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

	Facebook
	Twitter
	Linkedin
	E-Mail

Generative AI > Google Gemini API > Sales Pipeline Agent

Sales Pipeline Agent

Author: Venkata Sudhakar

Sales managers at ShopMax India need real-time visibility into their B2B pipeline - which deals are on track, which are stalling, and whether the team will hit quota. A Sales Pipeline Agent consolidates deal data, identifies risk, and generates a clear weekly forecast with recommended actions - saving hours of manual CRM review.

This tutorial builds a Sales Pipeline Agent using Google ADK and Gemini 2.0 Flash. The agent analyses open deals by stage, flags overdue follow-ups, and produces a revenue forecast with confidence levels.

The below example shows the Sales Pipeline Agent analysing ShopMax India's B2B pipeline for the Mumbai region.

from google.adk.agents import Agent
from google.adk.tools import FunctionTool
from google.adk.runners import Runner
from google.adk.sessions import InMemorySessionService
from google.genai.types import Content, Part
from datetime import date, timedelta

# Simulated pipeline data
PIPELINE = [
    {"id": "D001", "company": "Reliance Retail", "value_lakh": 85, "stage": "proposal",
     "probability": 60, "last_activity_days_ago": 3, "owner": "Priya"},
    {"id": "D002", "company": "Tata Motors Vendor", "value_lakh": 120, "stage": "negotiation",
     "probability": 80, "last_activity_days_ago": 12, "owner": "Arjun"},
    {"id": "D003", "company": "Metro Cash and Carry", "value_lakh": 45, "stage": "demo_done",
     "probability": 40, "last_activity_days_ago": 8, "owner": "Sneha"},
    {"id": "D004", "company": "Flipkart Seller Hub", "value_lakh": 200, "stage": "negotiation",
     "probability": 70, "last_activity_days_ago": 2, "owner": "Priya"},
    {"id": "D005", "company": "Hyderabad Medtech", "value_lakh": 30, "stage": "proposal",
     "probability": 35, "last_activity_days_ago": 18, "owner": "Arjun"},
]

def get_pipeline_summary(region: str) -> dict:
    """Return pipeline data for the given region."""
    total_value = sum(d["value_lakh"] for d in PIPELINE)
    weighted_value = sum(d["value_lakh"] * d["probability"] / 100 for d in PIPELINE)
    stalled = [d for d in PIPELINE if d["last_activity_days_ago"] > 7]
    hot_deals = [d for d in PIPELINE if d["probability"] >= 70]

return {
        "region": region,
        "total_deals": len(PIPELINE),
        "total_pipeline_lakh": total_value,
        "weighted_forecast_lakh": round(weighted_value, 1),
        "stalled_deals": [{"id": d["id"], "company": d["company"],
                            "days_inactive": d["last_activity_days_ago"],
                            "owner": d["owner"]} for d in stalled],
        "hot_deals": [{"id": d["id"], "company": d["company"],
                        "value_lakh": d["value_lakh"],
                        "stage": d["stage"]} for d in hot_deals]
    }

def get_stage_breakdown() -> dict:
    """Return count and value by deal stage."""
    stages = {}
    for d in PIPELINE:
        s = d["stage"]
        if s not in stages:
            stages[s] = {"count": 0, "total_lakh": 0}
        stages[s]["count"] += 1
        stages[s]["total_lakh"] += d["value_lakh"]
    return stages

pipeline_tool = FunctionTool(func=get_pipeline_summary)
stage_tool = FunctionTool(func=get_stage_breakdown)

pipeline_agent = Agent(
    model="gemini-2.0-flash",
    name="sales_pipeline_agent",
    instruction="""
        You are a Sales Pipeline Agent for ShopMax India.
        When asked for a pipeline review:
        1. Fetch the pipeline summary for the region
        2. Get the stage breakdown
        3. Produce a structured weekly report:
           - Total pipeline and weighted forecast in Rs lakhs
           - Stage funnel summary
           - Stalled deals list with recommended follow-up actions
           - Hot deals to prioritise this week
           - Overall quota risk assessment (Green/Amber/Red)
        Use clear headings. Keep it actionable.
    """,
    tools=[pipeline_tool, stage_tool]
)

session_service = InMemorySessionService()
runner = Runner(agent=pipeline_agent, app_name="pipeline_review", session_service=session_service)

session = session_service.create_session(app_name="pipeline_review", user_id="sales_mgr")
message = Content(role="user", parts=[Part(text="Generate weekly pipeline review for Mumbai region. Quota target is Rs 300 lakhs this month.")])
for event in runner.run(user_id="sales_mgr", session_id=session.id, new_message=message):
    if event.is_final_response():
        print(event.content.parts[0].text)

It gives the following output,

WEEKLY PIPELINE REVIEW - MUMBAI REGION
========================================

Pipeline Summary
  Total Deals: 5
  Total Pipeline Value: Rs 480 lakhs
  Weighted Forecast:    Rs 305 lakhs

Stage Breakdown
  Negotiation : 2 deals | Rs 320 lakhs
  Proposal    : 2 deals | Rs 115 lakhs
  Demo Done   : 1 deal  | Rs  45 lakhs

Stalled Deals (no activity > 7 days)
  D002 - Tata Motors Vendor (12 days) - Owner: Arjun
    Action: Schedule negotiation call; risk of competitor entry
  D003 - Metro Cash and Carry (8 days) - Owner: Sneha
    Action: Send ROI summary; push for commercial discussion
  D005 - Hyderabad Medtech (18 days) - Owner: Arjun
    Action: Qualify or disqualify - 18 days silence is a risk signal

Hot Deals to Prioritise
  D004 - Flipkart Seller Hub | Rs 200 lakhs | Negotiation stage
  D002 - Tata Motors Vendor  | Rs 120 lakhs | Negotiation stage

Quota Risk: AMBER
Weighted forecast (Rs 305 L) slightly exceeds quota (Rs 300 L),
but 3 stalled deals reduce confidence. Close D004 this week.

The Sales Pipeline Agent can be scheduled to run every Monday morning and post results to a Slack channel or email distribution list. Connect it to your live CRM via API to replace the simulated data, and extend it to compare pipeline health across regions, teams, or product lines.

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