A free, login-gated course for FP&A analysts and finance operators. 14 lessons teaching you to build version-controlled financial models with an AI agent — starting with a field card and a free lesson, no paid tier required.
14 lessons across four phases — from diagnosing spreadsheet failure to delivering board-ready outputs with full lineage.
Course dependency map: 14 lessons across four phases — Diagnosis (1–2), Agent OS (3–5), Construction (6–10), and Operations (11–14).
14-module curriculum
DiagnosisLessons 1-2
01
The Fatal Excel Model
Why spreadsheet models fail and what it costs the business.
02
Models Are Systems
Treating financial models as software: inputs, logic, outputs, and dependencies.
Agent operating systemLessons 3-5
03
Building With the Agent
The four agent loops: build, explain, refactor, and review.
04
Architecture and Separation of Concerns
Four-layer model architecture: data, assumptions, logic, and reporting.
05
Reading the Patterns
Translating Excel formulas into recognizable Python patterns.
Model constructionLessons 6-10
06
The Driver Layer
Data vs drivers vs derived values -- and how to source defensible rationale.
07
Revenue Engines
Customer roll-forward, ARR, expansion, and net new revenue modeling.
08
Cost Models and Unit Economics
Headcount, COGS, sales capacity, and CAC payback.
09
Financial Statements as Outputs
P&L to cash flow to balance sheet -- the connected three-statement model.
10
Testing Financial Logic
Seven tests every financial model should pass before it goes to the board.
Team operations and reportingLessons 11-14
11
Version Control for Finance Teams
Commits, branches, pull requests, tags, and rollbacks for finance.
12
Code Review for Forecasts
Classifying, diffing, and approving forecast changes as a team.
13
Scenarios as Branches
Base, bull, bear, sensitivity sweep, and variance bridge.
14
Reproducibility, Lineage, and Reporting
Run IDs, board packs, metric lineage, and driver rationale.
The shift
From opaque cell references to named dependency graphs
The same model logic — expressed two ways. On the left: Excel cell references that require tracing every formula to understand a number. On the right: named nodes where every dependency is explicit and every output is traceable.
Left: Excel cell references require tracing every formula to understand a number. Right: named nodes make every dependency explicit and every output directly traceable.
What you’ll be able to do
Six things you can do when you finish
→Build a version-controlled financial model that any team member can run and reproduce from a single command.
→Direct an AI agent to write, refactor, and review model logic — without writing Python from scratch.
→Separate data, assumptions, logic, and reporting so every board-pack number traces back to a defensible source.
→Write a test suite that catches the seven most common financial modeling errors before the forecast goes to review.
→Run base, bull, and bear scenario branches with git — no duplicated workbooks or manual reconciliation.
→Present board-ready outputs with a run ID and full driver lineage, auditable on demand.
Who this course is for
Built for finance professionals, not software engineers
FP&A lead
Your team manages production models in Excel or Sheets. You want version control, auditability, and faster iteration cycles — without rebuilding everything from scratch.
CFO / Controller
You sign off on the numbers. You want every board-ready output to trace back to a defensible rationale, with a full audit history no spreadsheet can provide.
Strategic finance
You build revenue, cost, and scenario models to support decisions. You want reproducible runs, clean review workflows, and a board pack that explains itself.
Included
What you get with every lesson
14 written lessons
Self-paced written lessons covering the full arc from spreadsheet diagnosis to reproducible reporting.
Field cards
One field card per lesson: key concepts, the core diagram, and the prompts you need to apply the lesson in Bridge Town.
Prompt templates
The four essential prompts of financial model engineering and a cheat sheet of 20 reusable modeling patterns.
PR review checklists and sample pull requests for classifying and approving forecast changes.
Written and maintained by the Bridge Town team at Knightian Labs. Built for the practitioners who live in forecasts — not for software engineers.
Free with loginNo paid tier required14 lessons32 downloadable files
FAQ
Common questions
Is the course free?
Yes. The course is free and accessible immediately after login. All 14 lessons and every downloadable artifact are available with a free Bridge Town account — no paid tier required.
What do I need to know before starting?
Finance domain knowledge. You do not need to know Python. The agent writes the model code; your job is to direct it, review the outputs, and understand what the code does.
What will I be able to build when I finish?
A complete three-statement financial model as version-controlled code, with a test suite, scenario branches, and a reproducible board pack — all built with an AI agent.
What do I get to download?
Each lesson ships a field card (key concepts and prompts), plus cheat sheets, starter packs, and sample pull requests — 32 downloadable files in total across the 14 lessons.
How long does the course take?
Each lesson takes 18–28 minutes to read. The full course is roughly 14 hours of reading. Apply the concepts in Bridge Town between lessons and the total time depends on you.
Start lesson 01 — free
Login with your Bridge Town account to unlock all 14 lessons, field cards, starter packs, and review checklists. No paid tier, no time limit.