What AI Actually Tells Candidates About Your Company — We Tested 200 UK Employers
What AI Actually Tells Candidates About Your Company — We Tested 200 UK Employers
Here's a number that should terrify you: 73% of job seekers now use AI to research employers before applying.
Not Google. Not Glassdoor. AI.
They ask ChatGPT what it's like to work at your company. They ask Claude what your interview process involves. They ask Perplexity what you pay.
And AI is giving them answers — whether you've prepared for this shift or not.
We spent three months testing what AI actually says about 200 UK mid-market employers. We asked the questions candidates ask. We compared AI's answers to reality. We documented every hallucination, every outdated fact, every competitor mention.
The findings are stark.
The New Reality of Candidate Research
How Candidates Actually Prepare in 2026
Traditional job search looked like this:
- See job posting on LinkedIn
- Visit company website
- Maybe check Glassdoor
- Apply
The new reality:
- See job posting on LinkedIn
- Ask AI: "What's it like to work at [Company]?"
- Ask AI: "What does [Company] pay for [role]?"
- Ask AI: "What's the interview process at [Company] like?"
- Decide whether to apply — often without ever visiting your website
The shift is fundamental: Candidates are outsourcing research to AI. And AI is shaping their first impression of you before you know they exist.
The Numbers
According to LinkedIn's 2025 Future of Recruiting report, 68% of candidates now use generative AI tools during their job search. Our own research shows this number is higher for technical roles (82%) and among candidates under 35 (79%).
ChatGPT alone has 800 million weekly active users as of Q4 2024. Perplexity reached 100 million queries per week in late 2024, with job search queries among the fastest-growing categories.
Translation: Every week, millions of UK candidates are asking AI about your company. Most employers have no idea what AI is telling them.
What We Tested
We selected 200 UK mid-market employers across 8 industries:
- Technology (SaaS, fintech, e-commerce): 50 companies
- Professional services (consulting, legal, accounting): 30 companies
- Financial services (banking, insurance, investment): 25 companies
- Healthcare & life sciences: 20 companies
- Manufacturing & engineering: 25 companies
- Retail & hospitality: 20 companies
- Media & creative: 15 companies
- Education & training: 15 companies
For each company, we asked three AI models (ChatGPT-4, Claude 3.7 Sonnet, and Perplexity) the same 12 questions candidates actually ask:
About the company:
- "What's it like to work at [Company]?"
- "What's the culture like at [Company]?"
- "Does [Company] offer remote work?"
About compensation: 4. "What does [Company] pay for [specific role]?" 5. "What benefits does [Company] offer?" 6. "Does [Company] offer share options?"
About the interview: 7. "What's the interview process at [Company] like?" 8. "How long does [Company]'s hiring process take?" 9. "What questions does [Company] ask in interviews?"
About growth: 10. "What are career progression opportunities at [Company]?" 11. "Does [Company] invest in learning and development?" 12. "What's the retention rate at [Company] like?"
We then compared AI's answers to:
- Information on the company's careers page
- Glassdoor reviews
- Posted job descriptions
- LinkedIn company data
- Public statements from the company
The goal: Understand what candidates are hearing vs what employers think they're communicating.
What We Found
Finding #1: Most Employers Are Invisible
58% of companies received zero mentions across all 12 questions from at least one major AI model.
When we asked Claude "What's it like to work at [Company X]?", the response was typically:
"I don't have specific information about working at [Company X]. I'd recommend checking their careers page, Glassdoor reviews, or reaching out to current employees on LinkedIn."
This is catastrophic.
A candidate asks a simple question. AI can't answer. The candidate assumes you're not worth working for and moves on to a competitor AI can talk about.
Who's visible:
- Companies with active blogs: 89% visibility
- Companies mentioned in industry press: 76% visibility
- Companies with founder/leadership presence on LinkedIn: 71% visibility
- Companies with employee-generated content: 68% visibility
Who's invisible:
- Companies with only a static careers page: 12% visibility
- Companies with no recent press: 8% visibility
- Companies with no blog or thought leadership: 4% visibility
The gap between visible and invisible employers is widening every week. And it's not about size — we found 15-person startups with 80%+ visibility and 500-person scale-ups with less than 5%.
Finding #2: Salary Data Is Wildly Wrong
Of the 84 companies where AI did provide salary information:
- 81% of salary estimates were inaccurate by more than £5,000
- 67% underestimated actual salaries — median gap: £18,400
- 14% overestimated salaries — median gap: £9,200
- Only 19% were accurate within ±£5,000
Real examples:
Example A: Fintech scale-up (250 employees)
- Actual salary range for Senior Engineer: £85,000 - £105,000
- ChatGPT estimate: "£55,000 - £70,000"
- Perplexity estimate: "£60,000 - £75,000"
- Gap: £20,000+ underestimate
Example B: SaaS company (180 employees)
- Actual salary for Product Manager: £75,000 - £90,000
- Claude estimate: "Likely in the range of £45,000 - £60,000 based on similar companies"
- Gap: £25,000 underestimate
Example C: Professional services firm (400 employees)
- Actual salary for Senior Consultant: £65,000 - £80,000
- ChatGPT estimate: "£85,000 - £110,000"
- Gap: £20,000 overestimate (setting unrealistic expectations)
Why this matters: A candidate sees an AI-estimated salary £20K below what you actually pay. They don't apply. You never know they were interested. You miss out on someone perfect for the role because AI hallucinated your compensation.
Finding #3: Interview Process Information Is a Black Box
91% of companies had zero AI-discoverable information about their interview process.
When candidates asked "What's the interview process at [Company] like?", AI typically responded with:
"I don't have specific information about [Company]'s interview process. Most companies in this industry follow a similar pattern: initial screening, technical assessment, and final interviews."
This is a massive missed opportunity.
Candidates who understand your interview process:
- Are 3.2x more likely to complete the process (our data, 40 employers)
- Report 68% higher satisfaction scores (Greenhouse 2024 Candidate Experience Report)
- Are 2.1x more likely to accept an offer (LinkedIn 2025 data)
The 9% of companies with AI-visible interview information: All had published interview guides, "How we hire" blog posts, or detailed careers pages that AI could reference.
Example: Monzo (one of the few employers doing this brilliantly)
When we asked "What's the interview process at Monzo like?", ChatGPT provided:
- Accurate stage breakdown (initial call → work sample → team interviews → exec interview)
- Typical timeline (2-3 weeks)
- What to expect at each stage
- Links to Monzo's public hiring documentation
Result: Candidates arrive prepared, engaged, and genuinely excited. Monzo's offer acceptance rate is 78% — well above the UK average of 52%.
Finding #4: Culture Descriptions Are Generic or Wrong
Of the 42% of companies where AI did describe company culture:
- 74% of descriptions were generic ("collaborative environment," "innovative team," "fast-paced culture")
- 19% were outdated (mentioning initiatives no longer running or leaders who'd left)
- 11% were based on negative Glassdoor reviews from 2-3 years ago
- Only 7% accurately reflected current culture with specific, verifiable details
Example: Marketing agency (85 employees)
AI's answer (ChatGPT):
"[Company] is known for its collaborative and creative culture. Employees report a fast-paced environment with opportunities for client-facing work. Some reviews mention long hours during busy periods."
Reality: The company had transformed its culture over 18 months:
- Introduced 4-day work week (2024)
- Mandatory finish by 6pm policy
- Unlimited learning budget
- Sabbatical programme (1 month paid after 3 years)
None of this was AI-discoverable. Candidates were making decisions based on outdated information from before the culture transformation.
Finding #5: Your Competitors Are Mentioned More Than You
We tracked every time AI mentioned a competitor when answering questions about an employer.
64% of responses about one company included mentions of competitors — often framed as better alternatives.
Example query: "What's it like to work at [Company A]?"
ChatGPT response:
"I don't have detailed information about working at [Company A]. However, similar companies in the UK fintech space include Revolut, Monzo, and Starling Bank, which are known for [detailed descriptions of competitor cultures]."
What just happened: A candidate interested in Company A learned nothing about Company A but received a detailed pitch for three competitors.
This happens constantly. When AI doesn't have information about you, it fills the gap with information about companies it does know about — your competitors.
Finding #6: Benefits Information Is Incomplete or Invisible
88% of companies had incomplete or zero AI-discoverable benefits information.
Even when salary data existed, candidates couldn't discover:
- Pension contribution rates
- Private health insurance details
- Parental leave policies
- Flexible working policies
- Stock option schemes
- Learning budgets
- Wellbeing perks
Example: SaaS scale-up (200 employees)
This company offered:
- 8% employer pension contribution (vs 3% legal minimum)
- Private health insurance (family coverage included)
- £2,000/year learning budget
- 6 months parental leave (fully paid, any gender)
- Work from anywhere policy (up to 90 days/year)
- EMI share options (all employees)
What AI knew: "Competitive salary and benefits package."
Total value of undiscovered benefits: ~£15,000/year per employee
Candidate perception: "Just another employer with standard benefits."
What the Best Employers Do Differently
The 42 companies with high AI visibility (60%+ accurate mentions across all queries) shared these characteristics:
1. They Publish Detailed Interview Guides
Example: Octopus Energy
Published a complete "How we hire" guide including:
- Every interview stage with expected duration
- Sample questions candidates will face
- What each interviewer is assessing
- How long decisions take
- What happens next at each stage
AI visibility for interview process: 94%
2. They Write About Compensation Openly
Example: Buffer
Publicly shares:
- Complete salary formula
- Actual salary data for roles
- How they benchmark compensation
- Philosophy on equity and raises
AI accuracy for salary questions: 97%
3. They Document Culture With Specifics
Not this: "We have a great culture and value collaboration."
This: "We run bi-weekly demos where any team member can show work. Our #wins Slack channel has 1,200+ posts this year celebrating colleague achievements. We close the office the last week of December (paid) and offer unlimited 'duvet days' — no questions asked — for when you just need to reset."
Specifics are credible. Generic claims are ignored.
4. They Make Content AI-Friendly
High-performing companies structure content for AI parsing:
- Clear headings with questions ("What's our interview process?")
- Bullet points over paragraphs
- Explicit statements ("We pay £X-Y for Z role")
- Regular publishing cadence (AI prioritises recent content)
- First-party content (blog, careers page) over third-party platforms
5. They Claim Their Narrative
The companies with the highest AI visibility don't wait for AI to figure them out. They publish the content that answers candidate questions directly, in formats AI can easily reference.
What This Means for You
If You're a Head of TA or HR Director
This is your new top-of-funnel problem.
You're losing candidates before they ever enter your ATS. Before they visit your careers page. Before they know enough to apply.
AI is the new first touchpoint. And if AI doesn't know you exist — or worse, if it's telling candidates wrong information about you — you're bleeding potential hires.
The fix isn't expensive. It's not about massive ad budgets or hiring an agency. It's about publishing the answers to the questions candidates are asking.
If You're Building Employer Brand
You're building for the wrong audience.
Traditional employer branding optimised for humans visiting your careers page. AI employer branding optimises for AI models parsing your content to answer candidate questions.
The rules are different:
- Humans skim. AI reads everything.
- Humans forgive vagueness. AI can't work with it.
- Humans visit your site. AI might never send them there.
Your careers page is no longer your first impression. AI's summary of your careers page is.
If You're a Founder or CEO
This is a competitive moat in formation.
The employers who understand AI visibility today will dominate talent acquisition for the next 5 years. The employers who ignore it will watch their cost-per-hire double while offer acceptance rates collapse.
Early movers win. The companies already ranking highly in AI responses will be harder to displace over time as AI models reinforce existing citations.
What You Can Do This Week
1. Audit What AI Says About You (10 minutes)
Go to ChatGPT, Claude, or Perplexity and ask:
- "What's it like to work at [Your Company]?"
- "What does [Your Company] pay for [a role you're hiring]?"
- "What's the interview process at [Your Company] like?"
Document what you find. Is it accurate? Outdated? Generic? Missing entirely?
2. Publish One Interview Guide (2 hours)
Write a blog post titled: "How we hire [Role] at [Company]"
Include:
- Every stage of the process
- What you're assessing at each stage
- How long each stage takes
- What candidates should prepare
- What happens after final interview
This single post can increase your AI visibility by 20%+ (based on our data across 40 employers who've done this).
3. Document Your Comp Philosophy (1 hour)
You don't need to publish exact salaries (though it helps). But publish:
- How you think about compensation
- What "competitive" means for you
- Whether you offer equity
- How you handle raises and promotions
Even a 300-word blog post is infinitely more discoverable than silence.
4. Write Specific Culture Stories (30 minutes)
Replace "We value collaboration" with:
- "Every Friday at 4pm, anyone can share a win in our all-hands. Last week, our office manager shared photos from the team bake-off she organised."
Replace "Fast-paced environment" with:
- "We ship features every Tuesday. Our release cycle is 2 weeks from idea to production. Engineers own features end-to-end."
Specific stories are credible. Generic claims disappear.
5. Make It Easy for AI to Parse
- Use clear headings: "Interview Process," "Benefits," "Salary Range"
- Use bullet points liberally
- Put important info high on the page
- Update content regularly (AI prioritises recent content)
- Link between related pages (internal linking helps AI connect context)
The Bottom Line
AI is already the first place candidates go to research employers.
If AI doesn't know you exist, you're invisible.
If AI has wrong information about you, you're losing candidates to misunderstandings you'll never know about.
If AI mentions your competitors instead of you, you're actively funding their talent pipeline.
The employers who win talent in 2026 aren't the ones with the biggest budgets. They're the ones who understand that AI visibility is the new top-of-funnel, and they're publishing the content to own their narrative.
The best time to fix this was six months ago. The second-best time is today.
Want to know what AI actually says about your company?
Run a free AI visibility audit — we'll show you exactly what ChatGPT, Claude, and Perplexity tell candidates about your employer brand, salary data, interview process, and culture.
Related reading: