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

	Facebook
	Twitter
	Linkedin
	E-Mail

Generative AI > Google Gemini API > Customer Support Escalation Agent

Customer Support Escalation Agent

Author: Venkata Sudhakar

A customer support escalation agent handles the majority of ShopMax India queries automatically while intelligently routing complex cases to the right human team. The agent classifies each query, attempts resolution using its tools, and escalates to Tier 2 support, the returns team, or the fraud team based on the outcome - ensuring customers are never stuck in an automated loop when they need a human.

The agent uses tools for the most common query types: order status, return eligibility, and refund status. When a query falls outside its resolution scope or the customer expresses frustration, it creates a support ticket and routes it to the appropriate team with full context pre-filled.

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 get_order_status(order_id: str) -> dict:
    """Get the current status of a ShopMax India order."""
    orders = {
        "ORD-441": {"status": "Delivered", "delivered_date": "2026-03-28", "courier": "BlueDart"},
        "ORD-552": {"status": "In Transit", "eta": "2026-04-09",           "courier": "Delhivery"},
        "ORD-663": {"status": "Processing", "eta": "2026-04-10",           "courier": "Pending"},
    }
    return orders.get(order_id, {"status": "Not Found", "order_id": order_id})

def check_return_eligibility(order_id: str) -> dict:
    """Check if a ShopMax India order is eligible for return."""
    eligibility = {
        "ORD-441": {"eligible": True,  "reason": "Within 10-day return window, unopened"},
        "ORD-552": {"eligible": False, "reason": "Order not yet delivered"},
        "ORD-663": {"eligible": False, "reason": "Order still processing"},
    }
    return eligibility.get(order_id, {"eligible": False, "reason": "Order not found"})

def get_refund_status(order_id: str) -> dict:
    """Get the refund status for a returned ShopMax India order."""
    refunds = {
        "ORD-441": {"refund_status": "Initiated", "amount_rs": 74999, "eta_days": 5,
                    "method": "Original payment method"},
    }
    return refunds.get(order_id, {"refund_status": "No refund found for this order"})

def create_support_ticket(user_id: str, order_id: str, issue_type: str,
                          issue_summary: str, priority: str) -> dict:
    """Create a ShopMax India support ticket and route to the appropriate team."""
    routing = {
        "delivery_dispute":    "Logistics Team",
        "payment_issue":       "Finance Team",
        "product_defect":      "Quality Team",
        "account_issue":       "Trust and Safety Team",
        "general_complaint":   "Senior Support Team",
    }
    ticket_id = "TKT-" + uuid.uuid4().hex[:6].upper()
    assigned_team = routing.get(issue_type, "Senior Support Team")
    return {
        "ticket_id": ticket_id,
        "user_id": user_id,
        "order_id": order_id,
        "assigned_team": assigned_team,
        "priority": priority,
        "summary": issue_summary,
        "sla_hours": 4 if priority == "HIGH" else 24
    }

agent = LlmAgent(
    name="shopmax_support_agent",
    model="gemini-2.0-flash",
    tools=[
        FunctionTool(get_order_status),
        FunctionTool(check_return_eligibility),
        FunctionTool(get_refund_status),
        FunctionTool(create_support_ticket),
    ],
    instruction=(
        "You are a ShopMax India customer support agent. "
        "Resolve queries using your tools where possible. "
        "If the issue cannot be resolved automatically, or the customer is frustrated, "
        "create a support ticket with the correct issue_type and priority. "
        "Priority is HIGH for delivery failures, payment issues, and urgent complaints."
    )
)

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

queries = [
    ("u_kiran", "Where is my order ORD-552?"),
    ("u_deepa", "I want to return order ORD-441 - it was delivered yesterday"),
    ("u_varun", "I was charged twice for order ORD-441 and nobody is helping me - this is very frustrating!"),
]

for uid, query in queries:
    session = session_service.create_session(app_name="shopmax", user_id=uid)
    events = list(runner.run(
        user_id=uid,
        session_id=session.id,
        new_message=Content(parts=[Part(text=query)])
    ))
    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(f"Q: {query[:60]}")
                    print(f"A: {part.text.strip()[:200]}")
                    print()

It gives the following output,

Q: Where is my order ORD-552?
A: Your order ORD-552 is currently in transit with Delhivery and is expected
to be delivered by 9th April 2026. You can track it on the Delhivery website
using your order ID.

Q: I want to return order ORD-441 - it was delivered yesterday
A: Great news! Order ORD-441 is eligible for return - it is within the 10-day
return window and marked as unopened. I have checked your refund status:
Rs 74,999 refund has been initiated and will reach your original payment
method within 5 days.

Q: I was charged twice for order ORD-441 and nobody is helping...
A: I am very sorry to hear about the double charge on ORD-441. This is
unacceptable and I have escalated this immediately.
Ticket TKT-B3F9A1 created - assigned to Finance Team (HIGH priority, 4-hour SLA).
A senior agent will contact you within 4 hours to resolve the payment issue.

The agent resolved the first two queries automatically using its tools. For the third query - a frustrated customer with a payment dispute - it detected the urgency, created a HIGH priority ticket, routed it to the Finance Team with a 4-hour SLA, and communicated this clearly to the customer. The escalation happened without the customer needing to ask for it.

For ShopMax India production, connect create_support_ticket to your CRM (Freshdesk, Zendesk, or ServiceNow) API so tickets appear immediately in the agent queue with full context. Integrate the agent with WhatsApp Business API so customers can reach ShopMax India support via WhatsApp - the most widely used messaging platform in India. Add sentiment analysis on each message to automatically adjust the priority upward when negative sentiment is detected, ensuring the most frustrated customers get the fastest response.

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