Session 1: Secure Foundations and Everyday AI Applications
Purpose: Establish safe working practices and equip delegates to use AI for research, communication, documentation and accounting tasks.
1. IT security and responsible AI use
- Identifying current technology use, including personal AI accounts and unapproved applications.
- Protecting client and firm information through secure access, multifactor authentication and appropriate file-sharing practices.
- Recognising phishing, fraudulent banking instructions, deepfakes and impersonation.
- Understanding confidentiality, data protection and the risks of uploading information to AI platforms.
- Establishing interim rules: approved tools, permitted tasks, restricted information, review requirements and escalation.
- Understanding AI limitations, including fabricated information, bias, outdated content and overreliance.
- Retaining professional judgement and responsibility for all work produced.
2. Effective prompting, research and productivity
- Writing prompts with a clear task, relevant context, constraints and required output.
- Asking AI to identify missing information and distinguish facts from assumptions.
- Drafting client emails, information requests, meeting summaries and action lists.
- Developing checklists, templates and internal guidance.
- Summarising documents without losing significant qualifications or exceptions.
- Using AI to assist technical research and verifying references against authoritative sources.
- Reviewing outputs for accuracy, completeness, tone and confidentiality.
Break
3. Accounting and financial reporting applications
- Drafting accounting memoranda from supplied facts and verified technical material.
- Preparing preliminary accounting policies and disclosure wording.
- Identifying generic or unsupported statements in AI-generated drafts.
- Developing management commentary and variance explanations.
- Using AI to assist with Excel formulas and reconciliation approaches.
- Independently checking calculations and technical conclusions.
4. Establishing the firm’s initial AI protocol — 30 minutes
- Identifying suitable initial use cases.
- Assigning responsibility for tool approval and output review.
- Agreeing what information may be entered into approved tools.
- Defining when employees must seek guidance or report an incident.
- Introducing an AI-use and review checklist.
Session 2: Practical AI in Audit, Tax and Data Analysis
Purpose: Apply AI to professional assignments, validate results and develop a practical implementation plan.
1. Audit procedures and documentation
- Using AI to summarise client background and identify matters for further enquiry.
- Developing preliminary risk considerations from supplied information.
- Linking risks, assertions and proposed audit procedures
- Improving the clarity and specificity of audit procedures.
- Summarising walkthrough notes and identifying documentation gaps.
- Supporting analytical review and developing follow-up questions.
- Reviewing working papers for inconsistencies and unsupported conclusions.
- Documenting significant AI use, validation and reviewer involvement.
- Recognising that AI-generated text cannot establish that audit work was performed or evidence obtained.
2. Tax research and client communication
- Structuring tax research questions around the relevant facts, jurisdiction and tax period.
- Using AI to assist with legislation summaries and research planning.
- Checking effective dates, legislative references and case citations.
- Drafting SARS correspondence, supporting-document requests and preliminary client explanations.
- Identifying missing facts and matters requiring specialist review.
- Reviewing calculations and advice before communication or submission.
Break
3. Data analysis and fraud indicators
- Preparing transaction data and checking completeness through record counts and control totals.
- Using AI to assist with formulas, queries and analysis steps.
- Identifying duplicate payments, unusual journals and unexpected transaction patterns.
- Testing analytical logic against known examples.
- Distinguishing anomalies from errors and evidence of fraud.
- Investigating false positives and documenting follow-up.
- Retaining reproducible analysis and independently checking selected results.
4. Implementation, governance and action planning
- Identifying the essential components of IT and AI acceptable-use policies.
- Establishing an approved-tool register and approval process.
- Defining responsibilities for confidentiality, review, incident reporting and monitoring.
- Selecting initial applications suitable for a controlled pilot.
- Developing a 30-, 60- and 90-day implementation plan.
- Assessing success through accuracy, time saved, review findings and rework.
