# Introduction to Bookbag

> Bookbag is the AI customer-support platform for ecommerce — an AI agent grounded in your store's real data that resolves tickets across chat, email, and social. Start here.

Welcome to Bookbag. Bookbag is an AI customer-support platform built for ecommerce. You connect your store's knowledge — help center articles, product data, policies, past tickets — and Bookbag builds an AI agent that resolves customer questions instantly, 24/7, across every channel your customers use.

Unlike a generic chatbot, a Bookbag agent is **grounded in your real data**. Every answer is retrieved from sources you control, with citations, so the agent resolves *order status*, *returns*, *shipping*, and *product* questions accurately instead of guessing.

> **TRY BEFORE YOU SIGN UP:** You can spin up a working agent on your own website in under a minute with the [live trial](https://app.bookbag.ai/try) — no account required. Paste your URL, we crawl it, and you can chat with a grounded agent immediately.

## What you can build

A single Bookbag agent can handle the entire routine support load for an ecommerce store:

- **Resolve tickets automatically** across a website widget, email, WhatsApp, SMS, Instagram, and Facebook Messenger.
- **Take real actions** — collect leads, escalate to a human, look up an order, or call your own API with custom actions.
- **Run a full help desk** — tickets, assignment, saved views, takeover, scheduling, and AI-assisted replies for the cases that need a person.
- **Reach out proactively** with outbound campaigns over WhatsApp.
- **Render interactive widgets** — order-status cards, product carousels, forms — right inside the chat.

## How it works

1. **Connect your knowledge** — Add data sources — crawl your website, upload files, paste text, or write Q&A pairs. Bookbag chunks and embeds them into a vector index.
2. **Bookbag builds an agent** — Describe what you want in plain language and the agent builder drafts a system prompt, picks sensible defaults, and grounds the agent in your sources.
3. **Test in the playground** — Chat with your agent, inspect the exact prompt and the sources behind each answer, and refine.
4. **Deploy everywhere** — Embed the widget on your storefront and connect email and messaging channels. The same agent answers consistently across all of them.

## Core concepts

| Concept | What it is |
| --- | --- |
| Agent | The AI assistant you build and deploy. Has its own knowledge, model, prompt, channels, and settings. |
| Data source | Something the agent learns from — a website crawl, file, text snippet, or Q&A pair. |
| Playground | Where you chat with and tune an agent before (and after) deploying it. |
| Channel | A surface the agent answers on — website widget, email, WhatsApp, Slack, Instagram, Messenger. |
| Action | A capability beyond answering — collect a lead, escalate, or call an API. |
| Help desk | The ticket inbox for conversations that need a human, with assignment and takeover. |
| Credit | The unit of usage. Each AI reply costs credits based on the model; your plan includes a monthly allotment. |

## What's next

- [Build your first agent](/docs/getting-started/your-first-agent) — A step-by-step walkthrough from zero to a deployed agent.
- [Data sources](/docs/agents/data-sources) — Every way to give your agent knowledge, and how to keep it fresh.
- [Best practices](/docs/getting-started/best-practices) — How to get accurate, on-brand answers from day one.
- [Deploy & channels](/docs/deploy/overview) — Put your agent on your storefront and connect messaging channels.