Assistant Services

Arif's service layer turns a user message into a governed business interaction. The public behavior is simple: gather context, reason with the configured model, run approved tools, ask for approval when needed, and return a clear result.

Core Services

Context Grounding

Before acting, Arif gathers relevant business context such as permitted records, record structure, current page context, organization knowledge, and recent conversation history.

Tool Routing

Arif discovers available tools from Qriib and approved connectors. When multiple sources expose similar capabilities, Arif routes calls to the correct configured tool.

Safety and Approvals

Tool calls that can change data pass through safety policy. The policy can allow, ask for approval, or block an action.

Session and Memory

Sessions keep conversation history, task state, attachments, generated artifacts, and rollback information. Users can return to previous sessions when they need continuity.

Observability

Organizations can monitor assistant behavior, tool usage, performance, and failure patterns through configured operational tooling. Public docs describe the behavior, not the internal telemetry implementation.

Integration Connectors

Connectors can extend Arif with approved business systems such as Drive, Helpdesk, communication platforms, project systems, reporting tools, and custom MCP servers.