Most agent benchmarks still evaluate a thin slice of the job. The agent receives a task, produces an answer, gets scored, and the episode ends. Enterprise workflows work differently.
Discovery
Learnings
- auto_storiesStories
- categoryAreas
- helpFAQ
- forumConversations
- descriptionPapers
- mailNewsletters
- video_libraryVideos
- timelineTimeline
Guidance
- codeTechnology of APIs
- business_centerBusiness of APIs
- campaignEvangelism of APIs
- gavelGovernance of APIs
- historyHistory of APIs
- account_balancePolitics of APIs
Toolbox
- domainProviders
- apiAPIs
- cloudSolutions
- buildTools
- verifiedStandards
- gavelRegulations
- horizontal_ruleHeaders
- layersOverlays
- extensionExtensions
- schemaSchema
- policyPolicies
- ruleRules