AI for Finance and Accounting Training Course

Participants can assess whether finance data is suitable for controlled AI use.

• Profile general ledger extracts for completeness, uniqueness, validity and timeliness
• Reconcile chart-of-accounts mappings to reporting hierarchies before model development
• Apply data lineage records from transaction source to financial report
• Classify structured, unstructured and semi-structured finance data sources
• Detect missing values, duplicates and outliers using data-quality rules
• Define data dictionaries for vendor, customer and journal-entry fields

Participants can create controlled prompts for repeatable finance and accounting tasks.

• Construct role, task, context and format prompts for finance analysis
• Use few-shot examples to standardise account-reconciliation narrative outputs
• Set delimiters and source boundaries to reduce prompt ambiguity
• Request assumptions, calculations and citations in generated management commentary
• Compare prompt outputs against approved policies and source documents
• Maintain prompt libraries with version control and approval ownership

Participants can use generative AI for document review and reporting drafts with appropriate safeguards.

• Extract obligations, dates and payment terms from supplier contracts
• Summarise accounting-policy memoranda while preserving qualified conclusions and exceptions
• Draft variance explanations from controlled financial data and reviewer notes
• Create board-report outlines using approved metrics, definitions and evidence
• Identify hallucinated facts, unsupported citations and omitted qualifications
• Apply human review checkpoints before external or regulatory communication

Participants can create reproducible finance analytics workflows with controlled reporting outputs.

• Build controlled Excel models with Power Query transformation steps
• Create Power BI variance dashboards from governed ledger datasets
• Automate recurring refreshes while retaining reconciliation control totals
• Use Python notebooks for reproducible finance-data transformation workflows
• Trace formula logic, query steps and report calculations for review
• Separate exploratory analysis files from controlled reporting outputs

Participants can identify and design AI-supported improvements for record-to-report activities.

• Map record-to-report activities suitable for AI-assisted close procedures
• Prioritise close bottlenecks using volume, risk, repeatability and judgement criteria
• Generate reconciliation support schedules from bank, subledger and ledger records
• Match transactions using reference keys, amounts, dates and confidence thresholds
• Route unresolved exceptions through documented preparer and reviewer workflows
• Measure close-cycle performance using ageing, exception and rework indicators

Participants can evaluate unusual journal entries using rules, anomaly models and documented investigation criteria.

• Define normal journal-entry patterns using account, user, period and amount features
• Apply Benford's Law as a screening indicator, not proof
• Compare rules-based exceptions with unsupervised anomaly-detection output results
• Set investigation thresholds using materiality, false-positive and capacity considerations
• Document anomaly rationale, evidence links and reviewer disposition
• Validate detection performance through precision, recall and sample review

Participants can design controlled AI workflows for invoice processing and payment-risk detection.

• Extract invoice fields using optical character recognition and validation rules
• Match invoices to purchase orders, receipts and contractual tolerances
• Detect duplicate payments through supplier, amount, date and reference comparisons
• Classify spend categories using supervised learning labels and confidence scores
• Route invoices by exception type, approval authority and payment risk
• Monitor straight-through-processing rates, exceptions and manual-touch costs

Participants can apply AI methods to receivables prioritisation, collection activity and cash-receipt forecasting.

• Predict payment timing from customer history, invoice attributes and disputes
• Segment receivables using ageing, risk scores and collection priorities
• Generate collection correspondence within approved tone, policy and legal boundaries
• Forecast cash receipts using scenario assumptions and probability-weighted outcomes
• Compare model forecasts with actual collections and revise drivers
• Escalate credit-risk signals through documented account-management procedures

Participants can build and challenge AI-supported forecasts using operational drivers and uncertainty measures.

• Build driver-based forecasts from operational volumes, prices and cost assumptions
• Use regression outputs cautiously when relationships lack economic rationale
• Produce scenario models for base, downside and upside planning cases
• Generate management commentary from variances, drivers and approved forecast assumptions
• Challenge forecasts through sensitivity analysis, back-testing and management judgement
• Present forecast uncertainty using ranges, confidence intervals and clear caveats

Participants can apply AI support to reporting work while preserving IFRS evidence and review requirements.

• Map AI-supported reporting tasks to IFRS disclosure and evidence requirements
• Draft flux analysis explaining period movements with traceable source evidence
• Check consistency between notes, primary statements and management-reporting narratives
• Identify judgments requiring controller review under IAS 1 and IAS 8
• Maintain evidence packs linking generated text to validated source data
• Apply materiality thresholds when reviewing automated reporting exceptions

Participants can use AI-assisted analytics while distinguishing reliable audit evidence from generated interpretation.

• Plan audit analytics around assertions, risks, populations and test objectives
• Use AI to select unusual transactions for targeted substantive procedures
• Distinguish audit evidence from AI-generated summaries and interpretations
• Evaluate model outputs for completeness, accuracy and reproducibility
• Document procedures consistent with ISA 315 and ISA 330
• Preserve audit trails for prompts, data extracts, outputs and reviews

Participants can design and test controls for AI-assisted finance processes.

• Design preventive, detective and corrective controls around AI-assisted processes
• Apply COSO principles to model ownership, monitoring and information quality
• Segregate data preparation, model operation, approval and override responsibilities
• Define control evidence for automated reconciliations and exception decisions
• Test access controls, change management and output-review procedures
• Record residual risks in an AI control register

Participants can establish governance requirements for finance AI models, data and decisions.

• Apply NIST AI RMF functions to finance-model governance decisions
• Classify finance AI use cases by impact, autonomy and data sensitivity
• Assess bias risks in credit, supplier and employee-related decisions
• Set retention, confidentiality and cross-border data handling requirements
• Create model cards covering purpose, limitations, inputs and monitoring
• Define escalation paths for material errors, incidents and policy breaches

Participants can select finance AI initiatives using documented value, risk and delivery criteria.

• Build an AI use-case backlog using value, feasibility and risk scoring
• Estimate benefits from hours saved, error reduction and working-capital improvements
• Specify pilot scope, baseline measures, acceptance criteria and exit conditions
• Assign process owner, data owner, model owner and control owner
• Plan user adoption through role-based guidance and feedback loops
• Prepare vendor due-diligence questions covering security, data use and support

Participants can assess AI vendors and platform options against finance control and integration requirements.

• Compare generative AI platforms by data isolation, auditability and administration
• Review service agreements for intellectual property, retention and liability clauses
• Assess third-party model documentation, testing evidence and incident procedures
• Define API integration requirements with ERP, consolidation and workflow systems
• Set procurement evaluation weights for functionality, controls, cost and scalability
• Document exit plans, portability requirements and supplier concentration risks

Participants can develop an implementation-ready design for an AI-assisted finance process.

• Create an AI-assisted close improvement design for a selected process
• Map current-state controls, data flows, exceptions and handoffs
• Specify future-state workflow, model boundaries and human review points
• Develop test cases covering accuracy, security, bias and failure modes
• Calculate benefit hypotheses, implementation costs and control-maintenance effort
• Prepare stakeholder communication for finance leadership, IT and internal audit

Participants can build, review and present a controlled AI-assisted finance process proposal with evidence, metrics and an implementation plan.

• Select a finance use case with defined decision and value criteria
• Prepare data requirements, process maps and control objectives
• Design prompts, analytics or workflow rules with human review
• Evaluate outputs against accuracy, materiality and exception-handling criteria
• Compile governance, risk, model-monitoring and audit-trail requirements
• Present business case, pilot plan and control design to reviewers