Something is changing in how South Africans find businesses online. A growing number of people — particularly in professional and corporate sectors — are asking ChatGPT, Google's AI Mode, Perplexity, and Gemini for recommendations instead of typing queries into traditional Google Search. "Which digital marketing agency should I use in Johannesburg?" is now a question being asked to an AI, not just Google.
This shift has enormous implications for visibility. If you're not in the sources these AI systems draw from, you simply don't exist in those answers. Here's what's actually happening and what you can do about it.
How AI Search Engines Choose Their Sources
AI search tools like ChatGPT (with search), Perplexity, Google AI Overviews, and Microsoft Copilot all work by retrieving information from crawled web content and synthesising it into a direct answer. They're not pulling from a private database of "best businesses" — they're reading and weighing the same web content that traditional search engines index, plus a few additional signals.
The factors that determine whether an AI system cites or recommends your business:
- Consistent mentions across authoritative sources — reviews on Google, citations in industry publications, listings on reputable directories
- Clear, structured content on your own site — FAQ sections, well-structured articles, schema markup that makes your information machine-readable
- E-E-A-T signals (Experience, Expertise, Authoritativeness, Trustworthiness) — author bios, credentials, certifications displayed prominently
- Fresh, regularly updated content — AI crawlers favour sites that are actively maintained
- llms.txt — an emerging standard (like robots.txt, but for LLMs) that gives AI crawlers a concise, machine-readable summary of your business
The llms.txt Standard
Several leading AI systems now look for a file at /llms.txt on your domain before crawling the rest of your site. This file gives AI systems a structured summary of who you are, what you do, and where to find your key content — without having to interpret your full site. Nines Media has published an llms.txt at ninesmedia.com/llms.txt as part of our own AI visibility strategy.
Your llms.txt should include: a one-paragraph description of your business and what makes it distinctive, your key services and service areas, links to your most important pages and blog articles, and your contact information. Keep it factual and concise — this is for AI parsers, not humans.
Structured Data: Talking to AI Directly
Schema markup is JSON-LD code embedded in your website that tells search engines (and AI systems) exactly what your content is. For businesses, the most important schema types are:
- LocalBusiness schema — your name, address, phone, opening hours, services, rating, and geographic service area
- FAQPage schema — marks up your FAQ section so AI systems can extract question/answer pairs directly
- Article schema with Person author — shows AI systems who wrote your content and their credentials
- Service schema — details the specific services you offer, their descriptions, and service areas
If your site doesn't have structured data, implement it. It's one of the clearest signals you can send to both Google and AI systems about what your business does and who it serves.
Content Strategy for AI Discoverability
Write content in the way AI systems look for information. This means: answer questions directly and early in your articles (don't bury the answer after three paragraphs of preamble), use natural question-and-answer formatting in FAQ sections, and write at the level of expertise you actually have — AI systems are increasingly good at detecting shallow, AI-generated thin content versus genuine expertise.
The questions that matter are the ones people actually ask AI systems: "Who is the best [service] in [city]?", "What should I look for in a [service provider]?", "How much does [service] cost in South Africa?" — these are the queries that AI systems answer, and they draw from articles that directly address them.
Third-Party Citations: The Trust Signal AI Relies On
AI systems give weight to businesses mentioned consistently across multiple credible sources. For South African businesses, this means: getting listed on Clutch.co, GoodFirms, or UpCity (for agencies), being featured in Business Day, Bizcommunity, or industry publications, getting testimonials published on platforms like Google and Hellopeter, and earning genuine press mentions.
The underlying logic is the same as traditional E-E-A-T: an AI system is more likely to recommend a business that multiple independent sources have validated than one that only claims credibility on its own website.