N Nexus Docs
Intelligence

Intelligence

AI-powered automation for pricing, analytics, advertising, reviews, and workflow orchestration.

Overview

The Intelligence layer adds AI-powered automation to the operational foundation. These features transform NexusCommerce from a multi-marketplace management tool into an AI-powered commerce platform that actively optimizes your business.

Intelligence features require:

  • At least one active marketplace connection
  • The Python FastAPI worker fleet running
  • An OpenAI API key (for AI Studio and review sentiment)
  • ClickHouse running (for analytics)

Key Concepts

AI Workers — Python FastAPI services that execute computationally intensive AI tasks asynchronously. Workers are dispatched by the NestJS API, execute against marketplace data and external APIs, and write results back to Supabase. The NestJS API polls job status and notifies the frontend on completion.

Job Queue — Jobs are queued in a Bull/BullMQ queue backed by Redis. The queue handles job prioritization, retries, and dead-letter handling for failed jobs.

ClickHouse — A columnar analytics database optimized for high-volume read queries. NexusCommerce ETLs order, inventory, and pricing data into ClickHouse for sub-second analytics queries that would be too slow against Supabase (PostgreSQL).

Buy Box — Amazon's (and some other marketplaces') mechanism for selecting which seller's offer appears as the default purchase option. Winning the Buy Box dramatically increases sales volume. NexusCommerce's pricing engine is designed around Buy Box win probability as a key optimization target.

Getting Started

Intelligence features are available once the Foundation is configured. No additional onboarding steps are required, but recommended first actions:

  1. Pricing — Create floor and ceiling price rules for your top 20 SKUs before enabling auto-repricing
  2. Analytics — Review the ClickHouse dashboard to understand your current sales distribution
  3. AI Studio — Run a pricing_recalculate job on a sample of SKUs to validate the pricing engine
  4. Reviews — Verify that review data is importing for your Amazon ASINs

Features

Pricing Engine

Rule-based and AI-driven pricing with competitor monitoring and Buy Box optimization. Supports floor/ceiling rules, competitor match strategies, and fully autonomous AI repricing.

Pricing documentation

Analytics

ClickHouse-powered analytics dashboard with revenue trends, order distribution, category performance, and marketplace comparison. Sub-second query times even on 1M+ order datasets.

Analytics documentation

Ad Automation

Cross-marketplace PPC campaign management. Create, manage, and optimize ad campaigns on Amazon Sponsored Products, eBay Promoted Listings, and other supported ad platforms from a single interface.

Ads documentation

AI Studio (FlowForge)

Visual workflow builder for composing AI job pipelines. Chain multiple worker jobs into sequences, set triggers (schedule, inventory event, pricing event), and monitor execution history.

AI Studio documentation

Review Analytics

Aggregate review data from all marketplaces with NLP sentiment classification, topic extraction, and trend tracking. Identifies patterns in negative reviews to guide product and listing improvements.

Reviews documentation

Configuration

Intelligence features share the following global settings under Settings > Intelligence:

SettingDescription
AI repricing modeManual (suggest only) or Auto (apply changes automatically)
Worker concurrencyMax concurrent AI jobs per worker instance
ClickHouse retentionHow many months of analytics data to retain
OpenAI modelGPT model used for AI Studio and sentiment analysis
Review import intervalHow often to fetch new reviews from marketplaces