The CIO didn’t Google you. She asked ChatGPT. And you weren’t in the answer.
This is not a hypothetical. Research from 6sense suggests that a large majority of B2B buyers now use a large language model during their buying journey – not as a novelty, but as a primary research tool. They ask it who the credible players are in their category, what the trade-offs look like and which providers have a track record worth examining.
The model answers.
And the answer is shaped by a combination of two things: what the model absorbed during training, and what it retrieves when it goes looking for fresh information in real time. Both pathways favour the same kind of source.
Here is the part most B2B marketers have not caught up with. The consumer-facing LLMs – ChatGPT, Gemini, Claude, Perplexity, Copilot – don’t simply regurgitate a training corpus. When asked a real buyer-research question, they increasingly combine what they already “know” with a live web search, then cite their sources. In both layers, established media platforms, news archives, research papers and attributed expert commentary punch well above their weight. Corporate blogs sometimes surface. Gated white papers rarely do. Social content lands inconsistently and tends to be treated as “low signal”.
There is a pattern in what gets cited: content with provenance; articles published on trusted platforms; named sources quoted in industry coverage; and expert perspectives attributed to real people on platforms with established credibility. The model doesn’t read your unpublished draft, it reads what a credible third party chose to publish.
The shift that matters
For 20 years, the game was SEO – get your website to rank on page one of Google. That game still matters, but a second game is now running alongside it, and for senior B2B buyers it may already be more influential. Call it generative engine optimisation (GEO): the question is no longer just “does my website rank?” but “does my brand appear when an AI is asked who the experts are in my space?”.
The Bundesverband Industrie Kommunikation’s Trendbarometer – an annual survey of B2B marketers published by Germany’s leading industrial communications body – reported that a significant majority of respondents now regard GEO as essential to content strategy over the next two years. That number will climb as the models improve and adoption accelerates among the exact senior audience most B2B companies are trying to reach.
South Africa’s B2B decision-maker audience is large, identifiable and concentrated. A platform like TechCentral reaches hundreds of thousands of readers each month, a significant share of whom are MDs, CEOs, business owners and senior executives. That is not a niche – it is a meaningful portion of the country’s enterprise buying power, gathered in one place.
And that makes the AI visibility question even more pointed.
When that audience turns to an LLM to research a vendor, the model – whether drawing on training data, a live web search or both – will surface whatever published record exists. If your competitor has a consistent presence on credible platforms, with published insights, named expertise and a visible point of view on the issues that matter, and you do not, the model knows them and doesn’t know you. In a market where senior decision-makers are reachable and actively using AI to inform their choices, being absent from the published record on platforms they trust is no longer a gap in your marketing plan – it is a gap in your pipeline.
The brands that appear in those answers share a pattern. They have a consistent publishing footprint on platforms the models trust. Their executives are quoted by name in industry analysis. They have built, over months and years, a body of attributed expertise that sits in the public record rather than behind a login wall or buried in a PDF.
This is not about volume, iit is about provenance. One well-placed article on a credible platform, quoting your CEO on a genuine market issue, carries more weight in the way an LLM answers a buyer question than a hundred social posts or a dozen self-published blog entries. The model – and its retrieval layer – doesn’t care how much you spent on the content. It cares whether a trusted source published it.
There is an uncomfortable truth for both brands and publishers. The content that earns AI visibility is the content that is genuinely useful – not the content designed to pass an internal approval chain or optimised for a campaign calendar. If a piece reads like marketing, the model treats it like marketing. If it reads like journalism, it gets cited like journalism.
The window is still open
In most South African B2B categories, the published landscape is thin enough that a determined company could establish a footprint strong enough to become the name the AI returns when the question is asked. A cybersecurity firm, a cloud provider, a fintech platform – any of these could own their category in the model’s answers within 12 months of consistent, credible presence on the right platforms.
But only if they start treating visibility on trusted platforms as infrastructure, not as a campaign.
The CIO will ask the question again next quarter. The model will answer with whatever it can find – in its training data, on the open web, or both. The only variable is whether your brand is part of that answer, or whether the AI has never heard of you.
Publishared is the exclusive B2B media partner for TechCentral, South Africa’s trusted business technology news platform. From branded content and executive positioning to events and podcasts, Publishared helps brands build credible, sustained visibility in front of the decision-makers who matter. If these are the people you need to reach, let’s talk.
- The author, Michelle Losco, is CEO of Publishared
- This article is produced in association with Publishared, TechCentral’s commercial partner, and is sponsored.