	In Browser
	StumbleUpon
	del.icio.us
	Google
	Google Buzz
	reddit
	LinkedIn

	Facebook
	Twitter
	Linkedin
	E-Mail

Generative AI > Google Gemini API > E-commerce Checkout Agent

E-commerce Checkout Agent

Author: Venkata Sudhakar

A checkout agent guides a customer through the complete purchase flow in one conversation - validating their cart, checking stock, applying discount coupons, confirming payment eligibility, and generating an order. For ShopMax India, this replaces a multi-page checkout form with a single agent that handles the entire flow and recovers gracefully from issues like out-of-stock items or invalid coupons.

The agent below uses five tools covering each stage of the checkout process. The model decides the correct sequence based on the customer query and tool results, handling both the happy path and edge cases like expired coupons or insufficient stock.

import uuid
from google.adk.agents import LlmAgent
from google.adk.tools import FunctionTool
from google.adk.sessions import InMemorySessionService
from google.adk.runners import Runner
from google.genai.types import Content, Part

def validate_cart(product_id: str, quantity: int) -> dict:
    """Validate a ShopMax India cart item - check product exists and quantity is valid."""
    catalogue = {
        "SKU-7821": {"name": "Samsung Galaxy S24", "price_rs": 74999, "max_qty": 2},
        "SKU-1001": {"name": "OnePlus 12",         "price_rs": 64999, "max_qty": 2},
    }
    item = catalogue.get(product_id)
    if not item:
        return {"valid": False, "reason": "Product not found"}
    if quantity < 1 or quantity > item["max_qty"]:
        return {"valid": False, "reason": f"Quantity must be 1 to {item['max_qty']}"}
    return {"valid": True, "product": item["name"], "unit_price_rs": item["price_rs"],
            "total_rs": item["price_rs"] * quantity}

def check_stock(product_id: str, quantity: int) -> dict:
    """Check if sufficient stock is available at ShopMax India warehouses."""
    stock = {"SKU-7821": 45, "SKU-1001": 12}
    available = stock.get(product_id, 0)
    return {"product_id": product_id, "requested": quantity,
            "available": available, "sufficient": available >= quantity}

def apply_coupon(coupon_code: str, order_total_rs: int) -> dict:
    """Apply a ShopMax India discount coupon to an order total."""
    coupons = {
        "SAVE10":   {"type": "percent", "value": 10, "min_order_rs": 20000},
        "FLAT2000": {"type": "flat",    "value": 2000, "min_order_rs": 50000},
    }
    c = coupons.get(coupon_code.upper())
    if not c:
        return {"applied": False, "reason": "Invalid coupon code"}
    if order_total_rs < c["min_order_rs"]:
        return {"applied": False,
                "reason": f"Minimum order Rs {c['min_order_rs']} required for this coupon"}
    discount = int(order_total_rs * c["value"] / 100) if c["type"] == "percent" else c["value"]
    return {"applied": True, "coupon": coupon_code, "discount_rs": discount,
            "final_total_rs": order_total_rs - discount}

def check_payment_eligibility(user_id: str, amount_rs: int) -> dict:
    """Check if a ShopMax India customer is eligible for EMI or COD for this order."""
    return {
        "user_id": user_id,
        "cod_eligible": amount_rs <= 50000,
        "emi_eligible": amount_rs >= 10000,
        "emi_options": ["3 months", "6 months", "12 months"] if amount_rs >= 10000 else []
    }

def place_order(user_id: str, product_id: str, quantity: int, total_rs: int, payment_method: str) -> dict:
    """Place a confirmed ShopMax India order and return the order ID."""
    order_id = "ORD-" + uuid.uuid4().hex[:6].upper()
    return {
        "order_id": order_id,
        "user_id": user_id,
        "product_id": product_id,
        "quantity": quantity,
        "total_rs": total_rs,
        "payment_method": payment_method,
        "status": "confirmed",
        "eta_days": 3
    }

agent = LlmAgent(
    name="shopmax_checkout_agent",
    model="gemini-2.0-flash",
    tools=[
        FunctionTool(validate_cart),
        FunctionTool(check_stock),
        FunctionTool(apply_coupon),
        FunctionTool(check_payment_eligibility),
        FunctionTool(place_order),
    ],
    instruction=(
        "You are a ShopMax India checkout agent. "
        "Guide the customer through checkout: validate cart, check stock, apply any coupon, "
        "confirm payment options, and place the order. "
        "Always validate and check stock before applying coupons or placing the order."
    )
)

session_service = InMemorySessionService()
runner = Runner(agent=agent, session_service=session_service, app_name="shopmax")

session = session_service.create_session(app_name="shopmax", user_id="u_raj")
events = list(runner.run(
    user_id="u_raj",
    session_id=session.id,
    new_message=Content(parts=[Part(
        text="I want to buy 1 unit of SKU-7821. I have coupon FLAT2000. Place the order with EMI payment."
    )])
))
for event in events:
    if hasattr(event, "content") and event.content:
        for part in event.content.parts:
            if hasattr(part, "text") and part.text:
                print(part.text.strip())

It gives the following output,

Your ShopMax India checkout is complete! Here is a summary:

Product: Samsung Galaxy S24 (SKU-7821) x 1
Original price: Rs 74,999
Coupon FLAT2000 applied: -Rs 2,000
Final total: Rs 72,999
Payment: EMI (3, 6, or 12 months available on eligible cards)
Stock confirmed: 45 units available at warehouse

Order placed successfully!
Order ID: ORD-A3F7C2
Estimated delivery: 3 days

Thank you for shopping with ShopMax India!

The agent validated the cart, confirmed stock, applied the coupon (reducing the total by Rs 2,000), checked EMI eligibility, and placed the order in a single natural language interaction. The customer did not need to navigate separate pages - the agent handled the complete checkout sequence from one request.

For ShopMax India production, add a payment gateway tool that initiates the actual payment transaction (Razorpay or PayU) and waits for a webhook confirmation before calling place_order. Use Firestore to store the order state so the agent can resume if the session drops mid-checkout. Add an order confirmation tool that sends an SMS to the customer mobile number via the India-specific SMS API after successful order placement.

Send your comments, suggestions or queries regarding this site to [email protected].