OpenAI was set to dethrone LinkedIn this summer. An in-house job platform, powered by AI, that matches candidates and employers based on what someone can do rather than what is in a CV. Announced on 4 September 2025, promised for mid-2026.
That moment has passed. No platform, no beta, no waiting list.
And yet in those twelve months more shifted than a launched product could ever have done.
What actually shifted in those twelve months
While the job board was slow to materialise, search behaviour itself crept into AI. Since February, the Indeed app has been in ChatGPT. You ask a question, you get job vacancies back, you compare employers, all within the same conversation. No ten blue links, no careers site, one answer.
ChatGPT now has around 900 million weekly users. A significant portion of them are looking for work-related information. Increasingly less on Google. More and more often within a conversation that knows them.
The questions on which the decision turns
The questions are concrete. Which employer in my field has a good culture, where can I genuinely progress, which organisation matches what I can do and what I want. Decisions are based on those exploratory questions, and that is precisely where most brands fall short.
In addition, OpenAI was building certifications. Employees obtain AI diplomas within ChatGPT itself; the goal is 10 million certified Americans by 2030. Walmart and Accenture were the first to jump on board, Accenture with more than 10,000 employees. The skill is measured, not claimed. And soon a model will match that measured skill against an open role.
That is the movement beneath the news. The product that made the headlines languished; the underlying habit kept growing.
The question that employers do not ask themselves
For employers, that changes the question completely. For years, the question was: how do I get to the top of Google, how do I drive traffic to my careers site. The question now is: what does the AI say about me at the moment someone is researching, without typing in my name.
That is a different playing field with different rules. On Google you win with a click. In an AI answer you win by being mentioned, even if a click never follows.
And that is where the shoe pinches. Most employers have no idea what a language model says about them. They see their vacancies, their campaigns, their figures on the careers site. They do not see the conversation in which a candidate asks ChatGPT which employer in their sector is a good fit, and in which three names are dropped that are not theirs.
What a model cannot read, does not exist for that model. An empty robots.txt, no schema, thin or slow pages: then you miss out on the answer, no matter how strong your brand otherwise is.
Two types of employers
As far as AI is concerned, there are only two types of employers left. Those who are readable and citable, and are structurally recommended. And those who are not, and simply never get a mention.
In the audits we carry out, we see just how big that gap is. In response to an exploratory query in a niche market, one employer scored 7.5% for visibility; his name simply did not feature in almost any part of the query. Once the basics were in place – technique, structure and quotable content – that figure rose to over 90%. The same brand, the same offering, a different result.
This is not a future scenario. It is happening now, every day, in conversations you never get to see. The candidate makes their choice before they arrive on your site, often before they know your name.
Take action rather than wait
Waiting for OpenAI's platform is therefore the wrong reflex. Whether that platform is there in six months or two years, the focus has already moved. What still seems like an experiment today will be ordinary in six months' time. What is still an advantage today will be the minimum in a year. And what still feels optional today will be the baseline in eighteen months.
Employers who realise this are already measuring what AI says about them. Which questions they turn up for and which they don't, on which platform, with what sentiment, and which competitor appears alongside them. They are structuring their site and content so that a model can read, understand and cite them. And they are doing this monthly, because AI answers shift.
There is also a second layer beneath that OpenAI news. If a model soon matches certificates and measured skills to employers, it will not just be about the candidate being discoverable. Then what counts is whether your organisation is described in the same language in which that match is made. An employer that talks vaguely about culture and core values gives a model nothing to match on. An employer that is concrete about roles, skills and context does. Employer branding thus also becomes a data question: is your story in a form that a model can pick up, or does it remain stuck in pretty sentences that land nowhere.
The job board that never came has made one thing painfully clear. You do not need a new platform to become invisible. That happens by itself, in the conversation the candidate is already having.
My name is Bryan Peereboom, founder of AI Rebels. We make employers visible in language models.