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 min
-
How Large Language Models Produce an Answer
40 min
-
Retrieval, Tools and Agents Behind the Product
40 min
-
A Field Guide to Accounting AI Tools and How They Are Used
45 min
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 min
-
Arithmetic, Tie-Outs and Long Ledgers
45 min
-
Reading the Evidence on AI Accuracy
40 min
-
Automation Bias, Algorithm Aversion and the Discipline of Doubt
40 min
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 min
-
Where AI Tools Send and Keep Client Data
35 min
-
Permitted Disclosures, the Contractor Exception and Taxpayer Consent
45 min
-
The Safeguards Rule, the Written Security Plan and the Data Decision
45 min
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 min
-
Reliance on Tools and Due Diligence in Tax Work
45 min
-
Competence, Supervision and Firm AI Procedures
40 min
-
State Boards, the Uniform Accountancy Act and Checking Your Own Rules
35 min
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 min
-
Assistant, Not Authority: What Counts as Authoritative
35 min
-
Verifying Every Citation Against the Primary Source
45 min
-
Documenting Research Provenance in the Workpapers
35 min
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 min
-
Supplying Figures and Authority Instead of Asking the Tool to Recall Them
40 min
-
Checking What You Paste, Then Iterating With Purpose
35 min
-
Prompt Patterns for Everyday Accounting Drafts
40 min
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 min
-
Tax Memos Under the Written-Advice Standard
45 min
-
Workpapers That Show Who Prepared, Who Reviewed and What Supports Each Conclusion
40 min
-
Recording AI Use, Validation and the Approval Record
35 min
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 min
-
Accrual Support and Flux Analysis With AI
40 min
-
Management's Responsibility, Officer Certifications and ICFR for Issuers
40 min
-
Controls Over GenAI in the Close and Segregation of Duties
40 min
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 min
-
AI Support for Audit Risk Assessment and Document Review
40 min
-
Investigating Items Flagged in Tests of Details
45 min
-
Documenting What the Audit Tool Did and Who Reached the Conclusion
35 min
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 min
-
Document Collection and Data Extraction
40 min
-
Bookkeeping, Client Accounting Services and What the Field Evidence Shows
40 min
-
Internal Knowledge Search, In-Firm Use of Return Information and Ranking Use Cases by Risk
45 min
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 min
-
Designing So a Wrong Output Never Becomes a Wrong Action
40 min
-
Designing the Review Step Against Automation Bias
35 min
-
Worked Workflows, the Workflow Canvas, Approvals and Records
45 min
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 min
-
Agent Risk Tiers From Read-Only to Posting and Paying
40 min
-
Human Control Points, Event Logs and Kill Switches
40 min
-
Indirect Prompt Injection, Excessive Agency and Segregation of Duties
45 min
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 min
-
Identity, Multi-Factor Authentication, Access, Logging and API Keys
40 min
-
Reading SOC 1, SOC 2 and SOC 3 Reports for AI Vendors
45 min
-
Overseeing AI Providers: Contracts, Subprocessors, Breach Notification and Exit
45 min
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 min
-
AI Inside the System of Quality Management: SQMS 1 and QC 1000
45 min
-
Voluntary Frameworks: The NIST AI RMF and ISO/IEC 42001
35 min
-
Running the Policy: Verification, Disclosure, Supervision, Records, Incidents and Review
45 min
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 min
-
Lab: Find the Fabricated Code Section and Fix the Weak Prompt
45 min
-
Lab: Review an AI Document Extraction and Design a Secure Workflow
45 min
-
Lab: Evaluate an AI Vendor and Build a Firm AI Use Policy
45 min
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 min
-
Controls, Owners and Review Points
45 min
-
Vendor Evidence, Data Decisions and Records
40 min
-
Pilot, Metrics, Failure Indicators and Improvement
45 min