> For the complete documentation index, see [llms.txt](https://docs.saleschat.pro/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.saleschat.pro/getting-started/readme.md).

# Introduction

What Saleschat is, how AI agents and AI workflows handle your WhatsApp conversations, and which pages to read first for your task.

Saleschat runs your business's WhatsApp conversations with leads, the people you sell to or support, from your own WhatsApp Business Account (WABA). An AI agent answers their questions, follows up, and qualifies them against criteria you set, then hands each lead to a human agent on your team when it has qualified them or can't help further. Campaigns, automations, and inbound rules that you configure decide when a conversation starts and whether the AI agent or your team takes it.

## Sales and support

In sales, the AI agent follows up on enquiries from your CRM, learns what each lead is looking for, and passes qualified leads to your sales team with that context. Enterprises often hold several enquiries for the same person, for example from different campaigns or products, and Saleschat keeps each one as a separate lead with its own history and data.

WhatsApp allows one conversation between a business number and a phone number, so one of those leads is active in the conversation at a time. Lifecycle rules, which you set per org, decide which lead is active and when the next one takes over. [Lead Lifecycle](https://docs.saleschat.pro/principles/lead-lifecycle.md) explains how to set these rules for several enquiries.

In support, the AI agent answers questions from your knowledge base and hands the conversation to your support team when it can't resolve the issue. Support teams usually identify people by phone number, so each number has a single lead and the default setup works without lifecycle rules.

## Key concepts

* **Org**: your business's Saleschat organisation. It holds your users (the people who sign in to Saleschat), teams, AI agents, WABAs, phone numbers, and API keys. Most businesses need only one org. An enterprise needs more when processes that use Saleschat run on separate CRMs, as described in [Planning Orgs, WABAs & Phone Numbers](https://docs.saleschat.pro/principles/planning-orgs-wabas-and-phone-numbers.md).
* **Lead**: the person you're talking to, together with Saleschat's record of them. The record is linked to a phone number, is usually created by your CRM, and holds the lead's conversation, attributes, and events.
* **Inbound message** and **outbound message**: an inbound message is one the lead sends to you, and an outbound message is one you send to the lead.
* **AI agent**: the Saleschat AI that talks to your leads. You configure its goals and qualification criteria, and give it tools such as a knowledge base or an MCP server.
* **AI workflow**: a single run of an AI agent on one lead, from the moment a journey starts it until it finishes or hands the lead over.
* **Human agent**: a user who handles leads, either by taking over from the AI agent or by messaging leads directly.
* **Journey**: anything that acts on leads without a human agent. A campaign sends outbound messages to a list of leads, an automation reacts to events from your CRM, an inbound rule decides who handles new inbound messages, and an App send delivers an outbound message from a connected platform such as your marketing tool. Campaigns, automations, and inbound rules can also start an AI workflow.

## How an AI workflow runs

An AI workflow starts when a journey triggers it, for example an inbound rule that hands new inbound messages to the AI agent. From then on the AI workflow holds the lead's conversation. Only one AI workflow can hold a conversation at a time, and while it does, campaigns and automations skip the lead and human agents can't send outbound messages.

The AI agent then talks to the lead and calls its tools when it needs information or has to take an action. The AI workflow ends when the AI agent meets its goals or hands the lead to your team, or when the lead stops replying. A user with the Admin role or above can also stop it at any time.

When the AI workflow ends, Saleschat records the outcome: a verdict on what happened, which goals were met, whether the lead qualified, and what the conversation was about. You can review each outcome in [AI Analytics](https://docs.saleschat.pro/application/ai-analytics.md).

## Where to start

Admins usually begin by setting up the org, developers by connecting Saleschat to their systems, and sales and support heads by measuring results.

### Set up your org

The Saleschat team creates your org and its first user with the Owner role. After that:

1. Generate an API key and connect your WABA through Meta's embedded signup, which Saleschat requires. During signup you can use a virtual number provided by Meta or bring your own phone number. [Onboarding](https://docs.saleschat.pro/getting-started/onboarding.md) covers both steps.
2. Submit your message templates for approval. See [WhatsApp (WABA) templates](https://docs.saleschat.pro/getting-started/whatsapp-waba-templates.md).
3. Group your human agents into teams, so that leads the AI agent hands over reach the right people. See [Teams](https://docs.saleschat.pro/application/teams.md).

### Connect your systems

The lead data your CRM sends to Saleschat is the context the AI agent works from. You can run Saleschat from the dashboard alone, but with a CRM integration your CRM decides which human agent each lead is assigned to. A CRM integration also protects phone numbers: Saleschat identifies the leads your CRM creates by their CRM ID, so human agents never see a lead's phone number. You can turn on masking as well, for phone numbers that leads type into their messages.

1. Allow the Saleschat domains on your networks, listed in [Whitelist URLs](https://docs.saleschat.pro/getting-started/whitelist-urls.md). If your firewall restricts which servers can call your APIs, add the addresses in [Outbound IP addresses](https://docs.saleschat.pro/getting-started/outbound-ip-addresses.md).
2. Create leads and send Saleschat their attributes and events. The [Basic CRM integration tutorial](https://docs.saleschat.pro/getting-started/basic-crm-integration-tutorial.md) walks through the Lead APIs, and [Events & Attributes](https://docs.saleschat.pro/principles/events-and-attributes.md) explains the data model.
3. Receive lead and conversation updates in your systems through [Webhooks](https://docs.saleschat.pro/getting-started/webhooks.md).
4. Send outbound messages from a marketing or notification platform as App sends. See [Apps on Saleschat](https://docs.saleschat.pro/principles/apps-on-saleschat.md).

For every endpoint and its parameters, see the [Leads API](https://docs.saleschat.pro/api-documentation/leads.md) and the rest of the API documentation.

### Measure results

* [AI Analytics](https://docs.saleschat.pro/application/ai-analytics.md) shows what your AI agent did in each conversation, whether it met its goals, and which leads qualified.
* [Analytics](https://docs.saleschat.pro/application/analytics.md) lets you build your own charts from leads, messages, journeys, and AI workflows.
* [Team Performance](https://docs.saleschat.pro/application/team-performance.md) and [Teams analytics](https://docs.saleschat.pro/application/teams/analytics.md) show response times and workload for each human agent and team.
