South Africa’s artificial intelligence policy remains in a reset phase. Cabinet approved the withdrawal of the draft national AI policy on 3 June 2026, saying the document would be reworked to ensure it achieves its intended goals and establishes national standards for the ethical use of AI. The Presidency said the draft had originally been published for public comment in March 2026.

The issue remains active. In September, Communications and Digital Technologies Minister Solly Malatsi said digital governance must be robust enough to adapt to rapid changes in artificial intelligence, cybersecurity, regulation and social norms. His GovTech 2026 address also linked digital governance to South Africa’s broader development and economic goals.

What the policy pause means for smaller businesses

The immediate message for owners and marketing teams is not to wait for a final policy before introducing basic AI controls. The policy position may still develop, but businesses are already using AI to draft content, summarise enquiries, qualify leads and suggest follow-up messages.

That creates a practical governance question: who checks the output, what information may be entered into a tool, and how does a customer know when an automated process is involved?

These questions are especially important when AI is connected to lead generation and customer data. A fast draft is useful, but an unchecked message can contain an incorrect product detail, an inappropriate promise or information that should not have been shared outside the business.

Three controls to introduce now

  • Define permitted uses. Start with low-risk tasks such as outlining an article, suggesting subject lines or turning approved notes into a first draft. Require human review before anything is sent to a prospect or customer.
  • Protect the source data. Create a simple rule for what staff may paste into AI tools. Avoid using unnecessary personal information, confidential pricing or sensitive customer notes when a less detailed prompt will do.
  • Keep an approval trail. Record who reviewed important AI-assisted content, what was changed and when it was approved. This is useful for training, quality control and resolving disputes about customer communication.

Make the process visible to the team

AI governance works better when it is built into the everyday workflow rather than left as a document that no one opens. A shared CRM can help teams keep leads and contacts in one place, assign selected contacts to an opt-in invitation, record consent and maintain subscriber lists.

From there, branded email design, timed nurture sequences and open and link tracking can create a clearer handover between marketing and sales. Content Studio can also help teams keep editable articles and social posts together, so an approved message is easier to update than a collection of disconnected drafts.

This does not remove the need for judgement. It reduces the risk of a useful lead, approval or correction being lost between spreadsheets, inboxes and individual tools.

A sensible next step

South Africa’s AI policy will need to be followed as it develops, but smaller businesses can act on sound operational principles now: use AI for assistance, keep people accountable, limit unnecessary data sharing and retain clear records.

For businesses reviewing their customer follow-up process, a connected email-sequence workflow can provide a practical place to apply these controls without turning every campaign into a manual handover.

Explore a connected lead nurturing setup to bring permission records, branded email and customer follow-up into one manageable workflow.