Chatbot Flows

Chatbot Flows are used to ask the customer questions in order, collect answers and advance the conversation through defined steps. In the visual editor you can add steps and change their order by dragging.

The Difference Between a Flow and Automation

Automation rules start one or more actions when an event occurs. Chatbot flows, on the other hand, organize conversation scenarios that wait for the customer's answer and use that information in later steps.

For example, "add a tag to a conversation that mentions the word return" can be an automation. "Ask the customer which product they need support for, get the answer and move to the relevant support step" is a conversation flow.

Creating a Flow

Open the flows area

Open the flows area in the Automation section of the panel.

Pick an example or create new

Choose a ready-made example or create a new flow.

Give it a name

Give the flow a descriptive name.

Start settings

Set the start settings and the trigger options offered.

Add steps

Add message, question, condition and action steps; arrange their order.

Set the variables

Specify which variables the information from questions will be written to.

Save and test

Save the flow. Activate it when you want to make it available and check the result on the relevant channel.

You can later edit, duplicate or deactivate flows.

Available Steps

Message: Shows an explanation or direction to the customer.
Question: Gets an answer from the customer. Quick reply options can be offered in suitable steps.
Condition: Determines how the next action proceeds based on the information received.
Action: Used for tasks such as tagging, handoff to a human agent, webhook, email or updating a variable.
AI reply: Gets an answer from the AI with the necessary context.
Wait: Adds a delay between steps.

Configure the steps based on the options offered on screen. Adding options to a question and setting the conditions of the next steps are different tasks; check all answer paths.

Variables and Conditions

You can hold a question's answer in a descriptive variable such as product_type or support_topic. Make sure the next condition checks the correct variable and compares it with the expected value.

Also consider what happens when the customer gives an unexpected answer or does not share the required information. Checking that the flow works only with ideal answers is not enough.

Using It Together With AI

Flows provide structured conversation steps; AI produces natural-language answers where appropriate. If you will use an AI reply in the flow, check that the required business information exists in the training records.

Try when the flow will start, how it proceeds while waiting for the customer's answer, and how the conversation continues when it completes. In steps that include a handoff, also check agent availability and handoff settings.

Using the Ready-Made Examples

Ready-made flows are a starting point. Change the product names, working hours, options and other example information to match your business. Complete blank fields such as the webhook address or email.

A result message does not guarantee the real action

A flow asking for a date and time does not by itself create a real appointment record. Use result messages such as "Your appointment has been created" or "Your order has been received" only when the relevant record or integration action is actually completed.

Check Before Activating

Try the start condition, all options, missing answers, actions and the closing message. Check that no wrong information is given and that the required actions take place. Keep in mind that activated flows can run in real customer conversations.