COMPLETE GUIDE: CONFIGURATION AND MANAGEMENT OF THE POLI BOT

This document centralizes the guidelines for creating, configuring, and managing automation and routing flows in the Poli Bot. The objective is to guide the structuring of self-service, chat distribution, and post-service message sequences.

WHAT IS THE POLI BOT AND HOW DOES IT WORK?

The Poli Bot is the platform’s native virtual assistant, designed to automate customer reception, perform service triage, and execute message sequences without the need for human intervention.

It is built through a visual flow system (decision trees), which ensures full control over what will be communicated to the customer.

HOW DOES IT WORK IN PRACTICE?

Unlike free Artificial Intelligence bots (which generate unpredictable responses), the Poli Bot is based on strict and predictable rules. It operates following a sequential logic of “Triggers, Validations, and Actions.”

1) THE ENTRY TRIGGER (THE BEGINNING)

The bot always “wakes up” based on a specific event. This can be when the customer sends the first message on WhatsApp (Inbound), when the platform sends a campaign (Outbound), or even when an agent closes a chat (to start a post-sales flow).

2) VALIDATIONS (BUSINESS RULES)

Before speaking to the customer, the bot can check the current scenario. The main validation is “Business Hours.” The bot checks whether your company is open. If it is closed, the customer is redirected to an absence message.

3) INTERACTION (MENUS AND MESSAGES)

During the flow, the bot guides the customer by offering clear options. It uses text messages, numeric lists, interactive lists, or clickable buttons to ask the customer what they want (e.g., “1 for Sales, 2 for Support”).

4) THE FINAL OBJECTIVE (ROUTING OR CLOSING)

Every Poli Bot flow has a destination. After the customer selects options in the menu, the bot performs its main function: organizing the queue. It automatically directs the customer to the correct department and hands the conversation over to an available human operator. If the customer’s issue has already been resolved through the menu, the bot may simply close the conversation.

MAIN OBJECTIVES OF THE BOT IN POLI

• Smart Triage: Works as a virtual receptionist, identifying the customer’s issue and directing them to the correct department.
• After-hours Support: Prevents customers from being left without a response overnight or during weekends.
• Follow-up Automation: Keeps the relationship active by sending automatic reminders days after a quote or negotiation.

1. BOT OVERVIEW AND CREATION

To start any automation on the platform, you must create a Bot.

• Access: Go to the Automations (Chatbot) menu.
• Creation: Click “Create bot” (or “New” in Self-service).
• Initial Setup: Define a title for the bot and select the communication channel that will be used.

2. ENTRIES (TRIGGER EVENTS)

Entries define the event that starts the bot flow. The Poli Bot has four main triggers:

1) Customer starts a chat

Triggered as soon as the customer sends the first message (Inbound).

Attention: This event requires the previous chat to be completely closed in the system. If the customer sends a message while the chat is still open, or seconds after closing (before the status update), the flow will not be triggered.

2) Message broadcast performed

Triggered automatically after the platform sends an active message to the contact (Outbound/Campaigns).

3) Action in the chat

Triggered by a classification or manual closing performed by the agent (e.g., Sale Completed, In Negotiation, Lost Lead). Used for post-service sequences.

4) Close a chat

Triggered at the moment of general chat closure.

Note: This trigger should not be used for routing/forwarding flows.

3. FLOW ACTIONS

After defining the Entry, the flow must be built using Actions. Access to the Actions tab is only enabled after configuring the Entry.

Actions are divided into three categories:

3.1 MESSAGE ACTIONS

• Send text message: Sends a standard message.
• Send WhatsApp template: Sends pre-approved templates (required for contacts inactive for more than 24 hours).
• Send options menu:
• • Numeric List: The customer responds by typing the number corresponding to the option (e.g., 1 – Support, 2 – Finance).
  • Interactive Buttons: Clickable buttons on the screen (created via Poli’s template builder).
  • Interactive List (WhatsApp): Visual menu in drawer/list format (created via template builder).

Available Variables (Only for Message Actions)

• {contact_name}: Name of the contact saved on the platform.
• {day_greeting}: Returns Good morning, Good afternoon, or Good evening.
• {contact_department}: Name of the contact’s current department.
• {department_online_users}: Number of users online in the selected department.
• {protocol}: Conversation protocol number.
• {rating}: Link to the satisfaction survey configured in the platform.

3.2 VALIDATION ACTIONS

A) CHECK BUSINESS HOURS

This is one of the most important validation actions in the Poli Bot and is usually the first recommended step in any self-service flow.

It acts as an initial filter, evaluating the exact moment the customer sent the message.

How it works in practice:
When adding this action, the bot flow will necessarily split into two paths:

• PATH 1: WITHIN BUSINESS HOURS (YES)
  The customer journey continues normally to the next steps.
• PATH 2: OUTSIDE BUSINESS HOURS (NO)
  The flow is redirected. The ideal configuration is to add a text message informing about the absence and then the “Close Conversation” action.

The platform allows different hours for different days of the week. This action prevents contacts from being routed to departments with no available agents.

B) CHECK IF THERE WAS A RESPONSE (WAIT TIME / TIMEOUT)

This action works as a timer waiting for customer interaction after the bot sends an options menu.

How it works in practice:
You define a maximum waiting time in seconds. The flow then splits into two:

• PATH 1: WITH RESPONSE (YES)
  If the customer interacts within the time limit, the flow follows the route corresponding to the selected option.
• PATH 2: WITHOUT RESPONSE (NO)
  If time expires, the flow is redirected. Best practice is to send a message informing about inactivity and then either “Close Conversation” or “Send to Available Operator.”

This ensures the bot journey has a clear ending and prevents customers from being stuck in the chat indefinitely.

3.3 MANAGEMENT ACTIONS

• Route to Department: Sends the customer to a specific queue respecting distribution rules.
• Send to Available Operator: Sends the conversation to any available operator.
• Send Satisfaction Survey: Triggers the survey module.
• Close Conversation: Ends the interaction and removes the customer from the bot flow.
• Attach Tag: Automatically applies a tag to the contact for organization, segmentation, and reporting.

4. SERVICE ROUTING AND DISTRIBUTION

For the “Route to Department” action to work properly, strict requirements must be met.

4.1 REQUIREMENTS TO RECEIVE CHATS

For a user (agent) to receive chats distributed by the bot, they must:

1. Be assigned to the selected department in the flow.
2. Be Online.
3. Be Available.
4. Have the Chat Operator profile active (even if they are Manager or Supervisor).

Queue behavior:
If there are no available agents in the department at the time of routing, the contact will enter the department queue. Distribution will occur automatically as soon as the first agent comes online.

4.2 DISTRIBUTION RULE CONFIGURATION

These rules define how the system assigns chats to agents.

• Rule Level: Can be configured at Company, Department, or Channel level.
• Lock to Department: Recommended as Yes for bot flows to ensure the contact stays restricted to that department.

Prioritization

• Prioritize account owner: Attempts to route to the customer’s account owner.
• Lock to account owner: Restricts the service exclusively to the account owner.
• Prioritize last agent: Attempts to route to the agent who handled the previous conversation.

Offline Behavior

Option to keep the contact assigned to the agent even if they go offline (prevents redistribution).

Distribution Type

• Balanced: Fair distribution respecting the maximum number of simultaneous chats per agent.
• Sequential: Follows a fixed queue order among agents.