Review · VerifiedUpdated 2026-05-03Verified against code 2026-05-03

The Nudg3 Loop

Nudg3’s value comes from a six-stage data flow that runs daily across every monitored workspace and loops back on itself via historical tracking. Each stage has its own doc.

[1] Collections & Prompts
       │
       ▼
[2] Mentions & Citations (per provider)
       │
       ▼
[3] Aggregated Stats
       │
       ▼
[4] Reports & Boosters
       │
       ▼
[5] Actions & Content
       │
       ▼
[6] Historical Tracking ──► loops into the next [4] Reports

The six stages

#	Stage	What it does
01	Collections & Prompts	Daily prompt × provider × workspace queries. Cloro queue, scheduler timings, provider list.
02	Mentions & Citations	Brand match detection, position, sentiment, key passages, source URLs, source classification. Per-provider parsing.
03	Aggregated Stats	Visibility score (0-100), audit score, share of voice, position premium, sentiment by provider. Deterministic scoring with unbranded queries only.
04	Reports & Boosters	Weekly Visibility Analysis reports + AI-generated Booster cards (opportunities, threats, quick wins, AI Model Strategy, Competitive Strategy).
05	Actions & Content	Booster → tracked Action → content brief → published URL → monitored for citations in future responses.
06	Historical Tracking	Week-over-week deltas, trend lines, change guards on methodology shifts and active-prompt rotation. Feeds back into the next report.

How to use this section

Sales/GTM: walk the loop top-to-bottom — it’s the story you tell prospects
Engineering: each stage maps to a clear subsystem in code (collection workflow, extraction service, scoring service, report workflow, actions API, change-guard logic)
Onboarding: read all six in order; you’ll have a working mental model in ~30 minutes

Cross-cutting capabilities (auth, multi-tenancy, agent, integrations, public API) live in platform/ — they’re not part of the loop, they support it.

Start Here 01 — Collections & Prompts