Top stories



Marketing & MediaYour dashboard is lying to you: Why brands keep missing the 1 trillion rands township economy
Kabelo Kale, The Media Krate 2 days


MO Agency, one of South Africa's leading revenue and AI-visibility consultancies, today published the South African AI Visibility Report 2026, the first study of its kind and the most extensive measurement of AI brand recommendations conducted in the country to date.
As South Africans increasingly begin their buying decisions inside an AI assistant rather than on a search engine, the report answers a question no local study has tackled at this scale: when a consumer asks ChatGPT, Claude or Gemini for the best bank, medical aid, car or insurer, which brands do the assistants actually recommend, and which do they leave out?
To find out, MO Agency'' used its proprietary AI-visibility platform, [[Getmd.ai, to put the real buyer-intent questions South Africans ask to the three leading assistants, ChatGPT (OpenAI), Claude (Anthropic) and Gemini (Google), each with live web search enabled. Twenty questions per industry, asked to three assistants and repeated three times each, produced 180 answers per industry and 3,240 in total, drawing on 21,433 source citations across 18 industries and 256 brands. Every brand was scored on a single AI Visibility Quality Score out of 100.
No South African brand scored above 70 out of 100 in any category. The strongest performer is Payfast in payment providers at 66.3, which means even the country's most AI-visible brand is recommended with only partial confidence.
Five of the 18 industries have no clear leader at all. Short-term insurance, life and long-term insurance, investment and asset management, private tertiary education, and IT managed services are all still open, leaving them winnable by whichever brand acts first.
The median category leader scores just 48.3 out of 100, so in most industries the AI's default choice is weak and there to be taken.
Several of the country's best-known names are mentioned often but recommended rarely, including Volkswagen, Absa, Hollywoodbets, Rain, and Peach Payments, visible in the conversation, but seldom the brand the assistant actually puts forward.
One of the report's clearest patterns is that in many categories the established market leader has naturally become the AI's default recommendation. Toyota leads car brands at 59.6, MTN leads mobile networks and Standard Bank leads commercial banking, both at 53.8, FNB leads business banking at 53.5, Discovery Health leads medical schemes at 50.8, and Capitec leads retail banking at 43.8. In these industries, years of brand equity, customer scale and content presence appear to have carried over into how the AI models describe the market, so the real-world leader has quietly become the AI-era leader too.
But the report shows that established market leadership carry-over is not guaranteed. In payment providers, Payfast is the closest any brand comes to owning its category at 66.3 out of 100, while a familiar name like Peach Payments is named often but rarely the one the assistant recommends. Household names such as Volkswagen, Absa and Hollywoodbets are named constantly yet rarely recommended, and in five industries – short-term insurance (car and home), life and long-term insurance, investment and asset management, private tertiary education, and IT managed services and systems integrators – no brand has emerged as the AI's pick at all. AI visibility, in other words, has to be earned in its own right.
"For the first time we can see exactly what AI is telling South African consumers, and the picture is striking," said Luke Marthinusen, founder and CEO of MO Agency. "No local brand yet owns its category in the eyes of the AI, and in five major industries the assistants have not picked a favourite at all. That is both a warning and an opportunity. The brands that understand this now will become the default recommendation in their market. The ones that ignore it will quietly disappear from the conversation at the exact moment a customer is deciding."
The shift driving these findings has a name, or rather three. AI search is the growing trend of asking an AI assistant for a recommendation instead of typing a query into Google and choosing from a page of links. The assistant replies not with blue links but with a short, confident shortlist of named brands.
Generative Engine Optimisation (GEO) and Answer Engine Optimisation (AEO) are the emerging disciplines that respond to it. They are the practice of making a brand visible, credible and recommendable to generative AI engines and answer engines, so that the assistant names the brand when a buyer asks. Where search engine optimisation worked to rank a website on a results page, GEO and AEO work to make the brand the answer itself.
The report also reveals how the assistants form their views. Across all 18 industries, they leaned on a small set of trusted sources, led by businesstech.co.za, cited 329 times across eight industries, followed by Hippo, MyBroadband, Wikipedia and Reddit. Earning presence on those sources is one of the most direct ways a brand can influence what AI recommends.
The South African AI Visibility Report 2026 is an extension of the AI-visibility work MO Agency already delivers for clients across financial services, technology, professional services and consumer sectors, monitoring and improving how their brands appear inside AI answers through the Getmd.ai platform.
"We have been doing this work for clients for some time, so this report is really a public view of something we live in every day," said Marthinusen. "Answer Engine Optimisation is roughly where search engine optimisation was fifteen years ago. The rules are being written right now, and South African brands have a rare chance to get ahead of them rather than scramble to catch up later. That is exactly where Getmd.ai and our team are focused."
The full South African AI Visibility Report 2026 is free to read at https://www.mo.agency/ai-visibility-report-south-africa, including the full brand rankings, recommendation-strength breakdowns and source analysis for each of the 18 industries.