Candidates no longer start their research on your careers page. They ask ChatGPT, Perplexity or Gemini what it is like to work for you, and they get a confident answer in seconds. That answer is now your first impression, and most HR teams have never read it.
That was the starting point for GEO: What It Means for Your Employer Brand, a myGwork panel on 12 May 2026 where Victoria Olsina joined talent and employer brand leads from Virgin Media O2, Lloyds Banking Group and Deutsche Bank. This post sets out the GEO side of that conversation: how AI systems form a view of an employer, why that view is hard to spin, and what HR, talent and marketing teams can do about it. For the fundamentals, start with her guide to generative engine optimisation.
What GEO means for HR and employer branding
Generative engine optimisation (GEO) for employer branding is the practice of shaping what AI assistants say when candidates ask about your organisation as a place to work. It covers which sources models retrieve, which facts they repeat, and whether your own content is clear enough to be quoted rather than replaced by third-party opinion.
Recruitment companies face the same problem twice: once for their own brand and again for every client they hire for. That is why Victoria Olsina runs a dedicated AI SEO and GEO service for recruitment companies.
Takeaway: your AI-generated reputation exists whether or not you manage it.
The questions the panel was built around
The session was planned around the questions employer brand teams are asking right now:
- When a candidate asks AI “What is it like to work at [company]?”, what does it say, and are employers paying attention?
- What trends, gaps and surprises show up in AI answers about employers?
- How do you measure the impact of AI search on talent attraction?
- How do inclusion and representation stories survive when candidates ask AI for inclusive places to work?
- What single action should teams prioritise in the next 90 days?
The rest of this post answers each one from the GEO side.
Why AI makes your employer reputation harder to spin
Victoria’s core point to the panel: it has never been this easy to check what a company is really like. One prompt gives a candidate a summary of reviews, forum threads, news coverage and employee posts, all read at once. No candidate could gather that much opinion that quickly before.
That shift moves power to candidates. A careers page and a recruitment campaign used to shape most of the story. Now an AI answer reflects whatever the wider web says, and as Victoria put it, the truth is the consensus. The more often something is said about you, the more likely a model is to repeat it as fact.
That leaves employers two jobs: do the work so the reality is good, and make sure the accurate story is published widely and consistently enough to become part of that consensus.
Takeaway: AI reports the consensus about you, not your campaign.
Ranking versus being cited
| Traditional search | AI search | |
|---|---|---|
| What the candidate sees | A list of links | One synthesised answer |
| What you compete for | A position on the page | Selection as a source |
| How you measure it | Clicks and impressions | Referral tracking plus prompt testing |
| How you lose | Sitting on page two | Being described inaccurately, or not at all |
Ranking first on Google does not guarantee an AI assistant will repeat your version of events. LLMs do not rank pages. They select sources.
Contradictions and stale facts
If your careers page says hybrid and a two-year-old article says office-first, the model either picks one or hedges. Old leadership names, outdated headcounts and past restructuring coverage surface the same way. Victoria’s rule for client brands applies directly: contradictions kill citations, consistency earns them.
How to measure AI visibility for your employer brand
Measurement was the question panellists found hardest, and it is still hard for most organisations.
AI tracking platforms are multiplying, and most SEO platforms now sell AI visibility tracking as a premium add-on. How accurate those numbers are is still debatable.
The cheapest starting point is already in your analytics suite. Create a dedicated channel in Google Analytics or Adobe Analytics that groups referrals from AI assistants such as ChatGPT, Perplexity, Gemini and Copilot. That shows you how many candidates click through from an AI answer to your careers site.
It has a clear limit. A candidate can read an AI answer, decide, and apply elsewhere without ever clicking. So pair the referral channel with a monthly prompt audit (step 1 below), which shows what candidates read even when they never visit.
Takeaway: track AI referrals now, and test prompts monthly to see the answers behind them.
Make sure your inclusion story shows up in AI answers
Candidates increasingly ask AI for inclusive employers: “Which banks are good employers for LGBTQ+ people?” Strong inclusion work gets lost if it lives in PDFs, internal comms or careers pages that AI crawlers cannot read.
What helps a model find and repeat your inclusion story:
- Specific, published facts. Name your employee resource groups, your policies and your benefits on crawlable pages, not just in brochures.
- Answer-first FAQs. A careers FAQ that answers “Is [company] an inclusive employer?” with facts in the first sentence.
- Independent recognition. Inclusion indexes, awards and partner platforms give models a source outside your own domain.
- Readable careers pages. Content buried in scripts or heavy code may never be read by AI crawlers at all.
Takeaway: if your inclusion work is not published in plain, crawlable language, AI cannot repeat it.
Where to start: five steps for the next 90 days
None of this needs a new budget line. It needs ownership and a monthly habit.
1. Audit what AI says about you today
Write 10 to 20 prompts a real candidate would use: “What is it like to work at [company]?”, “Does [company] offer remote working?”, “Is [company] a good employer for LGBTQ+ people?” Run them across ChatGPT, Perplexity, Gemini and Google AI Overviews. Record the answers and the sources each one cites, and repeat monthly.
2. Fix your entity
Use one consistent organisation name everywhere. Add Organisation schema to your site with sameAs links to your LinkedIn page, review profiles and other official accounts. If a model cannot confirm who you are, it fills the gap from whatever it finds.
3. Publish answers, not slogans
“We value our people” is unquotable. “Six months of fully paid parental leave for all parents” is quotable. The Princeton GEO study (Aggarwal et al., KDD 2024) found that adding statistics raised visibility in generative answers by around 41%, and citing sources lifted visibility for lower-ranked pages by 115%.
4. Build third-party evidence
Keep review profiles current and respond to reviews. Encourage employees to talk about their work on LinkedIn. Enter credible awards and appear on industry panels. Each creates an independent source a model can check outside your own domain.
5. Make it a shared job
Employer brand sits across HR, talent acquisition, marketing, communications and ERGs. Name one owner for the facts, one for the technical setup, and one for the monthly prompt audit and referral report.
Why regulated employers face a harder version
Banks, insurers and other regulated employers carry extra weight in AI answers. Regulatory news, fines and restructuring coverage tend to dominate what models retrieve, while the positive facts about working there sit on careers pages never written to be quoted.
The fix is the same five steps, applied with more discipline: tighter consistency across every official source, more specific published facts, and a prompt audit that includes the uncomfortable questions candidates actually ask.
Conclusion
Candidates are already asking AI about you, and AI answers with the consensus. You cannot buy your way into those answers, but you can make the accurate version the easiest one for a model to find, verify and repeat, then measure what reaches you.
The full panel, with perspectives from Virgin Media O2, Lloyds Banking Group and Deutsche Bank, is available to myGwork members on the event page. The entity and retrieval work behind it sits at the core of Victoria Olsina’s AI SEO and GEO service, and her free AI visibility checker shows whether AI systems can read your careers site.
Frequently Asked Questions
What is GEO in employer branding?
GEO in employer branding means shaping what AI assistants such as ChatGPT, Perplexity and Gemini say when candidates ask about your organisation as an employer. It focuses on consistent facts, quotable content and the third-party sources models use to build their answers.
How do I find out what ChatGPT says about my company as an employer?
Ask it directly, using the questions a candidate would type. Run the same prompts in Perplexity, Gemini and Google AI Overviews, record each answer and its cited sources, and repeat monthly to track changes.
How do I track candidates who come from ChatGPT?
Create a dedicated channel in Google Analytics or Adobe Analytics that groups referral traffic from AI assistants. It only captures clicks, so combine it with a monthly prompt audit to see the answers candidates read without clicking.
Can I correct inaccurate AI answers about my organisation?
Not directly. There is no edit button for AI answers. You correct them by fixing the sources: update your own pages, remove contradictions, refresh review profiles and publish specific, current facts that models can retrieve.
Who should own GEO for employer branding?
It works best as a shared responsibility. HR and talent teams own the facts, marketing or SEO teams own the technical structure and schema, and one named person owns the monthly prompt audit and AI referral reporting.
Want to know what AI says about your organisation?
Victoria Olsina helps brands become the trusted source ChatGPT, Perplexity and Gemini choose to cite. Book a free strategy session and she will show you where your brand stands in AI search today.
![Victoria Olsina, mentor at R[3]sidency x Construct](https://victoriaolsina.com/wp-content/uploads/2026/10/r3sidency-mentor-victoria-olsina-26239-300x158.png)
