Help center
How to sell your product catalog with an AI agent
With your catalog loaded, the TBit agent recommends products, answers about price and stock, sends photos and builds the cart right in the conversation. How to load products that actually sell.
5 min read

What it does
With your catalog loaded, the agent recommends products, answers about price and availability, sends photos and builds shopping carts inside the conversation.
How it works
The agent searches the catalog by meaning: if the customer asks for "something to gift my mom", it finds products even if none is named that. It recommends with photo, price and description, and builds the order with what the customer confirms.
How to load products that sell themselves
Clear name and detailed description: material, sizes, uses, who it's for. The description is what the agent uses to recommend.
Good photo: the agent sends it in the conversation.
Up-to-date price and availability: the agent answers with what it sees in the catalog.
Categories: they help the customer browse and the agent narrow things down.
Bulk import
You can import the whole catalog from a CSV or Excel file with columns for name, price, description, category and image. Useful for large catalogs or migrations from another platform.
The TBit catalog first
The Store catalog is part of the engine, not an accessory: search by meaning, recommendations, photos in the conversation and carts all work directly on it, without depending on external systems. It's the recommended option for selling with your agent. If you already sell on Shopify or Jumpseller, there are integrations that sync your products so you don't upload them twice, but the main path is the TBit Store: fewer pieces, fewer failures, and the agent always sees up-to-date inventory.
Where to set it up
Store menu: create and edit products, categories and import the file. The orders the agent builds appear in Operations.
What NOT to put in the prompt
Product lists with prices. They go out of date and clash with the real catalog. The prompt can define your sales style ("suggest add-ons", "offer the combo first"); the data lives in the Store.
Train with your own AI
Do you use ChatGPT or Claude? Connect them to your TBit agent and let them help you train it with your real data: on the platform go to Agent settings, Connect your AI section, generate your token and add the connector https://api.tbit.app/mcp. Your assistant can review the prompt, the memories and recent conversations, and suggest improvements.