How to Trigger a Bot Flow From a ZazzyAgent AI Agent
AI is excellent at understanding natural-language conversations.
Flow Builder is excellent when you need a structured process.
You don't have to choose between them.
An AI Agent in ZazzyAgent can understand a customer's request and then trigger an existing bot flow when the customer reaches a process that is better handled step by step.
Example
Customer:
I want to book a demo.
AI Agent understands:
Demo booking.
Instead of asking every question conversationally, it can trigger your existing booking flow.
The experience becomes:
AI understands → Flow handles the structured process
The action
Use:
##trigger_bot_flow## : 15
The number represents the selected bot flow.
Use the ## autocomplete to select the flow rather than manually entering an ID.
Example: Trigger an appointment flow
When the customer clearly confirms that they want to book an appointment,
trigger the Appointment Booking flow.
##trigger_bot_flow## : 15
Why combine AI with Flow Builder?
Some processes have predictable steps.
For example:
Appointment
Choose service
Choose date
Choose time
Submit request
AI doesn't need to reinvent that process every time.
It can recognize the customer's intention and hand the conversation to the structured flow.
Example: Lead qualification
Customer:
I'm interested in your enterprise plan.
AI recognizes that the customer needs qualification.
It can trigger:
Enterprise Lead Qualification
That flow can collect:
Name
Company
Requirement
Budget
Timeline
Example: Support
Customer:
I need to return my order.
AI recognizes:
Return request.
It can trigger a structured return flow that asks for the information your business needs.
Don't trigger flows too early
The condition matters.
Weak:
If the customer talks about appointments, trigger the appointment flow.
This may trigger when the customer is simply asking:
What are your appointment hours?
Better:
Trigger the Appointment Booking flow only when the customer clearly indicates they want to book an appointment.
Ask for missing information first
Suppose your booking flow requires a service.
Customer:
I want to book.
Your agent could ask:
Sure. Which service would you like to book?
Once the required information is available, the agent can trigger the flow if the flow itself doesn't collect it.
Trigger a flow after AI qualification
One powerful architecture is:
Customer
↓
AI Agent
↓
Understand request
↓
Ask necessary questions
↓
Trigger structured Flow
↓
Structured process
This creates a much smoother experience than forcing customers to navigate menus from the beginning.
Existing flows don't need to be rebuilt
If you already have a working Flow Builder process for something, you can keep it.
The AI Agent becomes the natural-language entry point.
For example:
You already have:
Appointment Flow
Instead of requiring the customer to type:
appointment
the AI can recognize:
I'd like to book something for next week.
and move them into the appropriate process.
Example System Prompt
You are the Appointment Agent.
Help customers understand our appointment services.
When a customer clearly wants to book an appointment,
trigger the Appointment Booking flow.
##trigger_bot_flow## : 15
If the customer is only asking general questions
about appointments, answer using the connected knowledge.
Do not trigger the booking flow until the customer
clearly wants to make a booking.
Triggering flows from AI vs keyword triggers
A keyword trigger requires the customer to use a predictable word.
AI can understand:
Can I schedule something for Friday?
and recognize that the customer wants an appointment.
The AI then becomes the natural-language router.
Common problems
Flow isn't triggered
Check:
Correct flow selected.
Flow is active/published.
Agent actually reached the condition.
Action is present in the System Prompt.
Correct account/channel is being used.
Wrong flow is triggered
Review the instruction that determines when the action should run.
Make the condition more specific.
Flow interrupts a normal conversation
The condition is probably too broad.
Tell the AI exactly when the flow should and should not be used.
A good architecture
Use:
AI for understanding
Flow Builder for structure
Actions for execution
This combination can create customer experiences that feel conversational while still following controlled business processes.