English
EnglishEspañolPortuguês
Implementation guide

Build an AI Support Operation

Build an AI support operation: tickets, WhatsApp routing, escalations, and an operational queue.

~40 min6 steps

What you will build

By the end of this guide, you will have a fully operational AI support system: it receives customer messages, classifies and creates tickets, routes to the right team, escalates when needed, and logs all activity.

Agent BuilderWorkflowsChannelsCRM TicketsMax Activity
1

Create the Support Agent

Build the AI agent that handles customer requests, FAQs, and ticket creation.

  1. Create the support agentOpen guide

    Open Agent Builder and create a new support agent with classification instructions.

  2. Write the support playbookOpen guide

    Define how the agent handles FAQs, collects ticket info, and identifies escalation triggers.

  3. Configure knowledge resources

    Attach your support documentation, FAQ files, and policy resources to the agent.

  4. Set response style

    Configure tone, language, and response length for customer-facing support.

2

Set Up the Ticket System

Create the CRM Ticket object with fields, categories, and priorities.

  1. Create the Ticket objectOpen guide

    Add category, priority, status, and resolution fields to the Ticket object.

  2. Define ticket categories

    Set up categories: billing, technical, account, delivery, escalation.

  3. Configure assignment rules

    Define which agent types handle which ticket categories automatically.

3

Build the Routing Workflow

Create the workflow that classifies incoming messages and routes to the right handler.

  1. Create the intake workflowOpen guide

    Trigger on incoming WhatsApp or channel message and classify intent.

  2. Add the AI classification node

    Use the support agent to classify: FAQ, billing, technical, or human needed.

  3. Create the ticket automatically

    On classification: create Ticket record with category, priority, and customer context.

  4. Route to the correct queue

    Branch by category: FAQ → auto-reply; billing → billing team; technical → tier-2.

4

Configure Escalation Logic

Define when and how the AI hands off to a human agent.

  1. Define escalation triggersOpen guide

    Detect: angry tone, repeated contact, unresolved after N messages, or explicit request.

  2. Build the escalation workflow

    On trigger: update ticket priority, notify the assigned human, transfer conversation.

  3. Configure SLA rules

    Set response time targets and auto-escalation if SLA is breached.

5

Connect Support Channels

Configure WhatsApp, Instagram, and Messenger as support entry points.

  1. Connect WhatsApp Business

    Link the WhatsApp number to the support agent and the routing workflow.

  2. Set up coexistence rules

    Define when AI handles vs. when a human agent joins the conversation.

  3. Test a full support interaction

    Send a test message, verify ticket creation, routing, and escalation path.

6

Final Operational State

Verify the complete support system end-to-end.

  1. Customer message arrives

    Incoming message triggers the classification workflow immediately.

  2. AI classifies and responds

    The support agent handles FAQs instantly, creates a ticket for complex cases.

  3. Ticket created and routed

    Ticket record created with context, priority, and assigned to the right team.

  4. Escalation fires when needed

    SLA breach or angry tone triggers escalation to human agent with full context.

  5. All activity in Max

    Every AI action, message, and status change visible in the Max activity timeline.