Describe deal get excel model is a workflow where you provide a plain-language description of a real estate transaction—including deal structure, cash flows, and key constraints—and receive a complete, working Excel file with all calculations built. This approach eliminates the manual setup of formulas, tabs, and logic blocks by translating your description directly into a functional model.
Relevant Articles
- Need to refine a complex model? See How to Iterate on Excel Models Without Starting Over.
- Want to understand how AI outputs full models? Review AI Outputs Real Excel Files.
- Curious about turning term sheets into models? Check Term Sheet to Excel Model with AI.
Working Example: Project "Mesa Ridge"
To see this workflow in action, let's use a specific deal scenario:
Every number referenced in this article will align with Mesa Ridge. If you see "$1,180,000 NOI," it refers to this deal. If a formula shows a 6.25% debt constant, it's Mesa Ridge's loan.
The Concept
The describe deal get excel model workflow inverts traditional Excel modeling. Instead of opening a blank workbook and manually building rows for cash flow, debt service, and returns, you start by articulating what the model needs to do. You describe the structure, constraints, and logic in natural language—then the AI constructs the Excel file for you.
This is not about generating formulas one at a time or copying cells from ChatGPT into an existing spreadsheet. The output is a complete .xlsx file with multiple tabs, properly linked references, and calculations that mirror the logic you described. You download it, open it, and verify the numbers. The model exists as a standalone artifact.
The core benefit is speed. In our experience reviewing analyst workflows, building a multifamily acquisition model from scratch takes 90 to 180 minutes depending on the waterfall complexity. With the describe-and-get approach, the initial model exists in under 5 minutes. You spend your time reviewing and refining logic, not typing cell references.
This workflow is most effective for deals with clearly defined structures. If you know the entry price, debt terms, waterfall tiers, and exit assumptions, you can articulate them directly. The AI doesn't guess—it builds what you specify. That's why the description step is the lynchpin. A vague description produces a generic model. A precise description produces a precise model.
The describe deal get excel model approach also separates the "what" from the "how." You define the financial logic (what the model should calculate), and the AI handles the implementation (how to structure the formulas and tabs). This division allows analysts to focus on the accuracy of the logic rather than the mechanics of Excel syntax.
How It Works
The workflow has three discrete steps: describe the deal, receive the Excel file, and open the file for review. Each step has specific inputs and outputs.
Step 1: Write the Description
You provide a text description that includes the deal parameters. This can be a structured bullet list, a paragraph of prose, or a mix of both. The critical requirement is completeness. If the waterfall has a lookback provision, state it. If the debt has prepayment penalties, include them. If the model needs a sensitivity table on exit cap and NOI growth, specify the ranges.
For Mesa Ridge, the description might look like this:
"Build a 6-year multifamily acquisition model for a 120-unit asset in Tempe, AZ. Purchase price: $18,600,000. Equity: $6,200,000 (90% LP, 10% GP). Debt: $12,400,000 at 6.25%, interest-only for 5 years, then 30-year amortization. Year 1 NOI: $1,180,000, growing 3.2% annually. Exit at a 5.4% cap rate in Year 6. Waterfall: 8% preferred return to LPs, then 70/30 LP/GP split above the pref. Include a cash flow tab, debt service schedule, and waterfall distribution tab."
This description gives the AI everything it needs to build the model structure. The numbers are specific. The logic is clear. There's no ambiguity about whether the debt is I/O or amortizing, or what the LP/GP split is above the pref.
Step 2: Receive the Excel File
The AI processes the description and outputs a .xlsx file. You download it directly. The file contains multiple tabs (e.g., Inputs, Cash Flow, Debt, Waterfall, Returns) with formulas already linked. The Inputs tab holds the parameters you described (purchase price, NOI, growth rates). The Cash Flow tab references those inputs and calculates annual NOI, debt service, and distributable cash. The Waterfall tab splits the proceeds according to the 8% pref and 70/30 structure.
For Mesa Ridge, the output would include:
- An Inputs tab with cells for $18,600,000 purchase price, $1,180,000 Year 1 NOI, 3.2% NOI growth, 6.25% debt rate, and 5.4% exit cap.
- A Cash Flow tab with columns for Years 0-6, rows for NOI, debt service, and distributable cash.
- A Debt Schedule tab showing interest-only payments for Years 1-5, then amortization in Year 6.
- A Waterfall tab calculating the 8% pref on the $6,200,000 equity, then splitting remaining proceeds 70/30.
- A Returns tab showing LP and GP IRR and equity multiples.
The formulas are standard Excel: =Inputs!B5 * (1 + Inputs!B8)^A3 for NOI growth, =Inputs!B6 * Inputs!B7 for annual debt service during I/O period. The model doesn't use macros or VBA. It's transparent and auditable.
Step 3: Open and Review
You open the file in Excel (or Google Sheets, if the file is .xlsx compatible) and verify the logic. Check that the debt service matches your expectations given the 6.25% rate. Confirm that the waterfall calculates the pref correctly. Trace a few formulas to ensure the references are correct. If the NOI in Year 3 should be $1,180,000 * (1.032)^2 = $1,256,469, confirm the cell shows that value.
This step is not optional. The describe-and-get workflow accelerates the build, but it does not eliminate the need for verification. In our models, we always run a zero test: set all revenues to zero and confirm that the exit proceeds equal zero. If the waterfall still shows distributions, there's a logic error.
When reviewing the Mesa Ridge model, you would check that the Year 1 debt service is $12,400,000 * 6.25% = $775,000 (interest-only), and that the Year 6 debt service includes principal paydown if the loan amortizes. You would verify that the waterfall's pref calculation uses $6,200,000 as the equity base and 8% as the hurdle. If any of these checks fail, you note the issue and refine the model—which is covered in a separate article on iteration.
The describe deal get excel model workflow assumes the first output will be close to correct but may require minor adjustments. The goal is to get 90% of the work done automatically, then spend your time on the 10% that requires judgment.
What to Include in Descriptions
The completeness of your description determines the accuracy of the model. A minimal description produces a minimal model. A detailed description produces a detailed model. The following categories should be included in every deal description.
Entry and Capital Structure
State the purchase price, total equity, and LP/GP split. If there's a GP co-invest or promote structure, specify it. For Mesa Ridge, the entry is $18,600,000 with $6,200,000 equity (90% LP, 10% GP). If the GP were contributing 1% of equity and receiving 10% of ownership (a GP catch-up), note that explicitly.
Also include the debt terms: principal, interest rate, amortization schedule, and any prepayment penalties or extension options. For Mesa Ridge, the debt is $12,400,000 at 6.25%, interest-only for 5 years, then switching to a 30-year amortization schedule. If the loan had a 2% prepayment penalty in Years 1-3, you would state: "Debt includes a 2% prepayment penalty if repaid before Year 3."
Operating Assumptions
Provide the Year 1 NOI and the growth rate (or a schedule of annual NOI if growth is irregular). If you have separate assumptions for revenue and expenses, break them out. For Mesa Ridge, we simplified to $1,180,000 NOI growing at 3.2% per year. If you had a pro forma with unit-level rent assumptions, you could describe it: "120 units at $1,200/month average rent in Year 1, growing 3% annually. Operating expenses: 40% of gross income."
Also specify any capital expenditures, leasing costs, or one-time investments. If Mesa Ridge needed $500,000 in Year 2 for unit renovations, the description would include: "CapEx: $500,000 in Year 2 for interior upgrades."
Exit Assumptions
State the hold period, exit cap rate, and any selling costs. For Mesa Ridge, the exit is Year 6 at a 5.4% cap rate. If selling costs are 3% of gross sale price, note that: "Exit in Year 6 at 5.4% cap, with 3% selling costs." This ensures the model deducts the correct amount from sale proceeds before distributing to LPs and GPs.
If the exit is assumed to be a refinance instead of a sale, describe the new loan terms. "Refinance in Year 5 at 70% LTV and 5.8% interest rate, distributing excess proceeds through the waterfall."
Waterfall Structure
This is where precision matters most. State the number of tiers, the hurdle for each tier, and the LP/GP split at each level. For Mesa Ridge: "2-tier waterfall. Tier 1: Return of capital plus 8% preferred return to LPs, with LP receiving 90% and GP receiving 10%. Tier 2: Remaining proceeds split 70% LP, 30% GP."
If the waterfall has a catch-up (where the GP receives 100% of proceeds until their promote equals a target percentage), describe it: "After 8% pref, GP receives 100% of next proceeds until GP has received 20% of total distributions (the catch-up), then 80/20 split thereafter."
Also specify whether the pref is compounded or simple, and whether it accrues on unreturned capital or the full equity amount. "8% pref compounded annually on unreturned capital balance." This level of detail prevents the AI from defaulting to generic assumptions.
Output Requirements
Tell the AI what tabs or sections you need. "Include an Inputs tab, Cash Flow tab, Debt Schedule, Waterfall, and Returns summary. Add a sensitivity table showing LP IRR across exit cap rates (4.8% to 6.0%) and NOI growth (2.0% to 4.5%)." If you need a specific format (e.g., monthly vs. annual cash flows), state it.
For Mesa Ridge, we requested annual cash flows because the hold is 6 years. If the deal involved construction with monthly drawdowns, you would request monthly periods: "Model construction over 18 months with monthly cash flows, then annual stabilized periods."
The describe deal get excel model workflow is garbage-in, garbage-out. If your description omits the debt amortization schedule, the model will default to interest-only for the full term. If your description doesn't specify exit costs, the model may assume zero. Be explicit.
Example Deal Descriptions
Below are three complete descriptions demonstrating different levels of complexity. Each is specific enough to generate a functional model.
Example 1: Simple Stabilized Multifamily
"Build a 5-year acquisition model for a stabilized 80-unit multifamily property in Austin, TX. Purchase price: $14,000,000. Equity: $4,200,000 (100% LP, no GP co-invest). Debt: $9,800,000 at 5.75%, 30-year amortization. Year 1 NOI: $920,000, growing 2.8% per year. Exit in Year 5 at a 5.2% cap rate with 2.5% selling costs. No waterfall—LP receives 100% of distributions. Include tabs for Inputs, Cash Flow, Debt Schedule, and Returns."
This description has no GP promote, making it the simplest structure. The model will show all distributable cash flowing to the LP, and the returns calculation will show a single IRR and equity multiple.
Example 2: Value-Add with Repositioning CapEx
"Build a 7-year value-add model for a 200-unit Class C multifamily property in Tampa, FL. Purchase price: $28,500,000. Equity: $9,500,000 (92% LP, 8% GP). Debt: $19,000,000 at 6.15%, interest-only for 3 years, then 25-year amortization. Year 1 NOI: $1,650,000. NOI grows 2.5% in Years 1-2, then 5.0% in Years 3-4 (post-renovation), then 3.0% in Years 5-7. CapEx: $1,200,000 in Year 2 for unit upgrades. Exit in Year 7 at a 5.6% cap rate with 2.8% selling costs. Waterfall: Tier 1 returns capital plus 9% pref (compounded annually on unreturned capital), split 92/8 LP/GP. Tier 2: Remaining proceeds split 75/25 LP/GP. Include tabs for Inputs, Cash Flow, CapEx Schedule, Debt Schedule, Waterfall, and Returns. Add a sensitivity table for LP IRR across exit cap (5.2% to 6.0%) and Year 3-4 NOI growth (4.0% to 6.0%)."
This description includes irregular NOI growth (to reflect the repositioning timeline), a CapEx deployment in Year 2, and a 2-tier waterfall with a compounded pref. The model will need to track unreturned capital to calculate the pref correctly.
Example 3: Development Project with Phased Equity Calls
"Build a 4-year ground-up development model for a 150-unit multifamily project in Denver, CO. Total development cost: $42,000,000. Equity: $14,000,000 (85% LP, 15% GP), called in three phases: $4,000,000 at Year 0, $6,000,000 at Year 1, $4,000,000 at Year 2. Debt: $28,000,000 construction loan at 7.5% (interest reserves funded upfront), converting to permanent debt at 6.0% (30-year amortization) at Year 2.5 (stabilization). Construction period: Years 0-2 (no NOI). Lease-up: Year 3 (60% of stabilized NOI). Stabilized: Year 4 onward at $2,200,000 NOI, growing 3.5% annually. Exit in Year 7 at a 5.0% cap rate with 3% selling costs. Waterfall: Tier 1 returns capital plus 10% pref (simple interest on total equity from date of each capital call), split 85/15. Tier 2: 70/30 LP/GP split. Include tabs for Inputs, Equity Call Schedule, Construction Budget, Debt Schedule (construction and permanent), Cash Flow, Waterfall, and Returns. Add an LP IRR sensitivity across stabilized NOI ($2,000,000 to $2,400,000) and exit cap (4.6% to 5.4%)."
This description introduces phased equity calls (requiring the model to track the timing of each capital contribution for pref calculation), a construction-to-permanent loan structure, and a lease-up period with partial NOI. It's the most complex of the three examples, but the description provides all necessary parameters.
Each of these examples would generate a distinct Excel model tailored to the specific deal structure. The first would have 5 columns (Years 0-4 plus exit), the second would have 7 columns with a CapEx row, and the third would have 8 columns with separate construction and operating periods.
The common thread: every parameter is quantified, every assumption is stated, and every structural choice is explicit. That's what "describe deal get excel model" requires.
Reviewing What You Get
When you open the Excel file, your first task is to verify that the model matches your description. This is not about checking every formula—it's about confirming the structure, key calculations, and overall logic.
Check the Inputs Tab
Open the Inputs tab and confirm that the hard-coded values match your description. For Mesa Ridge, you should see:
- Purchase Price: $18,600,000
- Total Equity: $6,200,000
- LP%: 90%
- GP%: 10%
- Debt Principal: $12,400,000
- Debt Rate: 6.25%
- Year 1 NOI: $1,180,000
- NOI Growth: 3.2%
- Exit Cap: 5.4%
- Hold Period: 6 years
If any of these values are incorrect (e.g., the model shows 6.5% debt rate instead of 6.25%), note the discrepancy. Small input errors cascade through the model, so fix them before reviewing the calculations.
Trace a Key Formula
Pick a high-stakes calculation and trace the formula. For Mesa Ridge, a logical choice is the Year 1 debt service. The model should calculate $12,400,000 * 6.25% = $775,000 (interest-only). Open the Cash Flow tab, find the Year 1 debt service cell, and check the formula. It should reference the Inputs tab: =Inputs!$B$6 * Inputs!$B$7 (assuming B6 is debt principal and B7 is debt rate).
If the formula is correct, the cell should display $775,000. If it displays a different number, either the formula is wrong or the inputs are wrong. Fix the error before proceeding.
Verify the Waterfall Logic
The waterfall is the most error-prone section because it involves conditional logic (IF statements or tiered splits). For Mesa Ridge, the waterfall should:
- Return $6,200,000 of capital to LPs and GPs (90/10 split).
- Pay an 8% preferred return on the $6,200,000 (or on the unreturned capital balance if compounded).
- Split remaining proceeds 70/30 LP/GP.
Open the Waterfall tab and check that Tier 1 shows the correct pref calculation. If the model assumes simple interest, the pref should be $6,200,000 * 8% * 6 years = $2,976,000. If compounded, the calculation is more complex: the pref accrues annually on the unreturned capital balance, so Year 1 pref is $6,200,000 * 8% = $496,000, Year 2 pref is ($6,200,000 - distributable cash from Year 1) * 8%, and so on.
Confirm that the LP receives 90% and the GP receives 10% of the Tier 1 distributions, and that Tier 2 splits the excess 70/30. If the model shows a different split, there's a logic error.
Run a Zero Test
Set all operating assumptions to zero (Year 1 NOI = $0, exit cap = 0%, or set exit sale price to equal the remaining debt balance). The model should show zero distributable cash and zero waterfall distributions. If the waterfall still shows positive distributions to the GP, the formulas are referencing a hard-coded value instead of a calculated result.
This test catches errors where the model accidentally includes a static number (e.g., "$500,000 GP promote") instead of deriving it from the cash flow.
Check the Returns Calculation
The Returns tab should show the LP IRR and GP IRR based on their respective cash flows. For the LP, the IRR should be calculated on the $5,580,000 equity contribution (90% of $6,200,000) as the outflow, and the LP's share of annual distributions plus exit proceeds as the inflows. For the GP, the IRR should use $620,000 equity as the outflow.
Use Excel's XIRR or IRR function to verify the result. If the model shows an LP IRR of 18.2%, manually calculate the IRR using the LP cash flows. If your manual calculation shows 17.9%, there's a formula error in the Returns tab.
The review process for Mesa Ridge should take 10-15 minutes if the model is correct. You're not rebuilding the model—you're auditing it. If the audit passes, the model is ready to use. If it fails, you move to the refinement step.
Refining the Output
Most models require at least one refinement after the initial output. The describe deal get excel model workflow delivers a functional model, but "functional" doesn't always mean "perfect." Refinements fall into three categories: input corrections, logic adjustments, and output enhancements.
Input Corrections
If the Inputs tab contains an incorrect value, you can fix it manually or regenerate the model with a corrected description. For Mesa Ridge, if the model shows a 6.5% debt rate instead of 6.25%, you have two options:
- Open the Inputs tab and change cell B7 from 6.5% to 6.25%. All formulas referencing B7 will update automatically.
- Revise your description to emphasize "Debt: $12,400,000 at exactly 6.25%" and regenerate the file.
Option 1 is faster for single-value corrections. Option 2 is better if multiple inputs are wrong, because it's easier to fix the description than to manually edit ten cells.
Logic Adjustments
If the model's logic doesn't match your intent, you need to specify the correction. For Mesa Ridge, suppose the model calculated the pref as simple interest, but you intended compounded interest. You would need to either:
- Manually adjust the Waterfall tab to compound the pref (replacing the simple interest formula with a running balance calculation).
- Update your description to state "8% preferred return, compounded annually on unreturned capital" and regenerate.
Logic adjustments are harder to make manually because they often involve restructuring multiple formulas. If the change is significant (e.g., switching from a 2-tier to a 3-tier waterfall), regenerating the model is more reliable than editing by hand.
However, for minor tweaks—such as changing the exit cap from 5.4% to 5.3%—manual edits are faster. Just update the Inputs tab and verify that the downstream formulas recalculate correctly.
Output Enhancements
Sometimes the model is logically correct but lacks a feature you need. For Mesa Ridge, suppose you want to add a sensitivity table showing LP IRR across different exit cap rates. You can:
- Manually build the sensitivity table using Excel's Data Table feature (referencing the exit cap input and the LP IRR output).
- Revise your description to include "Add a sensitivity table for LP IRR across exit cap rates from 5.0% to 6.0% in 0.2% increments" and regenerate.
Output enhancements don't change the core logic—they add reporting or analysis layers. If the base model is correct, adding a sensitivity table manually takes 2-3 minutes. If you're regenerating the model anyway (due to logic errors), include the enhancement in the revised description.
The key principle: small fixes can be done manually; large fixes should trigger a regeneration with a revised description. The describe deal get excel model workflow is most effective when you iterate on the description, not on the Excel file itself. Each regeneration produces a clean model with no accumulated manual edits. This keeps the model auditable and prevents the "Frankenstein" problem where half the formulas are AI-generated and half are hand-coded.
For Mesa Ridge, if the initial model has the wrong debt rate and a missing sensitivity table, the best approach is to revise the description to fix both issues and regenerate. The second output will be correct on both dimensions, and you'll have a clean file to work from.
Need to refine a complex model with multiple iterations? See How to Iterate on Excel Models Without Starting Over for the full workflow on managing changes without rebuilding from scratch.
