Agents are Polarion users
Each agent signs in as its own Polarion user. It sees only the projects it was hired into, and everything it writes is authored by it — like any colleague.
ALM Agents
Hire AI agents into your Polarion projects. Give them work the Polarion way. Keep every decision in human hands — and every step on record.
A team that works inside Polarion
Section titled “A team that works inside Polarion”Work the Polarion way
Assign a work item, and the agent starts. It reads documents, creates requirements and tests, links them, asks you when something is unclear, and submits its result for your approval.
You decide what matters
Changes to work people wrote, writes in connected systems such as GitLab, and every result wait for a person. Answer, approve or send back — right from the board.
Everything on record
Every run is a work item: model, cost, what it read, created and changed, which tools it called. The AI audit report is one click away.
What makes it different
Section titled “What makes it different”Agents learn from your feedback
When you send work back, an agent can turn your feedback into a project memory. You edit and approve it — and from then on the whole team gets it right the first time.
See how an agent thinks
The Agent Brain shows an agent’s skills, memories, permissions and track record — and the exact instructions it receives.
Connected, under policy
GitLab, GitHub and other systems join through an MCP gateway. You review each server’s tools once: read, ask first, or never.
Its own computer, no secrets
Coding agents work in an isolated sandbox per task — with no credentials inside and read-only internet limited to sites you approve.
Where to go next
Section titled “Where to go next”ConceptsAgent Pool, engagements, tasks, runs, memories — in five minutes.
Working with agentsHire, assign, answer, approve, teach.
GovernanceHow oversight works end to end.
ArchitectureComponents, data flow and security.