Using AI to create real estate pro formas involves describing property details (address, type, unit count), financial assumptions (purchase price, rents, expenses), and deal structure (debt terms, hold period, exit strategy) in a structured prompt. The AI generates a multi-tab Excel model with rent roll, operating budget, NOI calculations, and return metrics—delivering in seconds what traditionally requires hours of manual construction.
Need to learn more under the hood? See How AI Processes and Outputs Excel Files.
What Goes Into a Pro Forma
A real estate pro forma projects property cash flows over the investment hold period. Unlike corporate income statements that focus on profitability, pro formas focus on Net Operating Income (NOI)—the property's cash-generating capacity before debt service. NOI drives property valuation (Value = NOI ÷ Cap Rate) and debt capacity (lenders require minimum Debt Service Coverage Ratios).
The rent roll details rental income. For multifamily properties, this lists: Unit Number, Unit Type (1BR, 2BR, 3BR), Square Footage, Current Rent, Market Rent. For 150-unit properties, the rent roll has 150 rows. For office properties, the rent roll lists tenants: Tenant Name, Lease Start, Lease End, Rentable SF, Rent per SF, Annual Rent. Total rental income sums all units or tenants.
Vacancy assumptions reduce gross potential income. A property with $5M potential rent at 100% occupancy but 8% average vacancy produces $4.6M effective gross income. Vacancy rates vary by market and property quality: Class A multifamily in strong markets may assume 4-5% vacancy; Class C may assume 10-12%.
Operating expenses fall into categories: property management (typically 3-5% of effective gross income), repairs and maintenance ($400-$800 per unit annually for multifamily), property taxes (1-2% of property value, varies by jurisdiction), insurance, utilities (if owner-paid), administrative expenses. Total operating expenses typically run 35-50% of effective gross income, depending on property type and net lease structure.
NOI equals effective gross income minus operating expenses. This is the key metric. A property generating $4.6M effective gross income with $1.8M operating expenses produces $2.8M NOI. This NOI determines property value at exit (if exit cap is 6%, value = $2.8M ÷ 0.06 = $46.7M) and supports debt capacity (if lenders require 1.25 DSCR and debt service is $1.8M, max annual debt service = $2.8M ÷ 1.25 = $2.24M).
Cash flow after debt service distributes to equity investors. If NOI is $2.8M and debt service is $2.0M (principal + interest), cash flow available for distribution is $800K. This cash flow, plus sale proceeds at exit (after paying off debt), determines investor returns (IRR, equity multiple, cash-on-cash yield).
Capital expenditures reserve funds for future improvements. Lenders often require capex reserves: $250-$400 per unit annually for multifamily. These funds accumulate for future roof replacement, HVAC upgrades, unit renovations. Capex reduces distributable cash (investors receive cash flow minus capex reserve contributions).
Exit assumptions model property sale. Exit typically occurs after a hold period (5-10 years for value-add, 7-15 years for core). Exit value = Stabilized NOI ÷ Exit Cap Rate. If Year 7 NOI is $3.2M and exit cap is 6.5%, sale price = $49.2M. Sale costs (broker commissions, legal fees) are 2-3% of sale price. Net sale proceeds = Sale Price - Remaining Debt - Sale Costs. This goes to equity investors.
Specification (Defining Assumptions)
Specification is the art of precisely describing what you want the AI to build. Ambiguous specifications produce generic outputs. Clear specifications produce models that match your deal structure and firm conventions.
Define property characteristics first. Provide: Property Type (multifamily, office, retail, industrial, mixed-use), Asset Class (Class A, B, or C), Location (city, submarket), Total Units or Square Footage, Year Built, Current Occupancy Rate. Example: "Class B multifamily property, 200 units, Denver metro, built 1998, current occupancy 88%."
Specify acquisition economics. State: Purchase Price, Acquisition Costs (due diligence, legal, closing = typically 2-3% of price), Total Capital Required (price + costs). Example: "Purchase price $32,000,000, acquisition costs $960,000 (3%), total capital required $32,960,000."
Define current income. Provide: Current Average Rent (per unit for multifamily, per SF for commercial), Current Effective Gross Income (actual rent collected), Current Operating Expense Ratio (operating expenses as % of EGI). Example: "Current average rent $1,450/month per unit, current EGI $3.12M annually, current OpEx ratio 42%."
Describe growth assumptions. Specify: Annual Rent Growth Rate (2-4% typical for stabilized, higher for value-add), Occupancy Stabilization Target (if improving occupancy, state target and timeline: "Increase from 88% to 95% over 18 months"), Expense Growth Rate (often 2-3%, tracking inflation). Example: "Rent growth 3% annually, stabilize occupancy to 95% by Month 18, expense growth 2.5% annually."
Detail value-add plans if applicable. State: Renovation Budget (per unit or total), Renovation Timeline (how many units per month), Rent Increase Post-Renovation (how much renovated units command above market). Example: "Renovate 120 of 200 units at $12,000/unit over 24 months, renovated units rent for $1,700/month (17% premium over non-renovated)."
Specify debt structure. Provide: Loan-to-Value or Loan-to-Cost Ratio, Interest Rate, Amortization Period, Interest-Only Period (if applicable), Loan Term, Prepayment Provisions. Example: "75% LTV, 6.25% interest, interest-only for 2 years then 25-year amortization, 10-year term, no prepayment penalty after Year 3."
Define exit parameters. State: Hold Period, Exit Cap Rate, Sale Costs (as % of sale price). Example: "7-year hold, exit at 6.75% cap rate, 2.5% sale costs."
Describe return structure. Provide: Equity Split (LP and GP ownership percentages), Preferred Return (rate, compounding or simple, to which party), Promote Structure (IRR hurdles and split changes). Example: "90% LP / 10% GP, 8% annual pref to LP (compounding), promote: above 15% IRR, split shifts to 70/30 LP/GP."
Describing Your Deal to AI
Organize information into a structured prompt. Use clear sections and bullet points. The AI parses structured text more reliably than paragraph prose.
Example Prompt:
"Build a 7-year real estate pro forma for the following multifamily acquisition:
Property:
- Name: Horizon Apartments
- Location: Charlotte, NC
- Units: 180 (60 1BR, 90 2BR, 30 3BR)
- Year Built: 2005
- Current Occupancy: 89%
Acquisition:
- Purchase Price: $28,500,000
- Closing Costs: $855,000 (3%)
- Total Investment: $29,355,000
Current Income:
- 1BR Rent: $1,250/month
- 2BR Rent: $1,450/month
- 3BR Rent: $1,650/month
- Other Income (parking, pets): $45,000/year
- Current Effective Gross Income: $3,024,000
Operating Expenses:
- Property Management: 4% of EGI
- Repairs & Maintenance: $650/unit/year
- Property Taxes: $427,500/year
- Insurance: $120,000/year
- Utilities: $95,000/year
- Administrative: $60,000/year
Assumptions:
- Stabilize occupancy to 94% by Month 12
- Rent growth: 2.5% annually after stabilization
- Expense growth: 2.5% annually
- Capex reserve: $300/unit/year
Financing:
- LTV: 70%
- Loan Amount: $19,950,000 (70% of purchase price)
- Interest Rate: 6.0%
- Terms: Interest-only for 3 years, then 25-year amortization
- Loan Term: 10 years (balloon payment at end)
Exit:
- Hold Period: 7 years
- Exit Cap Rate: 6.25%
- Sale Costs: 2.5% of sale price
Returns:
- Equity: $9,405,000 (total investment - loan)
- Structure: 85% LP / 15% GP
- Preferred Return: 8.5% annually to LP (compounding)
- Promote: Above 16% IRR, split becomes 70% LP / 30% GP
Build a model with tabs for: Assumptions, Rent Roll, Operating Pro Forma, Debt Schedule, Cash Flow Waterfall, Returns Summary."
This prompt provides everything the AI needs: property details, current financials, forward assumptions, debt terms, exit parameters, and return structure. The AI generates a complete model matching these specifications.
Specifying Structure
Structure specification defines tabs, rows, columns, and calculation flow. Without structure guidance, the AI applies generic conventions. With structure guidance, the model matches your firm's standards.
Define tab names and sequence. "Create tabs in this order: (1) Assumptions, (2) Rent Roll, (3) Operating Budget, (4) NOI Summary, (5) Debt Schedule, (6) Cash Waterfall, (7) Returns." The AI creates tabs with these exact names in this sequence.
Specify column organization. "Use columns as follows: Column A = Line Item Labels, Column B = Year 0 (acquisition), Columns C-I = Years 1-7, Column J = Exit (Year 7 sale)." Or for monthly models: "Columns B-M = Months 1-12 (Year 1), Columns N-Y = Months 13-24 (Year 2)."
Define row organization. "Rent Roll should have these sections: (1) Unit Mix Summary (rows 5-10), (2) Current Rents by Unit Type (rows 12-15), (3) Market Rents (rows 17-20), (4) Projected Rents (rows 22-50, one row per unit type per year), (5) Total Rental Income (row 52)."
Specify calculation placement. "Place all growth rate assumptions in Assumptions tab, Column B, rows 10-20. Place all dollar value assumptions (purchase price, loan amount) in Column D. Place all percentage assumptions (LTV, occupancy) in Column F."
Request specific formula structures. "Calculate NOI as: Effective Gross Income - Operating Expenses. Calculate Effective Gross Income as: Potential Gross Income × (1 - Vacancy Rate) + Other Income. Use this exact structure." The AI replicates your specified formulas.
Define formatting preferences. "Format all currency as thousands with one decimal ($1,250.5K). Format all percentages with one decimal (8.5%). Bold all section headers. Use borders to separate sections."
Reviewing the Generated Model
Open the Assumptions tab. Verify every input: does purchase price match your specification? Does LTV match? Does interest rate match? If anything is incorrect, note it. Either correct manually or prompt: "Change purchase price from $28M to $28.5M."
Review the Rent Roll. Check unit counts: does it show 60 1BR, 90 2BR, 30 3BR as specified? Check rents: does 1BR show $1,250/month? Verify the rent growth formula: Year 2 rent should be Year 1 rent × 1.025 (for 2.5% growth). Confirm total rental income sums correctly.
Inspect the Operating Budget. Verify each expense category calculates correctly. Property management should be 4% of EGI: does the formula show =EGI*0.04? Repairs should be $650 per unit: does it show =180*650 for Year 1? Check that total operating expenses sum all categories.
Check NOI calculation. NOI = EGI - Operating Expenses. Verify this formula in each year. Spot-check one year manually: if Year 1 EGI is $3.15M and Operating Expenses are $1.42M, NOI should be $1.73M. Does the model show this?
Examine the Debt Schedule. Verify loan amount: 70% of $28.5M purchase = $19.95M. Check Year 1-3 show interest-only (principal = $0, interest = $19.95M × 6% = $1.197M). Check Year 4 shows amortization starting (principal > $0). Verify ending balance reduces each year after amortization begins.
Review the Cash Waterfall. Trace distributions for one year manually. If Year 5 cash flow after debt service is $900K, LP pref accrued is $500K, remaining $400K splits 85/15, verify: LP receives $500K + $340K = $840K, GP receives $60K. Confirm the model's formulas produce this.
Check the Returns Summary. Verify IRR calculation references the correct cash flow range: initial equity investment (negative) + annual cash distributions + sale proceeds. The dates should match cash flow timing. If the model uses XIRR, confirm the date array aligns with the cash flow array.
Test sensitivity. Change one assumption—e.g., increase exit cap from 6.25% to 6.50%. Verify dependent cells recalculate: exit value should decrease, sale proceeds decrease, final year cash flow to LP/GP decreases, IRR decreases. If the model updates correctly, formulas are linked properly.
From AI Draft to Final Model
The AI-generated model is a draft, not a final deliverable. Treat it as 80-90% complete. The remaining 10-20% requires human expertise: adding firm-specific nuances, adjusting for deal-specific features, verifying compliance with underwriting standards.
Customize for non-standard features. If the deal has a mezzanine loan (the AI generated only senior debt), add a row in the Debt Schedule for mezzanine: amount, rate, payment structure. If the property has a ground lease (land rent paid to landowner), add a row in Operating Expenses: ground rent, escalation terms.
Add sensitivity tables. The AI may generate a basic model without sensitivities. Prompt: "Add a two-way sensitivity table showing IRR sensitivity to exit cap rate (rows: 5.75%, 6.0%, 6.25%, 6.5%, 6.75%) and rent growth (columns: 2%, 2.5%, 3%, 3.5%, 4%)." The AI adds the table, correctly linked to the Returns Summary.
Insert charts for presentations. Request: "Add a bar chart showing annual cash flow to LP and GP." Or "Add a line chart showing NOI growth over the hold period." Charts make investment committee presentations more visual.
Annotate assumptions with sources. In the Assumptions tab, add comments or adjacent cells noting where assumptions came from: "Market rent based on CoStar comps, 3/15/2026," "Exit cap based on recent sales: 123 Main St (6.1%), 456 Oak Ave (6.4%)." This documentation supports underwriting rigor.
Run peer review. Have a colleague review the model. They check formulas you might have missed, verify calculations match deal terms, and confirm outputs are reasonable. Two sets of eyes catch more errors than one.
Prepare backup scenarios. Generate a base case, upside case (optimistic assumptions), and downside case (conservative assumptions). The upside might assume 3.5% rent growth and 6.0% exit cap. The downside might assume 2.0% rent growth and 6.75% exit cap. Present all three scenarios to investment committee.
Save the final model with proper naming. Use a convention: Property_Name_Date_Version.xlsx → "Horizon_Apartments_2026-03-15_v3.xlsx." Include the date so you know when the model was created. Include version number if iterating.
