Task Agents
Helping maintenance teams hand routine work to AI agents they can trust inside IBM Maximo.
Timeline
May 2026 – August 2026
Team
1 Product Manager, 1 Designer
Role
Product Design Intern
Skills
AI interaction design, Agentic UX, Prototyping, Design systems
What is IBM Maximo?
Maximo is IBM's asset management platform. Organizations use it to monitor and maintain physical assets, like plant equipment and infrastructure, and to plan and track the maintenance work around them.
TASk Agents project requirement
During my internship, my manager asked us to design task agents for Maximo: AI agents that handle routine maintenance work for users. This covered how people find, build, review, and trust an agent, and what engineering would need to build it. Our goal was to create a concept that we could continue iterating on after our internship and launch in 2027.
01
Context
IBM’s 9.3 roadmap moves the Maximo Assistant from answering questions to taking action. Task agents do that work on a user’s behalf.
02
Why it matters
By implementing task agents, users would be able to save hours every day during routine tasks.
03
Launch
Our goal was to create a concep to set this up for Maximo's 9.3 update in the following summer 2027.
The problem
The current Maximo AI assistant could answer questions, but it couldn't take action on a user’s behalf. Supervisors still drafted routine work orders by hand, and the software couldn't chagne real records on its own.
The Operational dashboard with the Maximo Assistant today. Image: IBM
How can users save time with task agents, while still managing all of them?
Business impact
We estimated that the briefing and task agents could save each user 55+ minutes a day, depending on the role. That adds up to 230+ working hours per user each year, or about $12.6K in value per user. By making routine workflows reusable, task agents give that time back to organizations every week.
“You actually did a project during your internship that is going to be implemented into our product in the next cycle.” — VP of Asset Lifecycle Management, IBM
Next up