MTS AI Commerce Assistant

AI Sales & Product Recommendation Engine for Adobe Commerce / Magento 2

MTS AI Commerce Assistant is a modular, scalable Magento 2 extension that delivers personalized product recommendations using machine learning, vector embeddings, and behavioral analytics. It supports OpenAI as the primary AI provider with graceful fallback to rule-based algorithms.

Package

magetech/module-ai-recommendation

Requirements

Magento 2.4.7+ · PHP 8.2+ · MySQL 8.0+ · Redis

Scale

169 source files · ~15,000 LOC

Technology Stack

7 AI-Powered Recommendation Engines

5 Placement Zones

• Homepage — Recommended For You (6 products)
• Category Page — You May Also Like (4 products)
• Product Page — Similar Products + Frequently Bought Together (6 products)
• Cart Page — Frequently Bought Together (4 products)
• Checkout Page — Complete Your Look (3 products)

Database Schema

6 custom tables with foreign key constraints and indexed columns for query performance:

• magetech_ai_recommendation — Core recommendation entity with confidence scores
• magetech_ai_vector_embedding — 1536-dimensional product vector embeddings
• magetech_ai_tracking_event — High-volume behavioral tracking events
• magetech_ai_prompt_template — Configurable AI prompt templates per engine
• magetech_ai_recommendation_analytics — Impression, click, purchase analytics
• magetech_ai_customer_preference — Customer behavioral preference profiles

API Support

• GraphQL — 3 queries + 1 mutation for headless commerce
• REST — GET /V1/magetech/recommendations/:placement, POST /V1/magetech/tracking
• AJAX — Frontend tracking and widget endpoints

Design Patterns

• Factory — Maps engine type strings to concrete classes
• Template Method — Abstract engine defines skeleton, subclasses implement algorithm
• Strategy — 7 interchangeable algorithms behind common interface
• Repository — 4 repositories with SearchCriteria/SearchResults
• Observer — 6 observers for behavioral tracking
• Plugin / Interceptor — 4 after-plugins on Magento core classes
• Publisher-Subscriber — 3 message topics via AMQP

Performance

• Redis Caching — Engine + service layer caching, reduces AI API calls by 90%+
• Async Embedding — Message queue for non-blocking product saves
• Batch Cron — 50 products per batch, every 6 hours for embedding generation
• Cache Warming — Pre-fetches all 5 placements every 30 minutes
• Data Cleanup — Removes tracking events >90 days, analytics >180 days

Theme Compatibility

• Luma / Default — PHTML + RequireJS + jQuery (out-of-the-box)
• Hyva — Alpine.js + Tailwind CSS (full compatibility)
• PWA Studio — React + GraphQL (via API)

Admin Panel

• 12 admin controllers with full CRUD
• Analytics dashboard with CTR and conversion reports
• Engine toggle controls (enable/disable each engine)
• Prompt template editor with test functionality
• Cache management with one-click flush
• Full ACL hierarchy with granular permissions