Real finance work, done with AI you can defend
A durable method for doing real finance and accounting work with AI, from close and reporting to memos, contracts, and valuation.
Free and self-paced, open to anyone. Bring your own AI; no account, sign-in, or API key needed, and your progress saves in this browser.
- Modules live
- 6
- Practice questions
- 97
- Key terms
- 47
- Est. time
- 12hrs
Download a real data folder
Each module gives you a scenario and a folder of realistic finance data to work from.
Run the pattern in your own AI
Use any tool you already have. The skill is the workflow and the validation, not one product's buttons.
Validate what you get back
A guided checklist and rubric prove the output ties to the numbers before it ships.
Practice tools
Curriculum
The Ten-Module Program
Module 0 teaches the method. Modules 1 through 9 drill it on a real workflow each, and the capstone chains them into one board-ready package. Modules open as they are released.
MODULE 0
The Finance AI Operating System
The method the whole course runs on. Why the same tool produces defensible work for one person and confident nonsense for another; the five moves that make the difference (map the task as a journey with a worked example of the destination, split deterministic work from non-deterministic, fuel the model with a minimal context folder, guard against data and governance risk, and review before anything ships); the deterministic split that keeps AI from imitating arithmetic; matching the model tier to the task; why context is the fuel and less is often more; and the human review checklist that no output skips. You leave with a personal red-lines card you can take to your own desk.
MODULE 1
Close and Reporting Acceleration
The first full workflow: turning a trial balance and a prior-period comparison into close-ready flux commentary that ties to the numbers. What close-ready means and why materiality and tie-out discipline come first; the flux pattern that has the template compute the variances (deterministic) and the model narrate only the drivers (non-deterministic); the minimal folder that briefs the model; the red-lines check that opens the lab; and a guided validation pass that catches the three ways AI flux commentary goes wrong (inventing plausible drivers, narrating immaterial noise, and rounding inconsistently). You produce commentary that reconciles to a real trial balance to the dollar.
MODULE 2
Variance Analysis and Driver Trees
Moving from what changed to why it changed and what to do about it. Building a driver tree (price, volume, mix, and cost), producing a primary explanation plus alternative hypotheses, and adding the decision-quality layer that turns analysis into a recommendation. Includes when the model is likely to be confidently wrong just past the edge of its competence, and how a second hypothesis guards against it.
MODULE 3
Industry Research and Benchmarking
Building a cited peer-set benchmark from public filings while keeping facts and inference strictly separated. Defining a peer set and metric definitions, extracting figures from filings, tagging every claim as fact (cited) or inference (reasoned), and validating that every source resolves. The discipline that keeps research memos from quietly fabricating a benchmark.
MODULE 4
Working Capital and 13-Week Cash
The deterministic-split showcase. The model builds the 13-week cash model (the machinery) from AR and AP agings and actuals, you validate the machinery once, and then the model narrates scenarios on top of it. Levers, sensitivity checks, and the actions a forecast should drive.
MODULE 5
Cost Takeout and Spend Analytics
Turning a vendor spend cube into a defensible opportunity-sizing analysis with an explicit assumptions ledger and a governance-safe procurement narrative. Sizing savings without fabricating benchmarks or defaming vendors, and documenting every assumption behind a number.
MODULE 6
Data Review and Anomaly Detection
Exception-based testing on a messy general-ledger extract, with an audit-trail artifact to back it. Because the anomalies are seeded, validation is fully objective: the lab is scored on what you found, what you missed, and what you flagged that was clean. Duplicate payments, weekend postings, round-dollar entries, and sign reversals.
MODULE 7
Technical Accounting Memo Workflow
Drafting a technical accounting memo from a fact pattern and guidance excerpts, then peer-reviewing a flawed AI memo the course provides. Issue spotting, guidance mapping, and the review that catches a wrong citation, an overreached conclusion, or a missing alternative treatment. Includes preparing for audit-committee questions.
MODULE 8
Contract Review to Accounting Implications
Extracting clauses from a contract set and mapping them to their accounting implications, with confidence ratings and an open-questions list. Revenue recognition under ASC 606, variable consideration, and bundled performance obligations, with every extracted clause quoting the contract language it came from.
MODULE 9
Valuation Model and Scenario Engine
The split-it thesis at full scale. The model builds a deterministic DCF skeleton, you validate the math against an assumptions ledger, and then the model writes the sensitivity narrative and the board pre-read framing. Assumptions traceability, sensitivities that bracket the base case, and narrative numbers that tie.
CAPSTONE
Capstone: The Close-to-Board Simulation
One integrated deliverable that chains the close commentary, variance explanation, exception review, and a technical memo into a single board package from one company dataset. Scored against a composite rubric covering all eight learning objectives. No new concepts; the assessment is the integration.
How this course works
1. Read one section at a time
Short, focused steps with worked examples and interactive calculators. Move with Continue or the arrow keys.
2. Check yourself as you go
Sections end with a knowledge check that explains every answer, right or wrong. Answers save as you go.
3. Finish with practice
Each chapter wraps up with a practice exam, and the course-wide Final Practice pools every question by topic.
Free and self-paced, open to anyone. Your progress saves automatically in this browser, so no account or sign-in is needed.
Go deeper
Sources & further reading
Everything behind the material: the research the chapters cite, the books worth owning, and the shows worth a commute. Open a shelf to browse.
Your instructor
Devon Coombs, CPA, MBA
Professor at Santa Clara University's Leavey School of Business. This is the same course he teaches on campus, adapted for self-paced online study.
“He provides every resource in the world to help his students, whether that be his own book, the videos he creates with great edits, or his own website, which helps you really drill down the information.”