GenAI Tools for Enhancing Client Engagement
Presenters : Nestene Botha
Overview
The bottleneck in client work is rarely the technical answer. It is the time needed to explain that answer properly: letters that clients do not act on, follow-up emails that are never written, and meeting preparation that gets skipped. Generative AI is most valuable where the practitioner supplies the facts and the tool improves the language. It is weakest, and riskiest, when it is asked to supply the substance.
This session shows how to start from a practitioner's own file note, describe the actual reader, and instruct the tool so that it adds no facts, figures, dates or legal references of its own. It covers what must still be checked before any AI-assisted letter goes out, how to keep a practice's own voice, and a worked example of an auto-assessment client letter. It then applies the same approach to meeting preparation and follow-up, the consent and POPIA implications of recording and transcribing meetings, and building a reusable library of engagement letters, onboarding sequences, standard SARS correspondence and proposals.
The final section addresses disclosure. It sets out what SAICA's GenAI guidance says on transparency and how that guidance relates to the binding SAICA Code of Professional Conduct, and it contrasts a narrower practical position with the guidance as written, identifying the guidance as the more conservative place to stand. It also covers engagement-letter wording informed by UK ICAEW and PCRT guidance, the hard line drawn by POPIA section 71 on automated decisions, and the principle that advice sent under a practitioner's name remains that practitioner's responsibility, whatever drafted it.
Topics covered
- Where GenAI genuinely helps: the real bottleneck in client work, the line between using the tool for language and relying on it for substance, and the vendor's own "draft and review" framing
- Explaining a technical position: starting from the file note, describing the reader, the three things to ask for every time, the no-new-facts instruction, what to check before sending, controlling tone, and a worked SARS auto-assessment letter
- Meetings, before and after: structured meeting preparation, what material to supply, follow-up emails that actually go out, and the consent, discoverability and section 72 implications of recording and transcription
- Proposals, onboarding and the client-facing library: avoiding communication volume for its own sake, building house versions of recurring letters, the first-six-weeks onboarding sequence, specific proposals, answering "did AI write this?", and the risk that AI-drafted content resembles a competitor's
- Disclosure and the limits: SAICA's GenAI guidance on transparency and its status relative to the SAICA Code, practical disclosure positions, engagement-letter wording, POPIA section 71 on automated decisions, the IESBA technology-related revisions and ICAEW guidance, and the tasks that should not be delegated to AI
Learning outcomes
Practitioners will be able to:
- Identify the client-facing tasks where generative AI adds value and those where it adds risk
- Rewrite a technical position into language that a specific client will act on
- Build a reusable prompt and context pack for recurring client communication
- Prepare for and follow up a client meeting more thoroughly in less time
- Explain where POPIA constrains automated decisions affecting clients
- State what SAICA's GenAI guidance requires on disclosure, and adopt a defensible practice position
Who should attend
Tax practitioners, accountants, practice owners and client-facing staff who write client letters, proposals and follow-up correspondence and want to use generative AI to do so more clearly and efficiently, within their professional obligations.
Related Webinars
Integrating Systems for Seamless Workflow and Operations: From Manual to Digital
Cybersecurity for Small Practices: Safeguarding Client Data in a Cloud Environment