IBM to launch “task-level” AI agents for engineers

IBM is preparing to launch IBM Engineering AI Hub v1.0 to help engineering teams work faster and more accurately.

The new product, which will be available from 14 October, will be available as an add-on to the company’s engineering lifecycle management (ELM) platform. It will introduce task-level AI agents directly into existing engineering tools as a “smart” feature, including DOORS Next for requirements management and Workflow Management for tracking and planning.

IBM touts that applying task-level AI agents will help address the risk of errors that often arise from general queries to large language models.

“Engineering teams are undergoing far-reaching digital transformation in complex engineering and product lifecycle management, and generative AI can help,” said Daniel Moul, IBM ELM Principal Product Manager, in a blog post.

“However, in engineering domains where accuracy and compliance are non-negotiable, and results must be correct, applying generative AI cannot be done naïvely. Even a single wayward requirement can lead to costly rework, schedule delays and low-quality outcomes that can risk product and sometimes company viability.

He added: “This is especially vital for safety-critical, regulated industries such as automotive, aerospace, or medical devices, where mistakes can impact product and business success. AI applications that lack enterprise-grade rigor or skip critical safeguards can undermine the promise of speed and automation they aim to deliver.”

First task for IBM AI agents for engineers

As part of its first release, the AI agents have been designed for use cases such as requirements quality analysis so that engineers can use agent-generated quality scores to address hidden risks among their requirement specifications.

Engineers will also be able to interact with their requirements using natural language to get answers to their questions, summaries, translations, and more, IBM said.

The company added the AI agents has been designed to quickly switch context, and can get up to speed even on long-running, complex tasks, defects and other work items.

IBM reassured sceptics that it is “keeping humans in the drivers’ seat” where AI recommends and people decide.

“Agents are promising, but they are not a panacea,” Moul said.

You might also be interested

• Google Gemini Enterprise enters race to become your company’s all-seeing AI
• What is Intel Xeon 6+ in Intel’s own words
• Exclusive interview with Kira Boyko, Product Line Director, on Intel Clearwater Forest: “You’re going to see a 750 kilowatt reduction and then 3.5x performance”
• Deloitte boondoggle highlights AI errors – and hype