An AI Skill for the A.CRE Apartment Development Model
I’ve been working on a project to make our library of Excel models AI-ready. The idea is straightforward: pair every A.CRE Excel model with an AI Skill, a packaged set of instructions and reference files that teaches an AI assistant how to operate that specific model on your behalf. The Apartment Development Model is the latest in that effort, and this post introduces the AI Skill we built to accompany it.
Think of this as a sister post to the Apartment Development Model post, which walks through the model itself: the tabs, inputs, outputs, and mechanics. If you haven’t seen that one yet, start there. This post focuses specifically on the AI Skill: what it does, how it works, and how to use it.
- While we refer to these as Claude Skills (the format is Anthropic’s), the instructions inside the Skill are largely platform-neutral. You can use the Skill with Claude, where it integrates most natively, especially via the Claude in Excel add-in, but also with ChatGPT, Gemini, or any other capable AI assistant. Just upload the SKILL.md file alongside the Excel model and the assistant can follow the same playbook. Some integrations are smoother than others, but the underlying knowledge transfers.
What is an AI Skill?
If you’re new to the concept, an AI Skill is a packaged set of instructions and reference files that an AI assistant loads alongside your file. It teaches the assistant things it wouldn’t otherwise know. In this case, that means every input cell on the Underwriting tab, the four user roles the model serves, how the model’s toggles change what each cell means, and the most common ways a ground-up apartment underwriting goes wrong before you ever get to a development spread.
The result is an AI assistant that can actually operate the model on your behalf, rather than one that talks about apartment development in the abstract.
For a primer with a short video tutorial, see our practical guide to Claude Skills.
What the AI Skill Does for You
The Apartment Development Model underwrites a ground-up multifamily project, with optional ground-floor retail, on a monthly basis. It covers the development budget and draw timing, equity plus construction and mezzanine debt sized to a target LTC, lease-up to stabilization, optional permanent debt, a sale at an exit cap rate, and a four-tier GP/LP waterfall with an optional second waterfall splitting the promote between two co-GPs. A few of its mechanics will trip up a general-purpose AI without specific instruction. Equity is solved by a macro so that actual LTC equals the target. Several blue input cells carry formula defaults that look like assumptions but are only conveniences. And the promote column takes the promote share alone, not the total GP split. The Skill knows all of this.
Role Identification
Before the Skill touches a single input, it asks which decision you are making:
- Developer / sponsor, go/no-go: deciding whether the project pencils. The Skill leads with the untrended development spread and yield on cost, then moves to the trended and sale spreads, and then to unlevered and levered returns. It can also use the Residual Land Value tool on the Summary tab to back into a maximum land price.
- Sponsor structuring the GP/LP or co-GP deal: setting fees, promote tiers, and the co-GP split. The Skill walks the four waterfall tiers, confirms how the GP share and promote interact, and flags whether the developer fee is included in GP returns.
- LP / equity investor: pressure-testing a sponsor’s assumptions and the net return. The Skill checks the sponsor’s blue formula defaults, the placeholders left in the file, and whether the number being quoted is the gross return or the LP’s net return.
- Lender, construction / perm: sizing the loans and testing the takeout. The Skill surfaces loan size against the target LTC, the interest carry, and the permanent loan coverage at takeout.
Populating Inputs
You can share a site plan, a land PSA or LOI, zoning and site data, an architect’s unit matrix, a GC budget, rent and absorption comps, a market report, an operating budget or opex comps, tax and insurance data, debt and JV term sheets, or an appraisal or IC memo. The Skill pulls the relevant assumptions and stages them for your review before writing anything.
Every blue cell in the model is treated as a required input you confirm, including the ones that ship with a formula default, such as hard cost at $140,000 per unit or the development fee at 3% of costs. The formula is a convenience, and keeping it or replacing it are both legitimate, but the Skill asks about each one.
For any assumption you have not sourced, the Skill follows a simple rule. If the template has a placeholder, it leaves it and tells you it did, asking you to confirm or replace it with a market-supported number. If there is no placeholder, it makes a defensible best guess and discloses it. It never writes a value without saying where the value came from, and it does not invent industry-norm ranges that age out within a year or two.
The Skill also reads the six toggle flags before anything else, because they change which cells are live. The opex toggle runs in reverse of what you might expect: 1 means Basic and 0 means Detailed. And every column K note in the file is placeholder narrative, so the Skill clears or rewrites each one rather than leaving old deal commentary behind.
Recalibrating After Changes
This model depends on VBA. The Recalculate macro solves the equity input so that actual LTC equals your target, and the toggle buttons, Add Item buttons, and Delete buttons all run macros too. The AI cannot fire any of them. So after the Skill writes inputs in a chat session, it hands you the updated file, and you open it in desktop Excel with macros enabled, click the recalculate button, save, and re-upload. The Skill then confirms the file reads as recalculated before it reports a single number, so you never get a stale result.
The same goes for the sensitivity tables, which do not refresh on their own because the workbook is set to calculate automatically except for tables. The Skill asks you to click Update Sensitivity Tables before it quotes them. In Claude in Excel, the Skill can run a fallback loop on the equity input and check the result, but it hands the job back to you if the check does not clear.
The Skill also does not write the toggle flags or add rows itself. You click the buttons, and the macros restyle, relink, and show or hide the right tabs.
Catching Common Mistakes
Eleven errors show up repeatedly on this model. Blue formula defaults taken as real assumptions. No hard-cost contingency in the template, so the budget is silently light until someone adds a line. Quoting the gross levered IRR at the top of the Underwriting tab as the LP return, when the LP’s net return sits in the partnership section further down. Typing the total GP split into the promote column, which double-counts, since a 20% promote with 10% co-invest already produces a 28% GP share. Setting the developer fee toggle to No, which drops the fee from GP returns and understates GP economics. Leaving the equity input stale after an edit, so actual LTC no longer matches the target. A detail-method budget row whose monthly cells do not sum to its total, which breaks sources and uses. A detailed lease-up row that does not reach exactly 100%, which silently pushes stabilization to the end of the grid. Cancelling the Solve for Mill Rate prompt, which sets property tax to zero. A garden product delivered building by building but modeled with a single delivery date, which overstates carry and delays income. And a construction loan with no maturity, which runs to the sale month when there is no permanent takeout.
Framing Outputs
The Skill always names which spread it is quoting: untrended, trended, or sale. It also names which return: the gross return, or the partnership net return after the waterfall. Every number it reports comes with the cell it was read from, for your own checking. For a developer, the headline is the untrended development spread alongside yield on cost. For an LP, the headline is the net return and equity multiple after the promote. For a lender, it is actual LTC against target and whether the permanent loan covers the construction payoff. The model does not calculate debt yield or LP-net sensitivities, so when you ask for them the Skill computes or reruns them in the conversation and says so.
Operating Contexts (Chat / Cowork and Claude in Excel)
The Skill works in two environments. You can upload the Excel file to a Claude conversation and have Claude populate the model via code execution. The Skill bundle includes a clean copy of the model, so no upload is needed to get started, and the macros are preserved in the file Claude hands back. Or, if you’re using Claude in Excel, operate the model live with Claude reading and writing to the workbook directly. In both cases, you run the macros in desktop Excel. And as noted earlier, the Skill is also portable to other AI assistants, though the integration may be lighter.
A Note on the Underlying Model
The Apartment Development Model is a monthly, ground-up multifamily development tool with optional retail income. All primary inputs live on the Underwriting tab, with the Retail Income and Detail Expenses tabs becoming input tabs when their toggles are on. It produces development spread, yield on cost, and unlevered and levered returns, and runs them through a GP/LP or co-GP promote waterfall. It does not model phased delivery, preferred equity, affordable or LIHTC deals, or after-tax returns, and it assumes a single delivery date and lease-up pace. For a for-sale condo, a mixed-use project, or an existing building, other A.CRE models are the better fit. The model is macro-enabled (.xlsm). See the model post for the full walkthrough and tutorials.
Note: This AI Skill is built for v3.0 of the model. If your file is on an earlier version, such as v2.98 (which added LTV-based permanent debt sizing) or v2.99 (which made retail expense growth a direct input), cell positions may differ, so check the Version tab and confirm key cells before running the Skill. The model requires desktop Excel (2013, 2016, or 365) with macros enabled. Excel Online, Google Sheets, and LibreOffice cannot run the Recalculate macro, and Mac is not officially supported. Windows blocks macros in downloaded files, so right-click the file, choose Properties, and tick Unblock before opening it.
Before You Use This AI Skill with the Apartment Development Model
A couple of notes worth surfacing before you download.
Who this Skill is for. This Skill is built for real estate professionals with a strong grasp of financial modeling, and ideally prior exposure to multifamily development underwriting, development budgeting and lease-up analysis, construction and permanent financing, and partnership waterfalls. It’s best suited to graduates of our A.CRE Accelerator real estate financial modeling program, or analysts comfortable building development models from scratch. AI assistants make mistakes, and the Skill assumes an analyst on the other side who can catch them. Treat its output the way you’d treat work from a sharp junior analyst: useful, fast, and always verified before it informs an investment decision.
License. The Skill is distributed under the A.CRE software license, with full terms in the LICENSE.txt file included in the bundle. The short version: use it for personal, organizational, and client-facing analysis, and don’t resell or redistribute it. Use by an AI assistant operating on your behalf is expressly permitted. That’s the whole point.
Download the Apartment Development Model + AI Skill
To make this model accessible to everyone, it is offered on a “Pay What You’re Able” basis with no minimum (enter $0 if you’d like) or maximum (your support helps keep the content coming – typical real estate development models sell for $200 – $500+ per license). Just enter a price together with an email address to send the download link to, and then click ‘Continue’. If you have any questions about our “Pay What You’re Able” program or why we offer our models on this basis, please reach out to either Mike or Spencer.
Your download includes three files: the Excel model, the AI Skill (.skill file), and a short README explaining how to use them together. The Skill bundle includes a clean copy of the model, so no upload is needed to get started with Claude.
We regularly update both the model and the AI Skill (see version notes below). Paid contributors receive a new download link via email each time either is updated.
Frequently Asked Questions about AI Skill for the A.CRE Apartment Development Model
Version Notes: AI Skill
Version 3.0
- Initial release of the AI Skill for the A.CRE Apartment Development Model
- Paired with v3.0 of the Excel model
- Supports both Chat / Cowork (Skill bundle includes a clean copy of the model, no upload needed) and Claude in Excel (operate the live workbook directly)
- Includes four-role identification (developer/sponsor go/no-go, GP/LP or co-GP structuring, LP/equity investor, construction and permanent lender), input population from site, budget, comp, and term sheet documents, toggle-state reporting, a dedicated recalculation hand-off for the macro-driven equity solve, and mistake-catching across blue formula defaults, missing contingency, gross versus net returns, promote entry, stale equity, lease-up and budget tie-outs, and loan maturity
- Portable to other capable AI assistants (ChatGPT, Gemini, etc.) via the SKILL.md file
