Automate financial models with an AI coding agent.

Manual underwriting in Excel eats a day. An AI coding agent builds the same model in minutes, surfaces assumptions you'd have missed, and re-runs itself when new transaction data lands.

What this buys you.

A financial model that runs itself. Drop the market data in, describe the deal, and the agent returns a structured model with cash flow, leverage, yield, and a multi-year IRR. Re-run the same prompt against new transactions and the numbers update without rebuilding the spreadsheet.

Context. Brainstorm. Build.

01

Gather the eight inputs

Recent sales and rental transactions, the counterparty's track record, floorplans, indicative pricing, location context, the brochure, and the payment or closing terms.

02

Ask the agent to plan first

Before asking for the model, ask how it would build one for this asset class and market. Read the plan and refine before building.

03

Say 'implement plan'

Two words. The agent builds the actual spreadsheet, formulas, and sensitivity charts against the uploaded context.

04

Link a refresh

Drop new transaction data into the same task, or connect a live data source, so the model updates without rebuilding from scratch.

The brainstorm prompt

How would you best build a financial model for this residential investment? What should be included and how can we forecast returns? Market data and project information attached.

What it still gets wrong.

i.

It's a coding agent, not a property tool

The output is only as accurate as the inputs. Get the transaction comps right before running anything.

ii.

Read the assumption layer

If a rent-free period or vacancy buffer looks off for your market, push back and ask for a revision.

iii.

The math is solid; the assumptions aren't guaranteed

Treat the first model as a draft to interrogate, not a final answer.

Next Playbook