What AI Says About Working at Deloitte — And How Much It Gets Wrong
What AI Says About Working at Deloitte — And How Much It Gets Wrong
Deloitte is the UK's largest professional services firm. Over 22,000 employees. One of the most searched employers in the country. The kind of company every graduate considers — and every mid-career professional has an opinion about.
So when candidates ask AI about working at Deloitte, the answer matters. We tested it.
What we asked
We put three standard candidate queries to ChatGPT, Perplexity, and Google AI Overviews:
- "What's it like to work at Deloitte UK?"
- "What does Deloitte pay a senior consultant in London?"
- "What benefits does Deloitte UK offer?"
Then we compared the AI responses against Deloitte's published data, verified job listings, and current employee information.
Salary: off by £25K
The most striking gap was compensation.
ChatGPT estimated a Deloitte UK senior consultant earns £45,000–£55,000.
Actual range (from active Deloitte job listings and verified data): £65,000–£80,000 base, plus performance bonus.
That's a £25K understatement at the midpoint. A candidate reading ChatGPT's answer would conclude Deloitte pays below market for the Big 4 — when in reality their compensation is competitive.
Perplexity was slightly better at £50,000–£65,000, but still undershot by £15K. It cited a 2023 salary survey that hadn't been updated.
Google AI hedged with "salaries vary widely" and provided no specific range — arguably more honest, but unhelpful to the candidate trying to evaluate whether to apply.
Culture: a five-year-old snapshot
All three AI models described Deloitte's culture in terms that would have been accurate around 2021:
- "Traditional professional services culture with long hours"
- "Hierarchical structure with clear promotion pathways"
- "Expectation of in-office presence five days a week"
The reality in 2026: Deloitte has invested heavily in flexible working, introduced hybrid policies (typically 2-3 days in office), expanded wellbeing programmes, and restructured career pathways. The "traditional Big 4 grind" narrative is several years out of date.
The problem? AI models are assembling culture descriptions from Reddit threads, forum posts, and blog articles — many of which describe the pre-pandemic working environment. Without updated first-party data, AI defaults to the loudest voices, which tend to be older complaints.
Benefits: partly fabricated
This is where AI hallucination gets dangerous.
ChatGPT listed several benefits that Deloitte does offer — private medical, pension, cycle to work scheme — but also confidently stated:
- "Unlimited annual leave" — Not accurate. Deloitte UK offers generous but defined leave.
- "Sabbatical after 3 years" — Partially true. The eligibility criteria are different from what was stated.
- "Free gym membership" — Incorrect. They offer subsidised, not free, gym access.
These aren't catastrophic errors, but they create mismatched expectations. A candidate who joins expecting unlimited leave will feel misled — and blame Deloitte, not the AI.
Why Deloitte should care
Deloitte receives an estimated 500,000+ applications per year in the UK. Even if only 10% of candidates use AI to research the firm before applying, that's 50,000 people forming impressions from inaccurate data.
The downstream effects:
- Salary perception gap: Candidates who believe Deloitte pays £45K for senior consultants won't bother applying. They'll target firms that AI says pay more — even if Deloitte actually pays more in reality.
- Culture mismatch: Candidates expecting a rigid, traditional environment self-select out. The people Deloitte most wants to attract — those who value flexibility and modern working — never apply.
- Benefits confusion: Candidates who join with AI-inflated expectations are disappointed from day one. That disappointment feeds back into negative reviews, which AI then cites, creating a vicious cycle.
What Deloitte (and every Big 4 firm) can do
The fix is technical, not editorial:
1. Publish an llms.txt file
A structured file at deloitte.co.uk/llms.txt that explicitly states current salary ranges, benefits, culture description, and working arrangements. This gives AI models authoritative first-party data to cite.
2. Add structured data to careers pages
JSON-LD schema markup on job listings with accurate salary ranges, locations, and employment types. This is machine-readable data that AI models trust over Reddit threads.
3. Ensure careers content is crawlable
If AI crawlers are blocked via robots.txt, the only information available is whatever third parties have written. Allowing GPTBot, ClaudeBot, and PerplexityBot to access careers content ensures first-party information is in the mix.
4. Monitor regularly
AI models retrain. What they say changes. A quarterly check isn't enough — weekly monitoring catches inaccuracies before thousands of candidates see them.
The bigger picture
Deloitte is just one example. We've audited hundreds of UK employers and found the same patterns across the Big 4, FTSE 100 companies, NHS trusts, and fast-growing startups:
- 78% of AI salary estimates are inaccurate
- 91% of companies have no llms.txt file
- 67% have no structured data on their careers pages
- Average AI Visibility Score across UK employers: 34/100
The bar is genuinely low. The companies that take AI employer visibility seriously now will have a structural advantage in talent acquisition for years.
Check your own company
We built OpenRole specifically to solve this problem. Run a free audit and see exactly what AI tells candidates about your company — salary, culture, benefits, and more.
It takes 30 seconds. No signup required.
→ Run your free AI employer brand audit
Sources: OpenRole audit data (March 2026), Deloitte UK careers site (verified March 2026), ChatGPT-4o, Perplexity, Google AI Overviews.