6-lesson course · about 2 hours
Explain reconciliation differences with evidence, or document them openly. No plugs.
For CFOs, controllers, FP&A teams, accountants and finance operations professionals
Everything for Finance teamsA focused 2-hour course for accountants and controllers: break a difference into its parts, use AI to propose causes from your actual records, test them, and leave a trail an auditor can follow.
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 levelThe course has six self-paced lessons. Later lessons are written for your level as you reach them, so you spend time on what you need rather than what you already know.
- An AI tutor you can ask anythingAsk questions about your own reconciliation situations, typed or out loud, and get explanations tied to the lesson you are working on.
- Quizzes and flashcards for each lessonShort quizzes and flashcards tied to each lesson help you check your understanding of steps like categorising components or spotting unreliable AI output.
- Progress tracking and a certificateSee where you are across both modules, and receive a certificate of completion when you finish.
Curriculum
The curriculum
Two modules, six lessons. The first breaks the difference down and prepares your records safely; the second tests the AI's candidate causes and turns the results into documentation a reviewer can follow.
Module 1Break the Difference Down Before You Explain It3 lessons
1Why Plugs and Unexplained Entries Hide the Problem
- Explain how a plug to 'other' makes a balance tie while leaving the underlying cause in the books
- Spot a correcting entry that a reviewer could not trace back to source records
- State the two acceptable results for any reconciliation difference
2Splitting a Difference into Component Items
- Compare two exported source records line by line to isolate unmatched and mismatched rows
- Assign each component item a provisional category: timing, duplicate, rejected or unloaded, currency or unit mismatch, or classification
- Show that the component items sum back to the total difference
- Record period, currency and units next to every component item
3Preparing Reconciliation Records for AI Safely
- Identify customer identifiers, account credentials and nonpublic reporting information in a reconciliation export
- Replace sensitive fields with neutral row keys while keeping a private mapping back to the source
- Choose which columns an AI tool needs to propose causes and which it does not
- Confirm that the tool in use is one your organization has approved for finance data
Module 2Test the Causes and Document the Result Honestly3 lessons
4Asking AI for Candidate Causes Tied to Specific Rows
- Write a prompt that gives the AI the prepared records, the difference amount, period, currency and units
- Ask for candidate causes that cite specific row keys and state what evidence would confirm or rule them out
- Spot AI output that asserts a cause as fact, proposes a balancing entry, refers to rows that are not in the data, or states amounts that do not recalculate
5Testing Each Candidate Cause Against the Source Records
- Trace each candidate cause back to the original source rows using the private key mapping
- Mark each candidate cause as supported, contradicted or unverified
- Record the specific evidence checked for each candidate cause
6Classifying Differences and Writing Open-Item Records
- Classify each component as resolved with a traceable explanation or as an open item, and state the evidence behind the classification
- Write an open-items log entry with amount, age, what was checked, what remains unknown and the next owner
- Write a reconciliation explanation a reviewer or auditor can follow without asking the preparer
- Use AI to draft or check documentation without letting it add facts the testing did not establish
Who it’s for
Made for you if…
Made for: CFOs, controllers, FP&A teams, accountants and finance operations professionals
Accountants who prepare bank, subledger-to-ledger or intercompany reconciliations
You want a repeatable way to investigate a difference that won't tie, and an honest way to record what you couldn't resolve by close.
Controllers who review reconciliations
You want explanations and open-item records you can follow without chasing the preparer, and fewer unexplained entries surfacing at audit.
Finance staff new to using AI for analysis
You want to use an approved AI tool on transaction-level data without exposing customer or nonpublic reporting information, and to catch AI output that guesses.
Questions
Good to know
What do I need to know before starting?
You should have done at least one account reconciliation, be able to export transaction-level data from a ledger or source system to a spreadsheet, understand basic journal entries and how period, currency and units affect a balance, and know which AI tools your organization has approved for finance data.
How long does it take?
About 2 hours across six self-paced lessons.
Will the AI tell me what entry to post?
No. The course teaches you to ask for candidate causes tied to specific rows, and to spot output that asserts a cause as fact, proposes a balancing entry, cites rows that aren't in your data or states amounts that don't recalculate.
How does the course handle sensitive data?
Lesson 3 covers identifying customer identifiers, account credentials and nonpublic reporting information in an export, replacing them with neutral row keys while keeping a private mapping, sending only the columns the AI needs, and confirming the tool is approved by your organization.
Does it cover accounting standards, materiality or write-off policy?
No. Interpreting standards, materiality thresholds and write-off policy is outside this course, as is fraud or misconduct investigation. The focus is investigating and documenting differences.
Will it help me pick or set up an AI tool?
No. Selecting, configuring or procuring AI tools isn't covered. You use a tool your organization has already approved.
Does it cover the full month-end close?
No. It doesn't cover redesigning the close calendar, AI agents that maintain a reporting cycle, scenario modelling or management commentary. It stays on reconciliation differences.