Professional buyers often know the task, material, or budget they have in mind but not the exact catalogue term or SKU. Traditional keyword search makes them translate intent into the store taxonomy before comparing suitable products. SETrade needed one released platform for the complete human buying journey while making the catalogue understandable to AI systems. Recommendations had to resolve to real products, respect explicit budgets, work in Ukrainian, and remain useful when AI was unavailable. Autonomous checkout and payment remained outside the released scope.
AI-Powered Dental E-commerce Platform with Agentic Product Search and Intelligent Matching
SETrade is a Ukrainian specialist retailer for dentists, clinics, laboratories, dental technicians, and milling centers, offering professional materials, instruments, equipment, 3D-printing and CAD/CAM products.
Country:
Ukraine
Services:
Industries:
Technologies:
Whether you're modernizing existing operations or building something entirely new, Clover Dynamics gives you the AI-native engineering support needed to move faster, automate deeper and stay ahead
Want the complete technical story? Download the full case study.
The Challenge
Accelerate ahead of competitors with machine customer capabilities, agentic commerce and AI-native systems built from the ground up for autonomy and scale
The Solution
Clover Dynamics delivered a modular commerce platform with AI embedded directly into product discovery. Buyers describe a need in natural language and receive a ranked shortlist of catalogue items with prices, match indicators, and concise explanations.
The storefront sends validated requests to MCIP through MCP over Streamable HTTPS. MCIP normalizes Vendure catalogue data, combines semantic search in Qdrant with lexical recall and deterministic reranking, and applies only explicit catalogue and price constraints. Returned items are reloaded from Vendure before display, keeping current product data under the commerce backend authority. Users retain explicit control of cart, delivery, checkout, and payment, with layered fallbacks preserving search if vector retrieval or MCIP fails.
The Features
Setrade — Natural-Language Product Assistant
Setrade — Complete Web and PWA Commerce Journey
Setrade — Grounded Recommendations
Setrade — Connected Content and Commerce Operations
The Result
- 01
Buyers can begin with a clinical need or product problem and reach relevant catalog options without knowing the exact product category.
- 02
Recommendations stay grounded in real products and include clear match explanations instead of unsupported AI suggestions.
- 03
The business gains a responsive web and PWA experience that connects assisted discovery with comparison, cart, checkout, accounts, and content.
- 04
The storefront, commerce source of truth, retrieval services, and MCP tools are separated for maintainability, reuse, fallbacks, and controlled evolution.
Full case study
More Case Studies
AI Assistants for Automating Follow-Ups Workflow
Platform that helps replace Google Sheets for SME businesses and helps financial analytics and financial specialists to build economic models effortlessly, connecting them to the company’s data. Run scenarios, manage cash and burn, and prepare for fundraising.
AI-Assisted Web Platform for Product Allocation Across Retail Stores
A Ukrainian household-goods retailer operating a central warehouse and a network of stores needed a consistent, explainable way to allocate products across locations.
