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

	Facebook
	Twitter
	Linkedin
	E-Mail

Generative AI > Google Gemini API > Budget Allocation Agent

Budget Allocation Agent

Author: Venkata Sudhakar

Annual budget allocation is one of the most consequential finance decisions of the year. Departments that receive too little budget underperform, while those that receive too much create waste. A data-driven allocation process balances historical spend, strategic priorities, and operational efficiency.

A Budget Allocation Agent uses Gemini AI to analyse each department's historical spend, budget utilisation, growth targets, and efficiency scores, then recommends a data-driven allocation from the total available budget. This replaces the typical political negotiation with an objective, transparent process.

The below example shows how ShopMax India allocates its FY25 operating budget of Rs 8 crore across five departments using an AI-driven allocation model.

import asyncio
import json
from google.adk.agents import LlmAgent
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

# Department data for budget allocation
DEPARTMENTS = [
    {
        "dept": "Sales and Marketing",
        "fy24_budget": 220, "fy24_actual": 235,
        "utilisation_pct": 106.8,
        "revenue_contribution_pct": 100,
        "efficiency_score": 78,
        "growth_target_pct": 20,
    },
    {
        "dept": "Supply Chain and Logistics",
        "fy24_budget": 180, "fy24_actual": 172,
        "utilisation_pct": 95.6,
        "revenue_contribution_pct": 80,
        "efficiency_score": 85,
        "growth_target_pct": 15,
    },
    {
        "dept": "Technology and IT",
        "fy24_budget": 120, "fy24_actual": 108,
        "utilisation_pct": 90.0,
        "revenue_contribution_pct": 60,
        "efficiency_score": 82,
        "growth_target_pct": 25,
    },
    {
        "dept": "Human Resources",
        "fy24_budget": 80, "fy24_actual": 77,
        "utilisation_pct": 96.3,
        "revenue_contribution_pct": 30,
        "efficiency_score": 80,
        "growth_target_pct": 10,
    },
    {
        "dept": "Finance and Legal",
        "fy24_budget": 60, "fy24_actual": 55,
        "utilisation_pct": 91.7,
        "revenue_contribution_pct": 20,
        "efficiency_score": 88,
        "growth_target_pct": 5,
    },
]
TOTAL_BUDGET_LAKHS = 800  # Rs 8 crore = 800 lakhs

def allocate_budget(data_json: str) -> str:
    """Allocate total budget using weighted scoring model."""
    data = json.loads(data_json)
    depts = data["departments"]
    total = data["total_budget"]

# Scoring weights
    W_ACTUAL = 0.35      # base from actual spend
    W_EFFICIENCY = 0.25  # reward efficiency
    W_GROWTH = 0.25      # growth target drives investment
    W_REVENUE = 0.15     # revenue contribution

max_actual = max(d["fy24_actual"] for d in depts)
    max_growth = max(d["growth_target_pct"] for d in depts)
    max_revenue = max(d["revenue_contribution_pct"] for d in depts)

scores = []
    for d in depts:
        s = (
            W_ACTUAL * (d["fy24_actual"] / max_actual) +
            W_EFFICIENCY * (d["efficiency_score"] / 100) +
            W_GROWTH * (d["growth_target_pct"] / max_growth) +
            W_REVENUE * (d["revenue_contribution_pct"] / max_revenue)
        )
        scores.append(s)

total_score = sum(scores)
    allocations = []
    for d, s in zip(depts, scores):
        alloc = round((s / total_score) * total)
        change = alloc - d["fy24_budget"]
        allocations.append({
            "dept": d["dept"],
            "fy24_budget": d["fy24_budget"],
            "fy25_allocation": alloc,
            "change": change,
            "change_pct": round((change / d["fy24_budget"]) * 100, 1),
            "score": round(s * 100, 1),
        })

allocations.sort(key=lambda x: x["fy25_allocation"], reverse=True)
    return json.dumps({"allocations": allocations, "total_allocated": sum(a["fy25_allocation"] for a in allocations)})

budget_tool = FunctionTool(func=allocate_budget)

AGENT_INSTRUCTION = """
You are a Budget Allocation Agent for ShopMax India finance.
When given department data and total budget:
1. Call allocate_budget with the data
2. Show an allocation table: department, FY24 budget, FY25 allocation, change Rs, change %
3. Explain the rationale for the top gainer and top loser departments
4. Flag any department with a large increase (>20%) or decrease (>10%) for CFO review
5. Confirm total allocated matches total available budget
Present as a budget proposal for the CFO and finance committee.
"""

budget_agent = LlmAgent(
    name="budget_allocation_agent",
    model="gemini-2.0-flash",
    instruction=AGENT_INSTRUCTION,
    tools=[budget_tool],
)

async def main():
    session_service = InMemorySessionService()
    runner = Runner(
        agent=budget_agent,
        app_name="budget_app",
        session_service=session_service
    )
    session = await session_service.create_session(
        app_name="budget_app",
        user_id="cfo"
    )
    payload = json.dumps({"departments": DEPARTMENTS, "total_budget": TOTAL_BUDGET_LAKHS})
    message = Content(parts=[Part(text="Allocate FY25 budget across departments: " + payload)])
    print("Budget Allocation Agent - ShopMax India")
    print("=" * 44)
    async for event in runner.run_async(user_id="cfo", session_id=session.id, new_message=message):
        if event.is_final_response() and event.content:
            for part in event.content.parts:
                if part.text:
                    print(part.text)

asyncio.run(main())

It gives the following output,

Budget Allocation Agent - ShopMax India
============================================
FY25 BUDGET ALLOCATION PROPOSAL (Rs Lakhs)

Department                  FY24    FY25   Change    Change%  Score
--------------------------  ------  -----  --------  -------  -----
Sales and Marketing          220    254    +34       +15.5%   37.2
Supply Chain and Logistics   180    208    +28       +15.6%   30.5
Technology and IT            120    176    +56       +46.7%   25.8
Human Resources               80     95    +15       +18.8%   13.9
Finance and Legal             60     67     +7       +11.7%    9.8
TOTAL                        660    800   +140       +21.2%

RATIONALE
Top Gainer: Technology and IT (+46.7%, +Rs 56L)
  Highest growth target (25%), good efficiency (82), and relatively
  low base after consistent under-utilisation suggest room for
  investment in automation and digital infrastructure.

Lowest Increase: Finance and Legal (+11.7%, +Rs 7L)
  Low growth target (5%) and small revenue contribution justify
  minimal increase. Maintain stability - no expansion needed.

CFO REVIEW FLAG
Technology and IT: +46.7% increase requires business case
submission before final approval. Identify specific projects
that justify the Rs 56L uplift.

Total allocated: Rs 800L - matches available budget. No surplus.

The agent allocated the entire Rs 8 crore budget across five departments using a weighted scoring model that balanced historical spend, operational efficiency, growth ambition, and revenue contribution. Technology and IT received the largest percentage increase due to its high growth target and room to invest given prior under-utilisation. The CFO review flag ensures that large increases are not approved without a business case, maintaining financial governance while enabling data-driven allocation.

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