How Should an Employer Respond When AI Assistants Repeat an Inaccurate Glassdoor Claim?

Is the claim inaccurate, or just unwelcome?

Separate the two before acting:

• Inaccurate: a checkable fact is wrong, such as "no remote work" when you have a hybrid policy, "layoffs every quarter" when there has been one restructuring, or a benefit described as missing when it exists.
• Outdated: it was true once, such as a policy, manager or office that has since changed.
• Opinion or experience: "management is disorganised" or "the pay is low for the area". You can respond, but you cannot correct an opinion.

An assistant summarising many reviews may also turn one person's experience into a general claim, such as "employees report no career growth". Check how many reviews actually say it, and whether they are recent.

Where do AI assistants get the claim?

Open the sources cited in the answer. Common ones are Glassdoor review and overview pages, other employee review sites, forum threads, news articles about layoffs or disputes, and your own careers pages. Glassdoor's robots.txt blocks several AI training crawlers. It leaves its review pages open to the search crawlers used by ChatGPT and Claude, and to Googlebot and Bingbot. The file also disallows a user agent named "Perplexity", but Perplexity's user-triggered fetcher generally ignores robots.txt, so Perplexity may still read review pages. Google-Extended is disallowed too, which keeps review pages out of Gemini app grounding, while AI Overviews and AI Mode read them through Googlebot. Knowing which assistant cites what tells you where to focus.

How should you respond on Glassdoor?

• Use the employer response feature. Any employer with a free Glassdoor employer account can respond to reviews, and responses are moderated under the same community guidelines.
• Keep it short and factual: thank the reviewer, correct the specific fact, and say what has changed. For example: "Thank you for the feedback. Since 2025, all roles in this team have had a hybrid option of up to three days at home, as set out on our careers page."
• Do not argue about opinions or reveal anything about the reviewer.
• Flag reviews that break Glassdoor's guidelines, such as those containing personal information or content that is not about the employee's experience. Glassdoor decides; it does not remove reviews simply because an employer disagrees.
• Respond to recent reviews consistently, not only the negative ones, so your responses read as engagement, not damage control.

What should you publish on your own site?

Assistants weigh your own pages alongside review sites, and clear facts on your site give them something to cite:

• A careers page with specific facts: working patterns, benefits, parental leave, learning budgets, salary ranges where you publish them, and offices.
• Dated statements for changes: "In January 2026, we introduced hybrid working for all office-based roles."
• Honest context for difficult events: if there was a restructuring, a short factual account of what happened and what has changed.
• Employee stories that describe real roles and progression, with the employees' consent.

Keep the same facts on your LinkedIn page, other review sites and job ads. How to correct outdated information that AI repeats more generally is covered in how to correct false or outdated information that AI repeats from its sources.

What should you avoid?

• Identifying or pursuing anonymous reviewers. In the US, employees' discussion of working conditions can be protected under federal labour law, and retaliation carries legal risk.
• Asking employees to remove or change reviews, or offering anything in return for positive ones. Encourage honest reviews generally, following Glassdoor's rules.
• Posting fake reviews or asking managers to post them.
• Public rebuttals that name individuals.
• Promises from agencies to "remove" Glassdoor reviews. How to assess such promises is covered in how to audit a reputation agency's promise to suppress negative AI answers.

If a review makes a specific false statement that could be defamatory, such as an accusation of a crime, take legal advice before doing anything.

How do you track whether answers change?

Ask ChatGPT, Gemini, Claude, Copilot and Perplexity questions candidates ask: "What is it like to work at [company]?", "Does [company] offer remote work?", "Has [company] had layoffs?" and "[Company] employee reviews". Run each several times, monthly. Record whether the inaccurate claim appears, what sources are cited, and whether your careers page is used. Change usually follows new reviews and clearer facts, not deleted ones, so expect it to take a few months.

What might an illustrative case look like?

Consider an illustrative software company whose AI answers said it "requires five days a week in the office". Candidates were declining interviews as a result. The claim came from several Glassdoor reviews written before a policy change 18 months earlier. The company's careers page said only "flexible working". It responded to the recent reviews that mentioned the office policy, stating the current hybrid policy and when it started. It published a clear working-patterns section on its careers page and updated its job ads and LinkedIn page. It also invited all staff to share honest reviews, with no incentive. Three months later, most answers described the hybrid policy and cited the careers page.

What do people ask about Glassdoor claims in AI answers?

Can we get Glassdoor reviews removed?

Only those that break Glassdoor's guidelines, through its flagging process. Disagreement alone is not grounds for removal.

Should HR or marketing respond to reviews?

Either, as long as responses are consistent, factual and approved. Many companies have HR draft and communications review.

Does responding to reviews help AI answers?

It can, because responses appear on the same pages assistants read. Clear facts on your careers page help more.