Modern Artificial Intelligence for Accounting Professionals
What the systems sold as "AI" actually are, and why that changes how you check them. This module separates rules, machine learning and generative AI, explains how language models produce an answer, opens up the retrieval, tools and agents inside accounting products, and maps tool categories to COSO's capability types and to what the evidence says about use.
-
Rules, Machine Learning and Generative AI
35分
-
How Large Language Models Produce an Answer
40分
-
Retrieval, Tools and Agents Behind the Product
40分
-
A Field Guide to Accounting AI Tools and How They Are Used
45分
What AI Can and Cannot Reliably Do With Financial Information
Fluent output can be wrong in ways an accountant must learn to predict. This module covers invented and stale authority, arithmetic, tie-out, rounding and sign errors on long ledgers, how to read accuracy studies for what they actually tested, and the twin traps of automation bias and algorithm aversion.
-
Confabulated Citations and Stale Authority
40分
-
Arithmetic, Tie-Outs and Long Ledgers
45分
-
Reading the Evidence on AI Accuracy
40分
-
Automation Bias, Algorithm Aversion and the Discipline of Doubt
40分
Client Confidentiality, Tax Return Information and Data Privacy
Before client data goes into any AI tool, you need to know what the data is, which rules protect it and whom they bind. This module covers tax return information under section 7216 and its regulations, the narrow contractor exception and consent, the FTC Safeguards Rule and the written security plan, the AICPA confidentiality rule, and a repeatable data decision.
-
Tax Return Information and Confidential Client Information
40分
-
Where AI Tools Send and Keep Client Data
35分
-
Permitted Disclosures, the Contractor Exception and Taxpayer Consent
45分
-
The Safeguards Rule, the Written Security Plan and the Data Decision
45分
Professional Responsibility and AI for Accountants
An AI tool changes how accounting work gets done, not who answers for it. This module covers the AICPA Code duties for members, the tax standard on relying on tools, Circular 230 due diligence, competence and firm procedures as the IRS Office of Professional Responsibility applies them to AI, supervision, and how to check your own state board rules.
-
The AICPA Code Duties That Do Not Move
35分
-
Reliance on Tools and Due Diligence in Tax Work
45分
-
Competence, Supervision and Firm AI Procedures
40分
-
State Boards, the Uniform Accountancy Act and Checking Your Own Rules
35分
AI-Assisted Tax and Accounting Research
AI can help an accountant find the way into a research question, but it is never the authority a conclusion rests on. This module builds a safe research workflow, explains what the GAAP hierarchy treats as authoritative, verifies every citation against the primary source, and records the research path so a reviewer can follow it.
-
A Safe Research Workflow From Question to Work Product
40分
-
Assistant, Not Authority: What Counts as Authoritative
35分
-
Verifying Every Citation Against the Primary Source
45分
-
Documenting Research Provenance in the Workpapers
35分
Prompt Engineering for Accounting Professionals
A good accounting prompt supplies the figures and the authority instead of asking the tool to recall them, pins down the entity and period, requires a tie-out, and ends with a reviewer. This module builds that structure, checks what may be pasted before prompting, shows how to iterate and when to stop, and gives patterns for everyday drafts.
-
The Anatomy of a Structured Accounting Prompt
35分
-
Supplying Figures and Authority Instead of Asking the Tool to Recall Them
40分
-
Checking What You Paste, Then Iterating With Purpose
35分
-
Prompt Patterns for Everyday Accounting Drafts
40分
Drafting Client Deliverables, Memos and Workpaper Documentation With AI
AI can produce a fast first draft, but every fact, figure and authority in it must trace to something a person checked. This module drafts letters and memos from a verified fact file, applies the written-advice standard to tax memos, builds workpapers that show who prepared and reviewed what, and keeps a proportionate record of AI use and approval.
-
Drafting Engagement Letters and Client Letters From a Verified Fact File
40分
-
Tax Memos Under the Written-Advice Standard
45分
-
Workpapers That Show Who Prepared, Who Reviewed and What Supports Each Conclusion
40分
-
Recording AI Use, Validation and the Approval Record
35分
AI in the Close, Reconciliation and Financial Reporting
AI can prepare reconciliations, accrual support, flux analysis and disclosure checklists, but it must stop before anything is posted, approved or certified. This module works through the numbers, sets out management's responsibility and the certification and internal control rules that apply to issuers, and designs controls over AI in the close.
-
AI-Prepared Bank and Account Reconciliations
40分
-
Accrual Support and Flux Analysis With AI
40分
-
Management's Responsibility, Officer Certifications and ICFR for Issuers
40分
-
Controls Over GenAI in the Close and Segregation of Duties
40分
AI in Audit and Assurance
What audit evidence standards ask of information a tool analyses, and why PCAOB standards for issuer audits and AICPA standards for non-issuer audits must never be mixed. The module covers AI support for risk assessment and document review, investigating items flagged in tests of details, and documenting what a tool did and who concluded.
-
Audit Evidence Under PCAOB and AICPA Standards When Technology Assists
40分
-
AI Support for Audit Risk Assessment and Document Review
40分
-
Investigating Items Flagged in Tests of Details
45分
-
Documenting What the Audit Tool Did and Who Reached the Conclusion
35分
AI for Firm Operations and Client Accounting Services
Where AI fits in the running of a firm: client intake and scheduling assistants, document collection and extraction, bookkeeping and client accounting services, and internal knowledge search. The module covers what the field evidence does and does not show, informing clients about providers, using return information inside the firm, and ranking use cases by risk.
-
Client Intake, Scheduling and Client-Facing Assistants
35分
-
Document Collection and Data Extraction
40分
-
Bookkeeping, Client Accounting Services and What the Field Evidence Shows
40分
-
Internal Knowledge Search, In-Firm Use of Return Information and Ranking Use Cases by Risk
45分
Building Repeatable Accounting AI Workflows
A chat that helps once is not a process. This module turns single AI tasks into staged workflows with classified inputs, a bounded AI step, independent checks, a named reviewer and a record. It shows how to stop a wrong output becoming a wrong action, how to design review against automation bias, and how to use the workflow canvas.
-
From Chat Session to Staged Accounting Workflow
40分
-
Designing So a Wrong Output Never Becomes a Wrong Action
40分
-
Designing the Review Step Against Automation Bias
35分
-
Worked Workflows, the Workflow Canvas, Approvals and Records
45分
AI Agents and Autonomous Systems in Finance and Accounting
An AI agent does more than answer: it calls tools, opens browsers, reads files and sends email. This module explains what agents are, how to rank agent uses from read-only to posting and paying, where human approvals and kill switches belong, why indirect prompt injection stays a residual risk, and how to keep segregation of duties.
-
What an AI Agent Does in Finance and Accounting Work
35分
-
Agent Risk Tiers From Read-Only to Posting and Paying
40分
-
Human Control Points, Event Logs and Kill Switches
40分
-
Indirect Prompt Injection, Excessive Agency and Segregation of Duties
45分
Cybersecurity, Data Governance and Vendor Evaluation
An AI vendor is a service provider that holds client data, so the firm has to test its claims, control who gets in, read its SOC reports correctly and oversee it by contract. This module covers vendor claims and changing terms, MFA and access, what SOC 1, SOC 2 and SOC 3 reports do and do not show, breach notification and exit.
-
Testing a Vendor's Security and Privacy Claims
40分
-
Identity, Multi-Factor Authentication, Access, Logging and API Keys
40分
-
Reading SOC 1, SOC 2 and SOC 3 Reports for AI Vendors
45分
-
Overseeing AI Providers: Contracts, Subprocessors, Breach Notification and Exit
45分
Firm AI Governance and Quality Management
A firm needs one written answer to which AI tools may be used, with what data, and who checks the result. This module builds that policy, places AI inside the system of quality management under SQMS 1 and QC 1000, uses voluntary frameworks without overstating them, and runs verification, disclosure, supervision, records, incidents and review.
-
The Firm AI Policy: Approved Tools, Prohibited Data and Acceptable Use
40分
-
AI Inside the System of Quality Management: SQMS 1 and QC 1000
45分
-
Voluntary Frameworks: The NIST AI RMF and ISO/IEC 42001
35分
-
Running the Policy: Verification, Disclosure, Supervision, Records, Incidents and Review
45分
Practical Accounting AI Lab
Four hands-on labs on fictional clients: find the planted errors in an AI-drafted bank reconciliation and client memo, expose a fabricated Code section and rebuild the weak prompt behind it, review an AI extraction that obeyed text hidden in a document, and grade an AI vendor before writing a firm AI use policy.
-
Lab: Find the Errors in an AI-Drafted Reconciliation and Memo
45分
-
Lab: Find the Fabricated Code Section and Fix the Weak Prompt
45分
-
Lab: Review an AI Document Extraction and Design a Secure Workflow
45分
-
Lab: Evaluate an AI Vendor and Build a Firm AI Use Policy
45分
Capstone: The Accounting AI Implementation Plan
Write a sixteen-section AI implementation plan for one real accounting task: choose a low-risk first use with a measured baseline, assign named owners and review points, file the vendor evidence and data decisions, then run a bounded pilot with metrics, return-on-investment tracking, stop signals and a review cycle.
-
Choosing a Low-Risk First Use and Scoping the Plan
45分
-
Controls, Owners and Review Points
45分
-
Vendor Evidence, Data Decisions and Records
40分
-
Pilot, Metrics, Failure Indicators and Improvement
45分