Previously, I talked about agentic AI doing work, not just answering questions. Here’s what that might look like in practice.
I wanted to combine my interest in AI, my project-management experience, and my curiosity about current development initiatives in Sarawak.
After a discussion with my AI agent, it suggested PCDS 2030. So I built a project tracker.
It tracks Sarawak’s PCDS 2030 development strategy. In version one, it covers seven projects across six sectors. Each project has milestones, status indicators, budget figures, and links to public sources.
Normally, a tracker like this would take quite some time to build: research projects, structure data, design the interface, write components, test, and iterate.
I didn’t do any of that directly.
I told the agent what I wanted. It researched projects from public reports and news outlets. It structured everything into clean data. It wrote the React code: project cards, milestone tracking, sector filters, and a progress dashboard.
Then I asked it to refine the work. Add progress bars. Make milestones expandable. Link every claim to a source. Each time, the agent made the changes and shipped them.
This is the shift: you stop building the thing. You start directing the build.
If you run projects, imagine this applied to your status reports, risk registers, stakeholder updates, and execution tracking.
The tools are ready to be useful.