How your employer brand shows up in AI search.
Candidates no longer open 10 tabs. They ask one assistant and read the answer as the truth. See how visible your employer brand is to ChatGPT, Claude, Perplexity and Gemini, and how to fix the gaps.
For 20 years candidate research meant Google, Glassdoor and your careers page. Now millions ask an AI assistant instead, and it replies in one paragraph. It names a few employers and skips the rest. You are in the answer, or you are invisible.
5 things decide whether an assistant can see you.
They compound. Miss the first and nothing else matters, because the model never reaches you. Get all 5 right and you become the answer rather than the omission.
Access
Can the AI crawlers actually reach your careers content, or is your robots file quietly turning them away.
Structure
Is your content marked up so a machine understands what it is reading, not just a human.
Knowledge graph
Does a clean, claimed entity for your organisation exist where the models look for facts.
Citation
Is your employer story corroborated across the wider web, not just asserted on your own site.
Social
Do the surrounding social and third party signals tell a model your brand is real and current.
These are the 5 dimensions The SetpointHQ Index scores, ordered here by dependency rather than by weight. In the published rubric, citation likelihood carries the heaviest weighting at 35%, followed by schema at 20% and the remaining 3 at 15% each. Read the methodology →
SEO was a game of ranking. AI visibility is a game of citation.
Under search optimisation, 10 results shared a page and you fought for position 3. Under AI, the model reads the web, decides who is worth naming, and writes one answer. Ranking 9th is invisible. Being named at all is the whole prize.
That changes what matters. A model cannot cite what it cannot reach, cannot understand content it cannot parse, and will not name a brand it has no corroborating evidence for. Visibility is earned across the 5 signals above at once, and most employers have never measured a single one of them.
Check it yourself in 5 minutes.
You don’t need a tool or a budget for a first read. Open ChatGPT, Claude, Gemini and Perplexity in turn, and ask the 4 questions candidates actually ask. Swap in your own company, sector, region and one named competitor.
- Broad. “What are the best companies to work for in [your sector] in [your region]?”
- Direct. “What is it like to work at [your company]?”
- Comparative. “Should I work at [your company] or [a named competitor]?”
- Practical. “What is the pay and progression like for [your role] at [your company]?”
Then read the answers against 4 outcomes. None is fatal and all are fixable, but only once you’ve seen them.
| What you see | What it means | What to fix |
|---|---|---|
| Named and accurate | You have real visibility | Nothing yet. Protect it and track it |
| Named but thin | The model knows you exist but can’t find good information | The depth of your content, and who else backs it up |
| Not named | Competitors appear and you don’t | Crawler access and your entity |
| Confused | You’re mixed up with another organisation of the same name | Claim and clean your knowledge graph entity |
One caveat. Answers change between sessions and between assistants, so a single run is a snapshot, not a verdict. It also won’t tell you which signal to fix first, or whether last month’s fix moved anything. That’s what the measurement is for.
Where ChatGPT gets its answer about you.
ChatGPT draws on 2 sources, and both have to work in your favour. What it learned in training decides whether you’re a known entity at all: a recognised organisation with a clean entity gets named with confidence, an unknown one gets a vague non-answer. What it retrieves live, when it searches the web, can bring current detail into the reply. But only if the crawlers can reach and parse your careers content.
Most employers miss for 4 dull, fixable reasons:
- Hidden content. Careers pages locked inside a job board widget the crawlers can’t see.
- No structure. No schema telling the model what your organisation and roles actually are.
- No entity. No claimed knowledge graph entity for the model to anchor to.
- Thin corroboration. Your story told on your own site and nowhere else.
None of this is about gaming the model. It’s about giving an honest account of your organisation in a form it can use.
Your careers site, as a crawler sees it.
If the crawlers can’t reach your careers content, it doesn’t exist as far as the model is concerned. Robots rules that block AI crawlers, content that only appears after JavaScript runs, jobs inside an embedded ATS widget: each one hides you. The fix is often a few lines of configuration. You just have to know to look.
Reaching the page isn’t understanding it, either. JobPosting schema on every live role, and Organisation schema for you as the employer, turn your content from something approximated into something cited correctly.
4 checks, run against the live site rather than the staging copy:
- GPTBot, ClaudeBot, PerplexityBot and Google-Extended are allowed in robots.
- Careers content is in the served HTML, not injected by a script.
- JobPosting and Organisation schema is present and passes the Rich Results Test.
- Your entity is claimed, and says the same thing your site does.
A fix you haven’t checked on the live site is a fix you don’t have. The free Careers Site Check runs those checks for you and comes back within 3 working days.
2 more ways in.
Generative engine optimisation for recruitment
The discipline that replaces SEO when candidates get one answer, not 10 links.
Read →Answer engine optimisation for recruitment
GEO gets you into the working material. AEO is about winning the answer itself.
Read →Recruiting agency, not an employer?
The 5 signals are identical. The buyer is not. An employer wants candidates to find them. An agency wants clients to find them, and a client asks a harder question: who fills roles like mine, in my sector, at my volume, and can that be checked. We cover that version separately.
Common questions
Is employer brand AI visibility the same as SEO?
Which AI assistants should we care about?
Can we control how AI describes our employer brand?
How is AI visibility measured?
We have a strong Glassdoor presence. Is that not enough?
We are a recruitment agency, not an employer. Does this apply to us?
Find out how visible your employer brand is to AI.
Our AI visibility audit measures your visibility across all 4 assistants, shows which signals are letting you down, and gives you a prioritised plan to fix them. Comparing providers first is sensible: here is how to choose an AI visibility audit, including the questions to ask us.