Back to case studies

Online Fashion Shop with AI-Generated Listings

  • Overview
  • Challenges
  • Services
  • Functional Modules
  • Technologies
  • Our Achievements

Client

An online fashion shop with a growing catalog and a team of sellers who keep adding new items.

Technologies

React / Next.js / TypeScript / Python (FastAPI) / PostgreSQL / pgvector / Redis / Telegram Bot API / LLM API / AWS / Docker / Terraform / Payment Gateway APIs

Adding a product meant writing the title, description and attributes by hand, and updating stock meant opening the admin panel on a computer. Every shopper saw the same catalog in the same order. The client wanted AI that speeds up listings and personalizes shopping, and a way for sellers to work from their phones.

Parallax effect image showing a scenic background

Challenges

  • Slow Listings: Each new item needed a title, description, category and attributes written by hand.
  • Seller Workflow: Sellers who work on the move needed to update stock and listings from a phone.
  • Generic Catalog: Every shopper saw the same products regardless of taste and history.
  • Customer Questions: Questions about sizes, availability and orders arrived at all hours.
  • Smooth Checkout: Payment and delivery options connected to the order flow.
  • Growth: An architecture that grows with the catalog and the traffic.

Services

  • Business Process Audit and Solution Design

    We talked to the shop's sellers and support staff and mapped how a product gets from an idea to the catalog. It showed which tasks take the most time: writing product cards, updating stock and answering the same customer questions. We agreed the scope and the order of work on that basis.
    Read more
    Business Process Audit and Solution Design

  • End-to-End Software Development

    We built the storefront and the admin panel as one product. Shoppers get a catalog with filters, product pages with sizes, colors, reviews and specifications, a cart and checkout. Staff manage products and orders in the admin panel.
    Read more
    End-to-End Software Development

  • AI Listing Generator

    A seller describes a product in plain words, for example from a phone. AI prepares the title, description, category and attributes such as material, fit and care instructions. The seller reviews the draft, edits it if needed and publishes it.
    Read more
    AI Listing Generator

  • Telegram Bot for Sellers

    Sellers manage the shop in a chat on their phone. They add new items, change stock and prices, and mark sold-out products without opening the admin panel. The changes appear in the shop right away.
    Read more
    Telegram Bot for Sellers

  • AI Recommendations and Assistant

    Recommendations are based on what a shopper browses and buys, and appear on the home page, product pages and cart. The AI assistant answers questions about sizes, availability and orders right in the chat and passes complex cases to staff.
    Read more
    AI Recommendations and Assistant

  • Payment and Delivery Integration

    We connected payment gateways and delivery services to the order flow. A customer pays and chooses delivery at checkout, and the order status updates along the way, so both the customer and the staff can follow it.
    Read more
    Payment and Delivery Integration

Functional Modules Delivered

  • Product Catalog

    • Catalog with content management for products, photos, colors, sizes and prices.

    • Filters by category, price and color, and search.

    • Product pages with rating, specifications and reviews.

  • AI Product Listing

    • A plain description becomes a complete product card.

    • Every draft is checked by the seller before it goes live.

    • New items reach the catalog much faster.

  • Seller Bot

    • Sellers add items, change stock and mark sold-out products from a phone.

    • Prices and new listings are handled in the same chat.

    • The bot works in Telegram, and the shop updates right away.

  • Personalized Recommendations

    • Suggestions from browsing and purchase history.

    • They appear on the home page, product pages and cart.

    • Each shopper sees a catalog that matches their taste.

  • AI Shopping Assistant

    • Chat that answers questions about sizes, availability and orders.

    • Product details are shown right in the conversation.

    • Complex cases are handed to staff.

  • Orders and Checkout

    • Order management with status tracking.

    • Customers choose payment and delivery at checkout.

    • Payment gateways and delivery services are connected to the order flow.

Technologies used

  • Frontend:

    React and Next.js with TypeScript.

  • Backend:

    Python with FastAPI and PostgreSQL.

  • AI:

    An LLM API for listings and chat, and pgvector for similarity search in recommendations.

  • Seller Bot:

    Telegram Bot API connected to the shop inventory.

  • Integrations:

    Payment gateway and delivery service APIs.

  • Cloud:

    AWS with Docker and Terraform.

Our achievements

  • Faster listings

    New products go live much faster with AI-prepared product cards.

  • Stock from anywhere

    Sellers update inventory from their phones, which keeps the shop accurate.

  • More relevant shopping

    Recommendations show shoppers items that match their taste.

  • Support around the clock

    The assistant handles routine questions at any hour and staff take over complex cases.

  • Smooth checkout

    Customers pay and choose delivery in a few steps.

  • Room to grow

    The architecture supports a larger catalog and more traffic.

Our achievements

Let’s talk

Have any questions or feedback? Send us a message!

Our contacts

View of Barcelona