29-lesson course · about 9 hours

Put Meta's Muse to work on your close, and still tie out every figure

For CFOs, controllers, FP&A teams, accountants and finance operations professionals

Everything for Finance teams

Set up Muse with the least access that works. Then run a variance review, a close tracker, a payout reconciliation and a commentary draft, and check each one before you sign.

What you’ll be able to do

By the end, you can…

Decide whether Muse fits your teamCheck the facts on Meta's official pages: where Muse is offered (US, Canada and Japan as of 2026-10-28) and the plans (Free with an unpublished usage limit, Power at $20/month, Maximum at $100/month). Also check that there is no team workspace, what the Terms say about financial advice and what your employer's AI policy allows. Then write a one-page fit decision.
Set up and connect Muse with the least access that worksUse a separate account and turn off training under Data controls. Set connectors and web access to Always ask, answer approval cards with Allow once or Deny, keep the wallet empty and share no Artifacts. Connect QuickBooks, then Stripe or Shopify, one at a time and retrieval-only where that option exists. Connect only after IT approves, and know why unpublished financials, payroll, customer PII and credentials stay out.
Run a variance review you can defendWrite the brief first: period, comparison basis, currency, units and thresholds. Run a read-only variance prompt that keeps measured variances apart from hypotheses marked [CHECK]. Trace the top flags to transactions, then recompute the material totals and percentages yourself in your workbook.
Keep the close calendar on trackBuild a close tracker Artifact with owners and BD1–BD5 due dates, and schedule business-day reminders. Then confirm under Upcoming that each one says what you meant, and fix any that drifted.
Reconcile payouts and draft commentary without letting Muse book anythingCompare Stripe or Shopify payouts with QuickBooks deposits and sort them into matched, unmatched and timing differences. Queue each unmatched item for your own disposition, and Muse creates no journal entries. Then draft commentary and a scenario labelled as a draft, using approved figures only, with causes kept as hypotheses until you have evidence.
Review what Muse did every weekGo through the Activity log, Upcoming, per-task approvals, Allowed websites, connector permissions and Memory. Correct or revoke anything that has drifted, and know when to disconnect, export your data or reset.

How it works

A tutor that meets you where you are

  1. Lessons that adapt to your levelEach lesson is a hands-on tutorial you follow with Muse open, practising on a sandbox company or anonymized exports. The lessons adjust to what you already know. A controller who has run a hundred closes and an analyst in their first year both start from where they are.
  2. An AI tutor you can ask anythingType your question or ask it out loud. You might ask why a setting matters, how to word a prompt for your own chart of accounts, or what to do when your screen doesn't match the lesson. Muse's screens change often, so expect some differences.
  3. Quizzes and flashcards for each lessonShort quizzes and flashcards tied to each lesson help the important points stick: the safe-start settings, the approval options and what never to connect.
  4. Progress tracking and a certificatePick up where you left off across the 29 lessons, see what's left in each module, and get a certificate of completion when you finish.

Curriculum

The curriculum

The course has 29 lessons in eight modules. You start by deciding whether Muse fits your work, then set it up and connect it safely. Next come four finance workflows, each a full tutorial, and the course ends with a weekly routine for reviewing what Muse did.

Module 1Is Muse right for your finance work?4 lessons

  1. 1What Muse is and what it is not
    • Distinguish Muse from Meta AI, Muse Spark and unrelated products named Muse
    • Describe how an agent that acts differs from a chat assistant in a finance setting
    • List the Muse features a finance team would use: chats, Artifacts, Connectors, Goals, scheduled tasks
  2. 2Availability, devices and plans as of October 2026
    • Check whether you are eligible for Muse in your country and age
    • Compare the Free, Power and Maximum plans for a monthly finance workload
    • Locate the official Meta pages to confirm current availability and prices
  3. 3What the Terms allow and what has gone wrong
    • Identify which finance uses the Terms prohibit or disclaim
    • Explain the responsibility you keep for every action and Artifact Muse produces
    • Connect each public incident to the setting that guards against it
  4. 4Decide whether Muse fits your team
    • Write a one-page fit decision for using Muse on finance standing tasks
    • List the questions to take to IT or compliance before connecting any work account
    • Choose a practice dataset that keeps live nonpublic results out of Muse

Module 2Set up Muse safely5 lessons

  1. 5Create a separate account for Muse
    • Create a Muse account with a separate email address
    • Explain what linking through Accounts Center would share
    • Name your agent during onboarding
  2. 6Turn off training and set permissions to Always ask
    • Turn off Help improve our AI models under Data controls
    • Set connector and web-access permissions to Always ask
    • Confirm the wallet holds no card and that no Artifact is shared
    • Verify all settings after saving
  3. 7Answer approval cards with Allow once or Deny
    • Read an approval card using See task details before deciding
    • Choose between Allow once, Allow for this task, Allow for this site, Always allow and Deny
    • Spot an approval request that goes beyond the task you set
  4. 8Give Muse your finance context through Memory and side chats
    • Write a context message covering role, close calendar and conventions
    • Open Identity > Memory and edit or delete entries
    • Use a side chat to keep one workstream's context separate
  5. 9Run a first read-only task and check the Activity log
    • Write a read-only first task with explicit stop points
    • Review the Activity log and System Files after the task
    • Compare what Muse did with what you intended

Module 3Connect finance services with the least access4 lessons

  1. 10What never to connect or upload
    • Classify proposed inputs as allowed, sanitize first, or never
    • Explain to a colleague why custom connectors and Full Disk Access to sensitive systems stay off
    • Write a short data rule for your team's Muse use
  2. 11Connect QuickBooks read-only to a sandbox company
    • Connect QuickBooks with the narrowest access offered
    • Ask Muse to state the connector's capabilities and what it will ask before doing
    • Run a read-only test query and verify it against QuickBooks
  3. 12Connect Stripe or Shopify read-only
    • Connect Stripe or Shopify as a second connector, only after the first is verified
    • Confirm the connector cannot create refunds, payouts or customer changes
    • Tie one payout amount and date to the processor dashboard
  4. 13Decide on Gmail and Mac access
    • Decide whether a Gmail connection is justified and approved
    • Configure Gmail for reading only, if you connect it
    • Set Mac local-app access to Off or Read only

Module 4Workflow: monthly variance review4 lessons

  1. 14Write the review brief before the prompt
    • Write a variance brief stating period, comparison basis, currency, units, source and thresholds
    • Note your own expectation of what moved before prompting
    • Save the brief as reusable text in a side chat
  2. 15Run the variance table with flags and [CHECK] marks
    • Run the read-only variance prompt against your sandbox data
    • Separate measured variances from hypotheses in the output
    • Spot period, currency or unit errors before reading any narrative
  3. 16Trace the biggest flags to transactions
    • Request transaction listings behind the top flags
    • Test each hypothesis against the listed transactions
    • Identify duplicates, misclassifications and cut-off issues for your own follow-up
  4. 17Recompute and review it like a first-year's work
    • Recompute totals, $ changes and % changes in Excel from a QuickBooks export
    • Record ticks, differences and open [CHECK] items on a review sheet
    • Decide what is fit to pass on and what returns for more evidence

Module 5Workflow: close tracker and reminders3 lessons

  1. 18Build a close tracker Artifact
    • Prompt Muse to build a close tracker Artifact from your checklist
    • Check the tracker against your source checklist line by line
    • Keep the Artifact unshared
  2. 19Schedule business-day reminders
    • Write a reminder prompt with day, time, time zone and business-day rules
    • Set approval for changes to scheduled tasks
    • Explain why business-day logic needs checking
  3. 20Confirm reminders under Upcoming and fix drift
    • Check every scheduled reminder under Assistant icon > Upcoming
    • Correct wrong dates, times or contents by chat
    • Delete reminders that are no longer needed

Module 6Workflow: payout-to-deposit reconciliation summary4 lessons

  1. 21Set the reconciliation scope and tolerances
    • Write a reconciliation scope with period, accounts, matching keys and tolerances
    • Anticipate common payout differences such as fees, refunds and in-transit timing
    • Record the control totals you will tie to
  2. 22Compare payouts with QuickBooks deposits
    • Run the payout-to-deposit comparison prompt with your scope
    • Check that every payout and deposit appears exactly once
    • Refuse any step that would create or edit entries
  3. 23Queue unmatched items for your disposition
    • Request evidence needs for the largest unmatched items
    • Record a disposition and owner for each exception
    • Distinguish a timing difference from an error that needs an adjustment
  4. 24Tie out the reconciliation summary
    • Foot and cross-foot the reconciliation in your workbook
    • Tie category totals to processor and QuickBooks reports
    • Sign off or return the summary with open items noted

Module 7Workflow: commentary from approved figures, and AP triage3 lessons

  1. 25Draft management commentary from approved figures only
    • Prepare an approved-figures input with period, currency and units
    • Run the commentary prompt and check every number against the input
    • Rewrite or remove any cause stated without evidence
  2. 26Build a labelled draft scenario Artifact
    • Prompt a scenario Artifact using only stated assumptions
    • Test its formulas against a hand calculation
    • Keep it labelled and unshared
  3. 27Optional: triage the AP inbox without replying or paying
    • Run an invoice-collection prompt on a read-only inbox
    • Check the table against the source emails
    • Keep payment and approval in your AP system

Module 8Review what Muse did each week2 lessons

  1. 28Run the weekly review
    • Walk through each review screen in order
    • Compare logged actions with your intended tasks
    • Record findings in a short weekly log
  2. 29Correct, revoke, export or reset
    • Revoke a website or connector and confirm it is gone
    • Download your Muse data before any reset
    • Choose the right response to a given finding

Who it’s for

Made for you if…

Made for: CFOs, controllers, FP&A teams, accountants and finance operations professionals

Controllers

You run the close and answer for every figure that leaves finance. You want an agent that tracks the checklist and flags variances, and you want it to have no way to post, edit or recategorize anything in the ledger.

FP&A analysts

You explain what changed each month and write the commentary. You want first drafts built only from approved figures, with every cause marked as a hypothesis until you have evidence.

Accountants and finance operations staff

You reconcile Stripe or Shopify payouts to QuickBooks deposits and work the AP inbox. You want the matching and gathering done for you, while dispositions, entries and payments stay with you.

CFOs

You have to decide whether a cloud AI agent belongs anywhere near your team's work. You want a clear view of its plans, limits, Terms and public incidents, plus the settings that keep nonpublic results and customer data out.

Questions

Good to know

Is Muse available where I live?

As of 2026-10-28, Muse is offered in the US, Canada and Japan, and you must be 18 or the age of majority where you live. Availability changes, so the course shows you where to confirm it on Meta's official pages. It does not cover ways around region or age limits.

Do I need a paid Muse plan?

Muse's own plans are separate from this course. As of 2026-10-28, Meta lists Free (with a usage limit it doesn't publish), Power at $20/month (500M Muse tokens per week) and Maximum at $100/month (3B Muse tokens per week). Meta doesn't publish how many tokens a task uses, so check the official page and watch your usage.

Can my whole finance team share one Muse?

No official Meta source describes a team workspace, shared agent or team plan. Each person needs their own Muse account and subscription. Lesson 4 covers what that means for your fit decision.

Will Muse post journal entries, pay bills or move money?

Not in this course. Connectors are set to retrieval-only where that option exists, every prompt says read only, no card goes in the wallet, and you post any adjustments yourself in your own systems under your own approvals.

Should I practise on live company data?

No. You practise on a sandbox company such as a QuickBooks sandbox, or on anonymized exports. Work accounts get connected only with your employer's or IT team's approval. Unpublished financials, payroll, customer PII and credentials stay out of Muse.

Will Muse give me tax, investment or audit advice?

Muse's Terms say it does not give legal, financial, investment or tax advice. The course doesn't teach any of these. It treats every output as prepared work that you check and sign off.

What do I need before I start?

You need working knowledge of a month-end close, variance analysis and bank or payout reconciliation, and enough spreadsheet skill to recompute totals, $ and % changes and reconciling differences. You also need eligibility for Muse, your employer's position on cloud AI agents, and a sample or sanitized dataset to practise on.

Does this course cover Factories, Dots or Claude Duet?

No. Multi-agent reporting Factories and other proactive agents such as Dots and Claude Duet are covered elsewhere in the Finance track. This course covers Muse alone.

Meta Muse for Finance Teams: safe setup and four standing tasks

$97

One payment · lifetime access

  • 29 hands-on lessons in 8 modules, each a step-by-step tutorial with copy-paste prompts written for finance
  • Safe-start settings, approval-card practice and a written data rule for your team's Muse use
  • Four workflows: variance review, close tracker with reminders, payout-to-deposit reconciliation, and commentary from approved figures
  • An optional AP inbox triage that collects invoices without replying, forwarding or paying
  • A weekly review routine covering the Activity log, Upcoming, Allowed websites, connectors and Memory
  • An AI tutor you can type to or talk to, quizzes and flashcards for each lesson, progress tracking and a certificate of completion

Secure checkout by Stripe. Have a promotion code? Enter it at checkout.

Refunds: Full refund if you ask before you open your first lesson. Once you’ve started, purchases aren’t refundable, except where the law gives you that right.

Already bought it? Sign in