of Manufacturing
In the manufacturing sector, IBM’s generative AI is poised to revolutionize plant operations. The Maximo Application Suite now features automated work order intelligence, which predicts work order failure codes using models that generate synthetic data. IBM estimates that this could save approximately 10,000 hours of productivity annually.
IBM’s open-source Granite time-series models are another significant advancement. The tiny time mixer (TTM) model, based on an encoder-decoder architecture, has shown a 3% to 40% accuracy improvement over other state-of-the-art models. This model enables better continuous monitoring of assets and processes, reducing false positives and enhancing operational efficiency.
Automating Office Work
IBM is extending the benefits of automation to everyday office tasks. The watsonx Orchestrate Assistant Builder, unveiled at Think 2024, helps businesses create assistants tailored to specific tasks in various departments such as HR, sales, and procurement. This tool allows for the quick building, enhancement, and validation of automation skills, transforming the user experience.
Generative interfaces for automation tasks are also in development, enabling users to describe their needs in plain English. These systems can invoke APIs, sequence multiple APIs, query databases, and summarize retrieved content using IBM’s Granite LLM. Early adopters like Sports Clips have already seen significant reductions in time to execute HR tasks, from hours to mere minutes.
The advancements presented at Think 2024 highlight IBM’s commitment to leveraging generative AI to enhance business automation. As the technology continues to evolve, it promises to bring about unprecedented efficiencies and capabilities across various sectors.
Image source: Shutterstock
. . .
Tags
