AI Review
Board view AI review → ticket tab AI draft. Read category, confidence, citations, and the draft. Accept, edit & send, reject, or re-screen.
Nexus is the multi-tenant hub for support tickets and work boards. Operators review AI drafts on live tickets, then curate gold replies into product knowledge so the next screen is smarter—not just faster.
Two jobs, two tabs: AI draft for live replies · KB review for product knowledge.
Nexus separates live AI draft handling from knowledge curation. Both show on the tickets board as dedicated views.
Board view AI review → ticket tab AI draft. Read category, confidence, citations, and the draft. Accept, edit & send, reject, or re-screen.
Board view KB review → ticket tab KB review. Grade the staff reply, assign product, pick a gold public answer, promote it into knowledge.
Promoted knowledge starts inactive (pending-curation).
Activate on Tickets → AI & knowledge so the next AI screen can cite better material.
From first ticket on the board to better drafts next week.
Open Tickets → AI review. Send or reject drafts so customers are not waiting on unreviewed AI.
Open Tickets → KB review. Score replies, set product, mark Done or Skipped.
Promote the best public reply into product knowledge (inactive until curated).
On AI & knowledge, activate snippets and adjust auto-send policy.
One foundation for humans in the SPA and agents on the audited HTTP surface—same domain rules, isolated per tenant.
Intake, email, board config, AI screening (heuristic or LLM), and tenant auto-send policy for safe categories.
Brands, products, workspaces, backlog, multi-board placements, lanes, custom fields, tags, tasks, and relations.
Opaque nxak_… keys bound to one tenant. Discover via GET /api/agent/about. Every call is audited.
Host, path, and header resolution with EF global filters. Branding tokens and settings are tenant-driven.
Start with the operator guide for AI Review and KB Review, then use Docs for setup, tenancy, and the Agent API.