Skip to main content

What is Nox

Audience: Sales · Management · Support · QA · Engineering · Where in app: the Nox bubble (bottom right, Cmd/Ctrl+K) → docked window with Call / Chat, full-screen Nox Live, the morning plan and Daily Brief · Plan availability: On for every company (no super-admin gate since 2026-08-20; settings-by-chat and Nox Activity since 2026-09-25; the launch-week NOX_ONLY_COMPANY_IDS test gate was removed 2026-11-03). What Nox can do follows the tenant's features — see feature awareness

Nox is the AI assistant the customer's own team talks to inside the Knock Knock dashboard. Where the widget's AI talks to visitors, Nox talks to the admin. You ask it questions in plain English — "What did John do on the site this week?", "How many of yesterday's leads came from Google Ads?", "Walk me through what someone did before they booked." — and it reads across visitors, sessions, chats, calls, bookings, lead scores, and the connected CRM to answer, citing its sources.

Naming. Customer-facing it is Nox. The codebase calls it Knox (every table, tool, and service is prefixed knox_/Knox). They are the same agent — when you see "Knox" in logs or code, read "Nox".

Two engines. Since the 2026-11-02 "NOX v3 (upgraded v2)" launch every company runs on the v2 engine (planner → researcher/writer subagents → guards, plus proactive triggers and the morning day plan). The intent-classifier pipeline described under How it works is the v1 fallback, used when GET /knox-v2/status is false or fails; v2 still reuses its classifier and knox_* tools.

Engineering deep-dive: this page is the functional/reference version. For the request-flow internals, sub-agent wiring, and memory architecture see Nox / Knox Admin Assistant and ms-ai.

What it does​

Nox started read-side first and now presents itself as an "AI sales manager": it plans the admin's day (the morning hunt), chases quiet deals, builds automations, coaches the customer-facing AI agents, reads the connected business mailbox, and runs work through a team of subagents — but anything that reaches a person still waits for the admin's OK (knox.actions.autonomous_send was retired 2026-11-01; saving settings writes it false). With the right entitlements it can also take a small set of pre-approved actions (place or schedule an outbound call, draft a follow-up, enrich a CRM contact, set a return-visit watch; on Nox funnel tenants: tag a contact, move a GHL pipeline card, trigger a workflow, book a meeting, mark a lead Won), but every customer-facing message goes through a review step unless the account has explicitly opted into autonomous send.

The launcher bubble opens the docked window (KnoxAssistant.vue, home view NoxHomeView.vue with Call / Chat); its expand icon opens Nox Live, the full-screen surface (NoxLive.vue, redesigned in 90536353): a voice area on the left and, on the right, the Today home (tiles: Drafts to review, Scheduled, Open in today's plan; starters; today's plan) that becomes the canvas once tool results arrive. The panel's Inbox holds threads, followed cases and Waiting for you (parked cards/questions). The Standup pill is retired (live/legacyStandup.js LEGACY_STANDUP_ENABLED = false) — the morning day plan replaced it. The brief never auto-plays: a brief-ready badge appears once a day (the per-device auto-play preference was removed). Voice calls: see Voice.

Typical questions Nox answers well today:

  • "Did John visit my site again after our meeting?"
  • "Walk me through everything John did last night."
  • "What questions did Sarah ask in her last conversation?"
  • "Of yesterday's 3 bookings, what was each person's journey?"
  • "How many of yesterday's leads came from Google Ads vs organic?"
  • "Where do I change my working hours?" (and it can offer to navigate you there)
  • "What can you actually do?" / "How do you know all this?"

For the full done/partial/future breakdown, see Capabilities.

How it works​

(v1 fallback — see v2 engine for the default path.) Every v1 question runs through the same pipeline:

  1. Hydrate prior turns. The conversation's earlier turns are pulled from the database (the source of truth), so Nox has the thread's context.
  2. Intent classification. The question is forced into exactly one intent: navigate, how_to, explain, analytics, compare, troubleshoot, action, draft_followup, greeting, or unknown. The classifier also returns a confidence score and, if the question is ambiguous, one clarifying question instead of guessing.
  3. Routing to a specialist sub-agent based on intent (see table below). Low-confidence or unmapped intents fall through to the generalist, which carries the full toolset.
  4. Tool selection. The specialist invokes one or more of 42 tools (see Tools) — visitor/session lookups, analytics, knowledge search, CRM queries, funnel reads/actions, etc.
  5. Synthesis. The LLM composes a natural-language answer, citing knowledge-base chunks where relevant with markers like [S1], [S2].
  6. Action card (optional). If the answer suggests navigating somewhere or sending something, it renders as a clickable card the admin confirms.

Intent → sub-agent routing​

IntentSub-agentHandles
navigateNavigation"take me to / open / show me the X page"
how_to, explain, compare, troubleshootKnowledgeproduct docs, how-tos, concept questions (RAG-grounded)
analyticsAnalyticsvisitor / session / metric / cohort questions
actionActionside effects: place call, enrich contact, propose navigation
draft_followupOutreachcompose a customer-facing email/SMS as a reviewable draft
greeting, unknown, low-confidenceGeneralistsmalltalk + catch-all with the superset toolset

The confidence threshold is 0.65: anything below it routes to the generalist rather than risk committing to a wrong specialist whose tool allowlist might not include what the question needs.

Configuration & options​

Most Nox behavior is controlled by the companies.knox.* config block (see Company Configuration):

  • knox.enabled — default true, no UI writes it; scopes background jobs only (Nox itself is on for every company).
  • knox.target_market — context the assistant uses when framing answers.
  • knox.tracked_pages[], watchlist_companies[], high_value_pages[] — what the tenant cares about.
  • knox.alert_thresholds.* — thresholds that flag anomalies in the Daily Brief.
  • knox.daily_brief.* — the morning brief: enabled, send_hour (default 9; funnel tenants 5), sections.* (see Daily Brief). knox.digest_schedule still exists but is legacy.
  • knox.baseline_metrics — nightly rollup of "what's normal" for this tenant.
  • knox.wake_voice — how proactive moments surface (quiet default / speak / off), overridable and lockable by the agency (agency.nox_wake_voice, nox_wake_voice_locked).

Identity (name, voice, persona, welcome message, look) is per company or set by the agency (agency.nox_name, nox_voice_id, nox_persona, nox_welcome_message, nox_look, and a super-admin-set own character nox_avatar ∈ chirpy | bee | nexus | chatsmart; backend/app/lib/noxIdentity.js). Agency clients may only change their outfit; it defaults from the onboarding trade (TRADE_OUTFIT: medical family front_desk/doctor/dentist/chiro/physio, home family dispatcher/electrician/builder/plumber/landscaper). Agencies request an own character from Agency settings → AI Assistant, which opens an ops ticket (type: 'avatar', label "Own character"). Per-user style (call_me, talk rules, greet_listen) lives in nox_rapport.

Per-tenant LLM/model config lives in knox_ai_models; custom behaviors in knox_tenant_config.

Behaviors & edge cases​

  • Tenant isolation is strict. Every tool scopes its query to the caller's company_id. Nox never sees another tenant's data.
  • Honesty about limits is part of the brand. If a question needs infra that isn't built (e.g. predictions, a B2B account entity, a CRM-stage join), Nox explains why and what it would need, rather than fabricating. See Capabilities.
  • Memory — durable per-admin facts (knox_remember/knox_recall, pgvector knox_admin_memory), plus on v2 a rulebook in the admin's words (nox_rule), a versioned business profile, recent-work/call memory and rolling thread summaries. Approved product notes ride with every page guide.
  • Presence is a snapshot. "Who's on the site right now?" reflects the moment of the query; it may change by the time you read the reply.
  • Citations can have false positives. A cited claim should be verified before acting on it (see Citations & Feedback).

Plan & limits​

  • Nox is on for every company; knox.enabled only scopes background jobs. Nox is told which features the tenant has and refuses unavailable ones with the real remedy (plan vs account manager).
  • Every question consumes AI credits / tokens; usage is budget-tracked in knox_usage_events (tool name, input, output size, duration, tokens, cost USD).
  • Action tools (place call, send follow-up) are individually entitlement-gated on top of the base Nox entitlement.
  • Nox does not make per-visitor predictions, never contacts a person without approval (autonomous send is retired; card-free action on Nox's own judgement needs KNOX_V2_GUARD_JUDGEMENT=on, off), and does not see data from before the company joined. Reversible changes are applied and reported with Undo (v2 autonomy auto).

Technical implementation​

  • v2 (default): ms-ai src/knox-v2/, entry POST /knox-v2/ask (SSE, BullMQ knoxV2RunQueue) — see v2 engine.
  • v1 (fallback): ms-ai src/knox-agent/. Orchestrated by knox-agent.service.ts; entry point POST /knox-agent/ask (streaming SSE).
  • Intent classifier: intent/knox-intent-classifier.service.ts (uses generateObject + a strict Zod schema — intent/knox-intent.types.ts).
  • Sub-agents: agents/knox-{navigation,analytics,knowledge,action,outreach,generalist}.agent.ts; routing in agents/knox-agent-router.ts.
  • Tools registered in tools/knox-tool-registry.ts; each tool is OTel-traced as knox.tool.<name> (visible in Grafana) and logged to knox_usage_events.
  • Conversations persisted to knox_agent_conversations (Postgres, ms-ai). Knowledge chunks in knox_knowledge_chunks (pgvector). See Data Stores.
  • Voice is a separate module, src/knox-voice/ — an ElevenLabs agent that delegates to v2 (see Voice).