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

	Facebook
	Twitter
	Linkedin
	E-Mail

Generative AI > Google Gemini API > Tax Provision Agent

Tax Provision Agent

Author: Venkata Sudhakar

Quarterly tax provisioning ensures that businesses set aside the correct amount for income tax liability throughout the year. Under-provisioning leads to a large unexpected tax payment at year-end, while over-provisioning ties up working capital unnecessarily.

A Tax Provision Agent uses Gemini AI to calculate taxable income from book profit, apply applicable deductions, compute the tax liability, and estimate advance tax instalments due under the Indian Income Tax Act. It also flags timing differences that create deferred tax assets or liabilities.

The below example shows how ShopMax India calculates its Q4 FY24 tax provision and advance tax requirement to avoid interest under Section 234B and 234C.

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

# FY24 financial data for tax provision
FY24_DATA = {
    "year": "FY2023-24",
    "book_profit_lakhs": 148.5,
    "permanent_differences": {
        "entertainment_disallowed": 2.8,
        "donation_80g_deduction": -3.5,
    },
    "timing_differences": {
        "depreciation_book_excess": 8.2,
        "provision_for_bad_debt": 4.5,
    },
    "tax_rate_pct": 25.168,  # 22% + surcharge + cess for domestic company
    "advance_tax_paid_lakhs": 28.0,
    "advance_tax_instalments": [
        {"due": "15-Jun", "pct_of_tax": 15},
        {"due": "15-Sep", "pct_of_tax": 45},
        {"due": "15-Dec", "pct_of_tax": 75},
        {"due": "15-Mar", "pct_of_tax": 100},
    ],
}

def calculate_tax_provision(data_json: str) -> str:
    """Compute taxable income and tax provision from book profit."""
    data = json.loads(data_json)
    book_profit = data["book_profit_lakhs"]
    perm = data["permanent_differences"]
    timing = data["timing_differences"]
    rate = data["tax_rate_pct"] / 100
    advance_paid = data["advance_tax_paid_lakhs"]

# Taxable income
    perm_adj = sum(perm.values())
    timing_adj = sum(timing.values())
    taxable_income = book_profit + perm_adj + timing_adj

# Current tax
    current_tax = taxable_income * rate

# Deferred tax liability (timing diffs create DTA/DTL)
    deferred_tax = timing_adj * rate

# Total tax provision = current tax
    tax_provision = current_tax

# Advance tax analysis
    tax_due = max(0, tax_provision - advance_paid)

instalments = []
    for inst in data["advance_tax_instalments"]:
        required = round(tax_provision * inst["pct_of_tax"] / 100, 2)
        instalments.append({"due": inst["due"], "required": required, "pct": inst["pct_of_tax"]})

return json.dumps({
        "year": data["year"],
        "book_profit": book_profit,
        "permanent_adj": round(perm_adj, 2),
        "timing_adj": round(timing_adj, 2),
        "taxable_income": round(taxable_income, 2),
        "tax_rate_pct": data["tax_rate_pct"],
        "current_tax": round(current_tax, 2),
        "deferred_tax_liability": round(deferred_tax, 2),
        "total_provision": round(tax_provision, 2),
        "advance_tax_paid": advance_paid,
        "balance_tax_payable": round(tax_due, 2),
        "instalment_schedule": instalments,
    })

tax_tool = FunctionTool(func=calculate_tax_provision)

AGENT_INSTRUCTION = """
You are a Tax Provision Agent for ShopMax India finance.
When given financial data for tax computation:
1. Call calculate_tax_provision with the data
2. Show the tax computation bridge: book profit -> permanent adjustments -> timing adjustments -> taxable income -> tax at rate
3. State the total tax provision required and advance tax already paid
4. Show the advance tax instalment schedule for the year
5. Warn if advance tax paid is below 90% of total tax liability (Section 234B risk)
6. Note the deferred tax liability created by timing differences
Format as a tax provision memo for the CFO and company CA.
"""

tax_agent = LlmAgent(
    name="tax_provision_agent",
    model="gemini-2.0-flash",
    instruction=AGENT_INSTRUCTION,
    tools=[tax_tool],
)

async def main():
    session_service = InMemorySessionService()
    runner = Runner(
        agent=tax_agent,
        app_name="tax_app",
        session_service=session_service
    )
    session = await session_service.create_session(
        app_name="tax_app",
        user_id="cfo"
    )
    message = Content(parts=[Part(text="Calculate FY24 tax provision: " + json.dumps(FY24_DATA))])
    print("Tax Provision Agent - ShopMax India")
    print("=" * 40)
    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,

Tax Provision Agent - ShopMax India
========================================
FY2023-24 TAX PROVISION MEMO

TAX COMPUTATION BRIDGE (Rs Lakhs)
Book Profit                       : 148.50
Add: Entertainment disallowed     :   2.80  (permanent)
Less: 80G donation deduction      :  -3.50  (permanent)
Add: Excess book depreciation     :   8.20  (timing)
Add: Bad debt provision           :   4.50  (timing)
                                    ------
Taxable Income                    : 160.50
Tax @ 25.168%                     :  40.39

TAX PROVISION SUMMARY
Total Tax Provision Required : Rs 40.39 Lakhs
Advance Tax Already Paid     : Rs 28.00 Lakhs
Balance Tax Payable (Mar 15) : Rs 12.39 Lakhs

ADVANCE TAX INSTALMENT SCHEDULE
Due Date  Required Cumulative  % of Tax
--------  --------------------  --------
15-Jun    Rs  6.06 Lakhs        15%
15-Sep    Rs 18.18 Lakhs        45%
15-Dec    Rs 30.29 Lakhs        75%
15-Mar    Rs 40.39 Lakhs       100%

SECTION 234B ALERT
Advance tax paid (Rs 28L) = 69.3% of total liability.
Minimum required: 90% = Rs 36.35L.
Shortfall of Rs 8.35L may attract interest at 1% per month.
Recommend paying Rs 12.39L by 15-Mar to clear liability.

DEFERRED TAX LIABILITY
Timing differences of Rs 12.70L create DTL of Rs 3.20L.
To be recognised in Balance Sheet under Deferred Tax Liability.

The agent computed taxable income by adjusting book profit for both permanent and timing differences, applied the correct effective tax rate including surcharge and cess, and generated a ready-to-use advance tax instalment schedule. The Section 234B alert is particularly valuable - it warns the CFO of a potential interest liability with enough time to make a payment before the March 15 deadline. The deferred tax liability note ensures the balance sheet treatment is correctly flagged for the CA.

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