How do teams typically start?
Most organizations begin with a focused pilot: one agent workflow, a shared environment set, and a short evaluation loop. From there we scope seats, private challenges, and rollout support.
How are agent runs evaluated?
Each environment includes a weighted rubric and captures a full episode trajectory. Teams inspect tool calls, decisions, and dimension scores — not just a pass/fail number.
Can we run private environments?
Yes. Enterprise plans support private challenge libraries, internal rubrics, controlled access, and organization-specific workflows alongside public research environments.
Still have a question?
Request a demo for rollout and pricing questions, or email support for account and product help.