DOCSEARCHbyAlgolia
Submit Docs

Quickchat AI Documentation

website·Indexed 35 days ago

https://docs.quickchat.ai

Build, deploy, and manage no-code AI Agents for customer support, sales, and internal operations. Features include a Knowledge Base for AI expertise, Custom and MCP Actions for external tool integration, human handoff, and smart data gathering. Includes a REST API for managing chatbot settings, conversations, and widget configurations, as well as integrations with Discord, HubSpot, Drupal, Email, Facebook Messenger, and Instagram.

Tokens
51.1K
Snippets
110
Records
372
Agent score
61%

What's inside Quickchat AI

  1. Overview of AI Insights aggregated analysis

    AI Insights provides a high-level aggregated view of conversation data, focusing on the latest aggregate state and per-conversation drill-downs. It integrates several key analysis metrics: Sentiment, Conversation Outcome, Topics, and Link Tracking. While the 'Advanced' tab shows distribution over time, AI Insights focuses on current aggregates.

  2. Overview of Quickchat AI Platform

    Quickchat AI is an enterprise-grade Conversational AI platform used to build, deploy, and manage no-code AI Agents for customer support, sales, and internal operations. The platform is built on a Four-Module System: AI Identity/Personality, AI Knowledge Base, AI Actions, and Conversation Insights.

  3. Overview of Quickchat AI Conversations API

    Quickchat AI captures and analyzes conversations across all channels (widget, Slack, WhatsApp, Telegram, etc.). The Conversations API allows programmatic access to this data for dashboards, analytics, CRM workflows, and BI tools. Every conversation includes AI-powered analysis such as sentiment detection, topic classification, resolution tracking, and automatic flagging. Data is distributed across four primary areas: listing conversations, retrieving detailed conversation info (including AI insights), accessing full message history/events, and managing custom conversation metadata.

  4. Overview of Human Handoff functionality

    Human Handoff allows an AI Agent to automatically transfer conversations to human operators. The process is AI-driven: the AI evaluates messages against generic and custom guidelines to trigger a handoff. When triggered, the visitor is asked for confirmation; if they agree, the conversation is transferred (optionally with an AI-generated summary). If the request occurs outside configured working hours, a custom message is shown instead.

    Additional handoff methods include:

    • Keyword triggers: Specific phrases (e.g., "speak to human") that prompt handoff.
    • Manual handoff: Operators can take over conversations via the Dashboard Inbox by changing the assignee.
    • Topic-based Routing: An advanced feature that routes handoffs to specific operators based on the conversation topic using routing profiles.

  5. Overview of Quickchat AI Agent Actions

    Quickchat AI Agents use Actions to perform tasks during conversations. There are two primary types:

    1. Quickchat AI Actions (Built-in): Pre-configured features that only need to be enabled, such as Human Handoff (transferring to a person) and Smart Data Gathering (collecting leads).
    2. Custom Actions: Tools that extend the AI's capabilities to external systems. These include:
      • API Actions: Call any REST endpoint via HTTP.
      • MCP Actions: Connect to Model Context Protocol (MCP) servers.
      • Google Sheets: A shortcut for logging data to spreadsheets.
      • HubSpot Action: Pre-built CRM integration for contacts, deals, and tickets.
      • Discord Action: Templates for support and moderation via the Discord API.
      • Shopify MCP: Allows the AI to search catalogs, store policies, and build carts. (Note: Order lookup requires a separate 'Check order status' API Action installed via the Shopify App Store).

  6. Overview of Knowledge Base Articles

    Articles are the fundamental building blocks of a Quickchat AI Agent's Knowledge Base. They represent pieces of knowledge such as FAQ entries, help pages, or product descriptions. Content is processed automatically in the background without requiring a manual retraining step, typically taking effect within seconds.

    There are three types of articles:

    • Article: Full documents with a title and body.
    • Paragraph: Short, standalone text entries without titles, used for brief facts or one-liners.
    • URL: Created automatically when importing web content; these contain scraped page content and the source URL.

  7. Overview of Quickchat Knowledge Base data sources

    The Knowledge Base is the foundation for a Quickchat AI Agent's understanding. Content can be added via several sources: manually written articles, atomic paragraphs, website imports, file uploads, Shopify integration, Intercom imports, and YouTube transcripts.

  8. Overview of Conversation Rating feedback

    Conversation Rating allows you to collect visitor feedback on their AI Agent experience by prompting them to rate the conversation at the end. Ratings are visible in the Dashboard under Conversations > Inbox and are included in conversation exports, helping to identify knowledge gaps and track satisfaction trends. This corresponds to the 'Feedback settings' in the Dashboard.

  9. Overview of AI Agent simulation testing

    The Testing feature allows you to run simulations against your AI Agent using curated datasets of user messages. This process helps catch regressions after editing the Knowledge Base, Identity, or Actions, stress-test edge cases (such as prompt-injection or out-of-scope requests), and compare different prompt versions. You can find this feature under Conversations → Testing in the left sidebar.

  10. Overview of AI Actions

    AI Actions extend an AI Agent's capabilities beyond the Knowledge Base by defining triggered behaviors. There are two primary types:

    • HTTP Request Actions: Call external APIs to perform tasks like checking order status, creating support tickets, or looking up account information.
    • MCP Actions: Connect to Model Context Protocol servers for advanced tool integrations (e.g., Shopify product lookups, database queries).

    Actions are primarily created and configured in the Quickchat Dashboard, while the API is used for listing, updating (enabling/disabling), and deleting actions, as well as specific MCP configuration.

  11. Overview of Conversation Metadata in Quickchat

    Conversation Metadata allows you to attach custom key-value data to conversations to bridge external systems (like CRMs or helpdesks) with your AI Agent. Metadata is used for:

    • Personalization: The AI Agent can reference metadata values (e.g., customer name or plan tier) to personalize responses.
    • AI Actions: Metadata values can be injected into AI Actions using the {{metadata_<key>}} placeholder syntax.
    • Visibility: Metadata is visible in the Inbox conversation detail view and included in conversation exports.
    • Validation: If a JSON schema is configured in the Dashboard, metadata is validated against it.

    Metadata can be provided via API, as initialization parameters in the widget/embed, or dynamically via the JavaScript SDK.

  12. Track AI-Assisted Revenue for Shopify stores

    AI-Assisted Revenue attributes Shopify orders to specific conversations where the AI Agent helped build the shopper's cart. This feature is only available when the AI Agent is connected to Shopify. If the store is later disconnected, historical attributed revenue data is preserved.

    An order is counted as 'assisted' only if the AI Agent added items to the cart during the chat and that specific cart was later checked out. If a shopper adds items manually after a conversation, the order is not attributed. Revenue figures are based on Shopify's order totals in the store's presentment currency, including tax and shipping as configured in Shopify settings.