The State of AI Search Visibility in South Africa (2026)

South African brands are largely absent from the moment AI decides what to recommend. In the most extensive study of AI search visibility conducted in the country to date, MO Agency analysed 3,417 AI prompt answers and 23,684 source citations across 19 industries and more than 260 brands - and found that no South African brand scored above 75 out of 100 in any category. AI assistants rarely recommend any local brand with full confidence, and in five of nineteen industries they have not settled on a leader at all. This is the current state of AI search in South Africa, and for brands willing to act, it is unusually winnable.

Key finding: the median category leader in South Africa scores just 48.3/100 on AI search visibility. In most industries, the AI default is weak - which means the brands that invest in AI search optimisation now can take positions that are still open.

What the results look like

Below is a sample of the results for the Retail Banking and Automotive industries in South Africa. View the SA AI Visibility Report 2026 for the full industry listings and reports.

The numbers at a glance

The data is drawn from the South African AI Visibility Report 2026, now expanded to a nineteenth industry - vehicle tracking.

How the study was built (methodology)

The method is identical across all 19 industries, so the results are directly comparable.

For each industry, MO Agency wrote 20 buyer-intent questions - the questions South Africans genuinely ask when choosing a provider - grouped into seven intents: ranking ("what is the best X in South Africa?"), ease ("easiest to sign up or claim?"), comparison ("Brand A vs Brand B"), segment ("best X for a student, family or business"), criteria ("cheapest, best value, most reliable"), trust ("most trusted"), and problem ("I need X, who should I choose?").

Each question was put to three AI assistants - ChatGPT (OpenAI), Claude (Anthropic), and Gemini (Google) - with live web search and grounding switched on, so the answers reflect current, sourced information rather than stale training data. Every question was asked three times per assistant to smooth run-to-run variation. That is 20 questions × 3 assistants × 3 runs = 180 answers per industry, and 3420 answers in total.

A second AI pass (Claude) then reads each answer and extracts every brand named, where it ranks, and whether it is an active recommendation, a neutral mention, or a caveated one. Brand-name variants (for example "VW" and "Volkswagen") are mapped to a single canonical brand so visibility is never split.

How the Quality Score works

Every brand gets a single score from 0 to 100 that rewards being named, weighted by how prominently:

• A brand scores 1 ÷ its rank in each ranked answer (1st = 1.0, 2nd = 0.5, 3rd = 0.33, capped at 10th).
• A passing mention with no ranked position scores just behind the last ranked item.
• A brand absent from an answer scores 0.
• The score is the average across every answer - including the ones where the brand was absent - multiplied by 100.

So a brand recommended first in every answer scores 100; a brand mentioned often but always near the bottom scores low. Crucially, this measures AI visibility, not market share, revenue, or product quality. A brand can be excellent and under-visible, or modest and over-visible.

Five findings that define the state of AI search in South Africa

1. Two kinds of market: Decided and undecided

Some categories already have a clear AI "default answer" - a brand the assistants reach for first, by a wide margin:

• Payments - Payfast (66.3, a 34-point lead)
• Car brands - Toyota (59.6, a 26-point lead)
• Vehicle tracking - Cartrack (57.6, a 19-point lead)
• Business & commercial insurance - Santam (43.9, a 25-point lead)
• Corporate banking - Standard Bank (53.8, a 24-point lead)
• SME banking - FNB (53.5, a 24-point lead)

Others are genuinely undecided - the leader is barely ahead of the chasing pack, so no brand owns the category in the AI's mind yet:

• Private tertiary education - Varsity College edges Eduvos by 0.7 points
• Short-term insurance - Santam edges OUTsurance by 1.6
• Investment & asset management - Allan Gray edges EasyEquities by 2.4
• Life insurance - Discovery Life edges Sanlam by 4.5

In a decided category the near-term goal is to win the second look, not to unseat the leader overnight. In an undecided category, the number-one position in AI search is there for the taking with a deliberate push.

2. Awareness is not endorsement - the household-name tax

The most consistent finding in the index is the gap between how often a brand is mentioned and how often it is actually recommended. Many of the biggest, most established brands are named almost everywhere but recommended in only a fraction of those mentions - Sanlam, Old Mutual, Nedbank, Standard Bank, Absa, Vodacom, Telkom, Volkswagen, Suzuki, RE/MAX, Seeff, Vumatel and Hollywoodbets among them.

These brands have already paid for awareness. The AI knows them. It just does not champion them - often deferring to a more focused challenger for the actual recommendation. That gap between being known and being chosen is the single most addressable opportunity in the dataset.

3. Challengers are punching above their weight

In category after category, newer or more focused brands out-rank far larger incumbents on AI recommendation:

• TymeBank is the #2 most visible SME bank, ahead of every traditional big bank.
• EasyEquities is the #2 investment brand, ahead of giants many times its size.
• Easybet is the #3 betting brand, beating bigger books on recommendation rate.
• Naked out-performs several established insurers in short-term cover.
• Chinese car brands (Chery, GWM, Haval) are firmly in the conversation.
• In IT services, specialists like Turrito Networks and Westech out-rank legacy enterprise integrators.

AI assistants reward clarity of positioning. A brand that is unambiguously "the easy one", "the cheap one" or "the one for X" tends to get recommended for that job, regardless of size.

4. The invisible-giant risk

The flip side: Several brands that matter in the real world are almost absent from AI answers, scoring near zero in their category - a sign their content and authority signals are not reaching the assistants. Invisible in AI means absent from the shortlist. In IT managed services the gap was so large it reshaped the study: the assistants recommended SMME-focused providers that were not even on the initial watchlist far more often than the enterprise integrators that were. The AI's view of "best" did not match the industry's view of "biggest".

5. Whose content the AI actually reads

Across the index MO Agency logged 23,684 source citations - the specific web pages the assistants drew on to build their answers. A small set of trusted South African domains does most of the work: BusinessTech was cited hundreds of times across multiple industries, alongside Hippo, MyBroadband, TechPoint Africa, Wikipedia and Reddit. AI visibility is not a black box - it is a function of what the models can find, read and trust about you. That is the lever.

Who AI recommends in each South African industry

The brand each set of assistants most often puts forward when a South African asks for the best option, scored 0-100.

Even the top brand in the country's strongest category is recommended with only moderate confidence. The picture is one of an open market, not a settled one.

Spotlight: Vehicle Tracking 

Vehicle tracking is one of the more decided categories in the index. Cartrack leads with a Quality Score of 57.6 - a commanding but not unassailable 19.2-point lead over the chasing pack. It is named in 98% of all answers and is the active recommendation in 73% of them, at an average rank of 1.76. Behind it, Tracker, Netstar, Matrix and Beame are widely mentioned yet under-recommended - the clearest opening in the category for a challenger that sharpens its AI-readable content and authority signals. The category was measured across 177 answers and 2,251 citations.

Why this matters now

South African buyers increasingly begin their research not on Google but inside an AI assistant: "What is the best provider for me?" The assistant answers with a short, confident list. The brands it names enter the consideration set; the brands left off never get a chance. Because the assistants are still forming their preferences in most South African categories, the cost of being invisible today is compounding - and the advantage of establishing AI search visibility early is durable. For the wider shift behind these numbers, see how AI search is changing traffic.

What brands should do about it

The response to this data is AI search optimisation - the discipline of getting named and recommended inside AI answers. In practice that means:

• Measure first. Benchmark where you stand across ChatGPT, Claude and Gemini before doing anything else. A free AI Visibility Audit is the fastest way in.
• Make your content extractable. Structure key pages for AI retrieval, with clear answer blocks, schema and clean machine-readable formats.
• Earn the sources AI reads. Prioritise presence on the South African domains that shape answers - BusinessTech, MyBroadband and sector titles.
• Close the endorsement gap. If you are widely mentioned but rarely recommended, the work is sharpening positioning and corroboration, not building awareness.
• Iterate across engines. Track citations, sentiment and share of voice monthly, because the assistants change how they cite.

For the full method, read how to rank in AI search results in South Africa.

Frequently asked questions

What is the South African AI Visibility Report?

It is a study by MO Agency measuring how the leading AI assistants - ChatGPT, Claude and Gemini - recommend brands to South African buyers. It analysed 3,417 AI answers and 23,684 source citations across 19 industries and more than 260 brands, scoring each brand on a 0-100 AI Visibility Quality Score.

How many industries does the report cover?

Nineteen, following the addition of vehicle tracking. They span banking, insurance and medical schemes, investments, payments, connectivity and IT, and consumer categories including automotive, vehicle tracking, estate agencies, private tertiary education, and betting and gaming.

How visible are South African brands in AI search?

Low, on the whole. No South African brand scored above 75/100 in any of the 19 industries measured, and the median category leader scored just 48.3/100. In five industries there is no clear AI-recommended leader at all, indicating most categories are still open.

Which brand has the highest AI visibility in South Africa?

Payfast, in payment providers, with a Quality Score of 66.3/100 - the highest of any brand in any category. It is followed by Toyota (car brands, 59.6) and Cartrack (vehicle tracking, 57.6).

Which South African websites most influence AI answers?

A small set of trusted domains does most of the work, led by BusinessTech and including Hippo, MyBroadband, TechPoint Africa, Wikipedia and Reddit. Earning presence on these sources is one of the most direct ways to influence what AI assistants recommend.

How can a brand improve its AI search visibility?

Through AI search optimisation - AEO and GEO: publishing clear, structured, citable content, strengthening entity and schema markup, and earning mentions on the authoritative third-party sources AI assistants read. MO Agency offers this as a measured, revenue-linked programme.

See where your brand stands

Find out where your brand appears in AI answers - and how to become the one ChatGPT, Claude and Gemini recommend. Get your AI Visibility Audit or explore the full South African AI Visibility Report 2026.

Cite this analysis: MO Agency (2026). The State of AI Search Visibility in South Africa. Based on the South African AI Visibility Report 2026.