Four features, one concession
One of the biggest names in employer branding has just built a product to measure whether AI still mentions you.
On 6 August, Symphony Talent launched its AI Search Visibility Solutions. Four components: an index that measures how you appear in ChatGPT, Gemini, Claude and Perplexity; an accelerator that fine-tunes your careers site and job openings; an amplifier that benchmarks you against competitors; and a layer that tracks movement over time.
The product is nice. The caveat underneath is the real news.
When the big boys build a dashboard, the question is answered
When an established agency builds a dashboard for your visibility in language models, the question of whether this is a fringe phenomenon has been answered. For two years, AI visibility was something for the early birds. From now on, it is in the big boys' toolbox. Next quarter, your competitor will have a metric in their hands that you might not yet be doing.
This is how a category grows up. First, a few people call it by name. Then, someone builds a tool around it. Next, it becomes a line item on the quote. Symphony takes that third step for employer branding, and with that, your AI visibility becomes something you'll soon have to explain if you aren't measuring it.
The figures that explain this haste
Google now displays its AI Overviews above 43 per cent of all search queries. A year ago that was 15 per cent. Nearly tripled in twelve months. The candidate checking out an employer increasingly gets their answer right there on the screen, summarised by a model, without ever clicking on your page.
So Symphony measures it. Rightly so. Measuring is where you start, because you can't steer what you don't see.
A measurement tells you where you are, but not how to get there
That is the gap that every tool leaves open. An index, an accelerator, a benchmark: they point out the void to you. Filling it is something they do not do. That filling is editorial work, engineering and patience, and that does not appear on any screen as a button.
A language model builds its judgement of you from everything it finds about you: your job vacancies, your reviews, an old press release, a forum thread from three years ago. What is missing is filled in. What is unclear is simplified. And what is nowhere to be found does not exist.
An index shows that you do not appear under “best healthcare employer”. Good to know. The question that matters comes after that: why not, and what do you put where so that the model does pick you up.
The story that you no longer write yourself
I spent fifteen years in recruitment. The whole time was about control over the narrative: what the job description says, what the hiring manager says, what’s on the careers site. That narrative had an owner, and that owner was you. Now a model reads everything lying around loose about you and compresses it into a single paragraph. That paragraph is the first impression of your next candidate, and you didn't write it.
Take your employment conditions. If they appear differently on your jobs page than they do on Glassdoor, and differently again in an old press release, the model will pick one, or it will leave you out. If your culture is only in a vibe video and nowhere in readable text, the crawler reads nothing. Clarity beats volume, every time.
And it doesn't stop at your own site. A model only really believes you when others back you up. A consistent mention on a review site, a trade medium that names you, an employee who shares how things work on a forum: that carries more weight than the finest sentence on your careers page. You no longer control the conversation alone; you fuel it.
Two types of employers
For a language model, there are only two types of employer left. The employer with a clear, consistent narrative that it can retell, and which appears in the answer. And the employer whose information cannot be found anywhere, and which it skips for lack of a foothold. There is little room between the two.
A dashboard from any agency whatsoever won't change a thing as long as the underlying source is shaky. You can measure your invisibility to the nth degree and still remain invisible. The figures on the screen will only shift when you tackle the places where the model reads you: your own domain, your job vacancies, the reviews and professional forums where others talk about you, the structured data you use to explain yourself.
For employers, this means three things. What is still an advantage today will be the norm in six months. What is still just a loose experiment today will be a fixed rule in the budget in a year's time. And what you haven't made readable for a model today will soon cost you the candidates you never saw passing by.
What can you do this week?
What you can do this week is less work than it sounds. Ask ChatGPT, Perplexity and Google's AI answers who the best employers in your sector are, and check whether you are among them. Then put the answers to the real questions – salary, career progression, what the application process is like, what it is like to work there – in plain text on your own domain. And make sure your story says the same thing everywhere, because as soon as you contradict yourself, the model will take the easiest summary, and that is seldom yours.
Symphony's launch is good news. The market is now saying out loud what candidates have been doing for months. The conversation about you happens inside the model, and the only question that matters is whether you are in it.
My name is Bryan Peereboom, founder of AI Rebels. We make employers visible in language models.