Skip to main content

Pizza Ordering Bot Tutorial

This tutorial demonstrates how to build a simple conversational flow using a pizza ordering bot. This example shows how to handle straightforward ordering processes with basic branching and order validation.

Flow Overview​

Here's the conversation flow we'll build:

What You'll Learn​

This tutorial covers:

  • Building simple conditional flows
  • Handling user corrections and order modifications
  • Using entities for structured data collection
  • Creating confirmation and validation steps
  • Managing conversation state across multiple turns

Building the Pizza Bot​

Step 1: Set Up Basic Structure​

  1. Create a new bot called "PizzaBot"
  2. Set up the welcome intent with a greeting:

    Welcome to Tony's Pizza! I can help you order delicious pizzas. What would you like to do today?

Step 2: Create Pizza Ordering Entities​

Create these custom entities:

Entity: pizza_size

  • Values: small, medium, large, extra-large
  • Synonyms:
    • small: personal, individual, 10-inch
    • medium: regular, 12-inch
    • large: family, 14-inch
    • extra-large: xl, jumbo, 16-inch

Entity: pizza_toppings

  • Values: pepperoni, mushrooms, sausage, peppers, onions, olives, cheese
  • Synonyms:
    • pepperoni: pepperoni, salami
    • mushrooms: mushroom, fungi
    • sausage: italian sausage, meat
    • peppers: bell peppers, green peppers

Step 3: Create the Order Intent​

Intent: order_pizza

  • Example Phrases:
    • "I want to order a pizza"
    • "Can I get a large pepperoni pizza?"
    • "Order pizza for delivery"
    • "I'd like to place an order"

Step 4: Build the Ordering Flow​

Story Node: Pizza Order Start

  • Entry response: "Great! Let's get your pizza order started. What size would you like?"
  • Form parameters:
    • size (type: @pizza_size) → prompt: "What size pizza would you like? We have small, medium, large, and extra-large."

Story Node: Toppings Selection

  • Entry response: "Perfect! Now, what toppings would you like on your $session.params.size pizza?"
  • Form parameters:
    • toppings (type: @pizza_toppings, allow multiple) → prompt: "What toppings would you like? You can choose multiple toppings."

Story Node: Order Confirmation

  • Entry response: "Let me confirm your order: One $session.params.size pizza with $session.params.toppings. Is this correct?"

Step 5: Add Confirmation Logic​

In the Order Confirmation Story Node, add these routes:

Route 1: Confirmation

  • Intent: confirm (create with phrases like "yes", "correct", "that's right")
  • Transition: Checkout

Route 2: Modification

  • Intent: modify_order (create with phrases like "no", "change that", "modify")
  • Agent Says: "No problem! Let's start over with your order."
  • Transition: Pizza Order Start

Step 6: Complete the Checkout​

Story Node: Checkout

  • Entry response: "Excellent! Your $session.params.size pizza with $session.params.toppings will be ready in 20-25 minutes. Your total is $15.99. Thank you for choosing Tony's Pizza!"
  • Add a route with condition true that transitions to End Session

Advanced Features to Add​

Once you have the basic flow working, try adding these enhancements:

Price Calculation​

  • Add price entities for different sizes
  • Calculate total cost based on size and number of toppings
  • Display the price before confirmation

Delivery Options​

  • Add entities for delivery vs pickup
  • Collect address information for delivery orders
  • Estimate delivery times

Special Offers​

  • Create intents for checking deals
  • Add conditional logic for applying discounts
  • Handle promotional codes

Order Tracking​

  • Create intents for checking order status
  • Store order information for later retrieval
  • Provide estimated completion times

Testing Your Pizza Bot​

Test these conversation paths:

  1. Happy Path: "I want a large pepperoni pizza" → confirm → complete order
  2. Modification Path: Order → change mind → reorder → confirm
  3. Unclear Input: Test with ambiguous requests to ensure fallbacks work
  4. Multiple Toppings: "Large pizza with pepperoni, mushrooms, and sausage"

Best Practices Demonstrated​

  • Clear confirmation steps: Always let users review their orders
  • Flexible modification: Allow users to change their minds
  • Comprehensive testing: Test all possible conversation paths
  • Graceful error handling: Provide helpful responses when things go wrong

This pizza ordering bot showcases how Botnex can handle real-world ordering scenarios with multiple decision points and user interactions.