AI chatbot & storefront build

An AI consultant that finds the right book — on a storefront we built from scratch

Role
Storefront build & AI feature
Team
3 engineers · QA · PM
Stack
LLM + semantic search
An AI consultant that finds the right book — on a storefront we built from scratch

Overview

Knygarium is a Ukrainian bookstore with a catalog deep enough that shoppers who didn't know the exact title often gave up browsing. We built the store itself, then added an AI consultant that turns a mood or genre into real matches from stock.

The AI consultant sits on every page as a persistent widget rather than a separate "ask us" page, because the moment someone gives up scrolling categories is exactly when they need it. It reads a loose description, searches the real catalog semantically, and only ever returns books actually in stock.

What this covered

  • Full storefront build — catalog, checkout and a dozen-plus genre categories
  • AI consultant live on every page, wired into semantic search across the full catalog
  • Recommendation conversations convert to a purchase about 70% more often than plain browsing
  • Catalog organised into a dozen-plus genre categories, plus a live "New arrivals" carousel on the homepage
  • Telegram and Viber contact options alongside the storefront for customers who'd rather message than browse
Knygarium product page for a book, with cover, price and specifications
Knygarium homepage new releases carousel with the AI shopping assistant button visible