6-lesson course · about 2 hours
Variance commentary where every cause traces back to a number
For CFOs, controllers, FP&A teams, accountants and finance operations professionals
Everything for Finance teamsUse AI to draft month-end commentary that cites a source figure for each claim, labels unproven causes as hypotheses with a named check, and passes a short evidence check before the CFO sees it.
What you’ll be able to do
By the end, you can…
How it works
A tutor that meets you where you are
- Lessons that meet you at your levelSix self-paced lessons. The later ones are written for your level as you reach them, so you aren't re-reading what you already know.
- An AI tutor you can ask anythingType or speak your questions while you work, such as how to word a hypothesis label or what counts as a precise source location.
- Quizzes and flashcards for each lessonShort quizzes and flashcards tied to each lesson help you check your understanding and remember the steps.
- Progress tracking and a certificateSee where you are in the course and get a certificate of completion when you finish.
Curriculum
The curriculum
The course has two modules. The first covers writing commentary that holds up on evidence alone, and the second covers drafting it with an AI tool and checking it before it reaches management.
Module 1Evidence Before Words3 lessons
1Anatomy of a Traced Variance Claim
- Spot the difference between a variance figure and a stated cause in an existing comment
- Write a variance comment in which each cause cites a specific source figure
- State the reporting period, currency, units and calculation method beside each cited figure
- Record the source location of each figure precisely enough that a colleague could find it without asking you
2Proven Causes Versus Hypotheses
- Choose whether a stated cause is proven by the traced figures or is a hypothesis
- Write a hypothesis label in a consistent format that management can recognise
- Name a specific data check that would confirm or reject each hypothesis
- Rewrite a comment that asserts a possible cause as fact so it reads as a hypothesis
3Flagging Stale, Provisional and Unreconciled Inputs
- Identify stale, provisional and unreconciled inputs in a set of month-end figures
- Write a short limitation note stating how each weak input affects a specific comment
- Decide whether a comment built on a weak input should be qualified, labelled as a hypothesis or held back
Module 2Drafting With AI and Checking Before It Ships3 lessons
4Preparing a Safe Figures Pack for an AI Tool
- Build a figures table where each row has an ID, value, period, currency, units, method, source and input status
- Remove or replace customer identifiers before sharing figures with an AI tool
- Check whether the chosen AI tool is approved for the data before using it
5Prompting an AI Tool to Draft From Supplied Figures Only
- Write a prompt that tells the AI tool to use only the supplied figures and cite a figure ID for each claim
- Instruct the tool to label unproven causes as hypotheses with the supplied check
- Review an AI draft line by line and reject any statement that cannot be linked to a supplied figure
6The Evidence Check and Late Corrections
- Run a short evidence check on a draft commentary and record the result for each claim
- Remove or relabel every claim that lacks a traced number
- Update commentary after a late correction so every affected sentence matches the revised figure
Who it’s for
Made for you if…
Made for: CFOs, controllers, FP&A teams, accountants and finance operations professionals
FP&A analysts who write the monthly commentary
You calculate the variances but often write causes like 'timing' without a clear source. This course gives you a way to cite a figure for each claim, or to say plainly that a cause is a hypothesis.
Controllers who review commentary before it goes to the CFO
You need to know which lines are backed by numbers and which aren't. A repeatable evidence check and clear hypothesis labels make that visible before sign-off.
Finance staff starting to use AI for drafting
You want AI to speed up the writing without inventing explanations. You'll learn to give it a safe figures pack and keep its draft tied to the figures you supplied.
Questions
Good to know
What do I need before starting?
Experience with monthly variance analysis against budget, forecast or prior period, working use of spreadsheet lookups, pivots or system exports, one month of actuals with a comparison basis to practise on, and your employer's permission to use your chosen AI tool with that data.
How long does it take?
About two hours across six self-paced lessons.
Does it cover reconciliations or fixing the data?
No. You'll learn to flag unreconciled, stale or provisional inputs and state how they limit your commentary, but reconciliation methods, correcting entries and data cleaning aren't covered.
Will I build a full management pack or slides?
No. The course is about the variance commentary itself. Pack assembly, slide design, scenario modelling and accounting policy interpretation are outside its scope.
How does the course handle confidential data?
You'll learn to remove or replace customer identifiers before sharing figures with an AI tool, and to check whether the tool is approved for that data before you use it.
What happens when a figure changes after I've written the commentary?
The last lesson covers late corrections, so that every affected sentence is updated to match the revised figure.
Do I need a particular AI tool?
No specific tool is required. The course teaches a drafting and review method, not how to choose a tool. Use one your employer has approved for the data.