> ## Documentation Index > Fetch the complete documentation index at: https://docs.lindy.ai/llms.txt > Use this file to discover all available pages before exploring further. # Smart Support Agent > Intelligent customer support automation with Lindy
## Quick Overview With the Smart Support flow, your agent will: * **Monitor** a support inbox (Gmail, Outlook—or swap in Help Desk APIs) * **Detect** whether the message needs a knowledge-base reply or human help * **Search** your docs, FAQs, and policies for accurate answers * **Respond** in your brand voice or draft a reply for approval * **Log** tickets and escalations to Slack, Zendesk, or HubSpot\ Start from the [email responder template](https://chat.lindy.ai/marketplace/?templateId=66bfc603ea73bec3fea906db) and tweak as below. ## Setting Up Your Smart Support Agent ### Step 1: Connect the Trigger to Support Inbox Gmail trigger Select `Gmail – Email Received` (could swap for other systems like `Zendesk – New Ticket`). ### Step 2: Determine If a Reply Is Needed Filter out non-support emails early to avoid responding to newsletters, confirmations, or internal messages. Requires reply condition `Condition` prompt: ``` Only continue if the sender asks a question about products, orders, or policies. ``` ### Step 3: Search the Knowledge Base Query your documentation to find relevant answers before crafting a response. Knowledge base search * Action: `Search Knowledge Base` * Attach FAQs, policies, refund rules, troubleshooting docs, and more ### Step 4: Verify an Answer Exists Check if we have sufficient information to provide a helpful response, escalating complex issues to humans. Found answer condition * If answer found → continue. * If not found → `Slack – Send Channel Message` the thread for human intervention * For `Send Channel Message` to work correctly, remember to add the Lindy app to your channel Found answer condition ### Step 5: Draft or Send the Reply Provide a helpful response using knowledge base information in your brand voice. Send reply Body prompt: ``` You are a professional customer support agent responding to a customer inquiry. Using the knowledge base information provided, craft a concise, helpful response that: - Directly answers their question using KB content - Maintains a friendly, professional tone - Provides clear next steps when applicable - Stays under 100 words - Includes relevant links or resources from the KB If multiple solutions exist, offer the most appropriate one based on their specific situation. Sign off professionally with your name/team. ``` ### Step 6: Log Support Interactions Track all support interactions in a Google Sheet for reporting and analysis, with optional integration to your existing ticketing system instead. Ticket update * Action: `Google Sheets – Append Row` to log ticket details, response time, and resolution * Update your support management system if applicable ## Advanced Customizations | Trick | What it does | | --------------------- | ------------------------------------------------------------------------- | | Multi-Channel Support | Swap Gmail for Zendesk, Intercom, or Slack slash-commands. | | Sentiment Escalation | Flag angry tone → route to human immediately. | | SLA Timer | `Wait` + Condition → alert if no answer within 2 h during business hours. | | Auto-Survey | Send CSAT form 24 h after ticket closure. | ## Best Practices Review policy and product docs weekly; outdated articles create wrong answers. Aim for one-screen replies; link out for deeper resources. Escalate order issues or refund disputes early to avoid churn. The same flow works for chat or SMS—just swap the trigger/action pair. ## Next Steps Handle support calls 24/7 Generate customer success reports Resolve questions in Slack channels Add oversight for complex support issues