How widespread is it?
Very, and shallow. McKinsey’s 2025 global survey of 1,993 organisations found 88% using AI in at least one business function, up from 78% a year earlier, with most organisations still in pilot mode and only a minority reporting a measurable effect on profit 1. The South African small-business picture matches: nearly everyone has tried a tool; far fewer have a process. The difference between the two groups is the subject of this article.
What are the first five jobs?
| Job | The leak it closes | Evidence |
|---|---|---|
| Answering the website, at any hour | Enquiries that wait until morning | An hour’s delay cuts lead qualification about sevenfold; 23% of firms never reply 2 |
| Filling the diary and reminding | Missed appointments | About 23% no-show on average; text reminders raise attendance 4 5 |
| Sending and chasing invoices | Late payment | 91% of SA small businesses paid late, 18 days past due on average 3 |
| Clearing the inbox | Slow replies | 80% of customers say experience matters as much as the product 6 |
| Writing the follow-up | Leads that go cold | Only 7% of 433 companies replied within five minutes 7 |
Why these five?
Because each has a number attached and each is rule-bound. Answering a repeated question, booking a slot, sending a reminder, issuing an invoice, drafting a follow-up: none of them needs judgement, all of them are done late or not at all by a busy owner, and all of them can be measured before and after. The jobs that need judgement, the complaint, the negotiation, the exception, stay with a person, and the businesses that try to automate those are the ones with the cautionary tales.
What separates the return from the pilot?
Three habits. The content is real: the chatbot is trained on the actual price list, not a draft. Someone owns it: one person reads the conversations and corrects the answers in the first week. One number is measured: reply time, no-show rate, days to get paid, before and after. Businesses that do these three see the effect in a month; businesses that install a tool and look away are the 88% with nothing to show.
What is specifically South African about it?
Language, connection and money. The customer writes in isiZulu or Afrikaans and expects an answer in it. The connection is mobile data, so a heavy page loses the visitor before any AI helps. And cash flow is the binding constraint: Xero’s report on late payment found 91% of small businesses affected and 46% spending one to two months a year chasing money 3. An invoice with a pay-now link sent from the van is a South African use case before it is an AI one.
Where to start?
With the leak you can measure. For most businesses that is the website: a chatbot trained on your content, live in about 30 minutes, month to month, with a free trial of fifty messages over seven days to check your own enquiry volume. Then the diary, then the invoices. One at a time, each measured.
Questions people ask
Is AI automation only for big companies?
No. The five jobs above are small-business jobs, and the products that do them are priced monthly in rands with no contract.
How much technical skill do I need?
None for these five. The setup is questions about your business; the install is one line of code or a plugin.
What should I measure?
One number per job: reply time for the website, no-show rate for the diary, days to payment for invoices, hours to reply for the inbox. Before and after, on your own data.
Sources
- McKinsey & Company, “The State of AI: Global Survey 2025” (1,993 respondents, 105 countries)
- Oldroyd, McElheran & Elkington, “The Short Life of Online Sales Leads”, Harvard Business Review, March 2011 (audit of 2,241 US companies)
- Xero, “South Africa Late Payments Report” (2025), reported by IOL, January 2025
- Dantas, Fleck, Cyrino Oliveira & Hamacher, “No-shows in appointment scheduling – a systematic literature review”, Health Policy 122(4), 2018 (105 studies)
- Gurol-Urganci et al., “Mobile phone messaging reminders for attendance at healthcare appointments”, Cochrane Database of Systematic Reviews, 2013 (8 trials, 6,615 participants)
- Salesforce, “State of the Connected Customer”, sixth edition (2023)
- Drift, “The 2017 Lead Response Survey” — a secret shopper against 433 B2B companies