26% of job seekers research employers with generative AI — twice the rate of review sites | a survey of 5,000 people in Japan
Itera, Inc. (head office: Tokyo; Representative Director: Takayuki Muto), which provides the AI search optimisation tool LLMOcheki, surveyed 5,000 company employees across Japan on how they gather information about employers during a job search.
Among the 880 respondents who had looked for a job in the past year, 26.02% said they had “asked a generative AI such as ChatGPT, Gemini or Copilot”. That is 2.08 times the share who used review sites such as OpenWork or Tenshoku Kaigi (12.50%).
Counting generative AI together with the AI summary at the top of Google search results (AI Overview), 314 people — 35.7% of those job hunting — encountered information about employers through AI. Where candidates form their view of a company is shifting.
Background
As generative AI spreads, the ways job seekers gather information about employers have multiplied. Yet there has been little large-scale data in Japan on how much job seekers actually use AI.
We therefore ran an online survey of 5,000 company employees nationwide to find out how information is gathered during a job search.
Key findings
- 26.02% of job seekers researched employers by asking a generative AI directly
- 21.59% read the AI summary at the top of Google search results (AI Overview)
- 314 people — 35.7% of job seekers — used generative AI or an AI summary
- Asking generative AI was 2.08× review sites (12.50%) and 2.83× social search (9.20%)
- The most researched topic via AI was “the company’s business and reputation” (59.24%)
- Next was “the going rate for salary and benefits” (53.82%), above half
Finding 1: How employer information is gathered (n=880, multiple answers)
We asked the 880 respondents who had job hunted in the past year how they researched the companies they applied to or considered.
| Method | Share |
|---|---|
| Job listings from job sites and agencies | 48.64% |
| The company’s official website | 34.32% |
| Asking a generative AI (ChatGPT, Gemini, Copilot, etc.) | 26.02% |
| The AI summary at the top of Google search (AI Overview) | 21.59% |
| Review sites (OpenWork, Tenshoku Kaigi, etc.) | 12.50% |
| Social media search | 9.20% |
Job listings from job sites and agencies came first (48.64%), but asking a generative AI (26.02%) approached the level of the company’s own website (34.32%).
Meanwhile the review sites long used for this purpose reached only 12.50% and social search 9.20% — both below generative AI.
Finding 2: What was researched with AI (n=314, multiple answers)
We asked the 314 respondents who used generative AI or an AI summary what they researched.
| Topic | Share |
|---|---|
| The company’s business and reputation | 59.24% |
| The going rate for salary and benefits | 53.82% |
| Help with writing documents | 40.45% |
“The company’s business and reputation” came first at 59.24% — above help with writing documents (40.45%), showing that AI is being used to evaluate the company itself.
“The going rate for salary and benefits” (53.82%) also exceeded half, indicating that candidates are checking pay levels with AI.
Finding 3: Job search activity (n=5,000)
880 people (17.6%) had looked for a job within the past year.
What the results suggest
Candidates start evaluating you before they apply
Recruitment has assumed a sequence: candidates find you on a job platform, then deepen their understanding on your careers site. This survey shows that a conversation with an AI now sits ahead of that.
35.7% of job seekers encounter information about employers through AI, and 59.24% of them research “business and reputation”. The impression formed at that stage plausibly shapes whether they apply at all.
Whatever you publish becomes the answer
The material an AI draws on when answering questions about a company is whatever exists on the web: the official site, job platforms, press releases, reviews. The less official information a company publishes, the more the AI assembles its answer from third-party sources.
The 53.82% researching “the going rate for salary and benefits” is worth noting too. If you do not publish your own salary bands and policies, an AI may answer with industry averages or inference.
What companies should do
We suggest four responses:
- Check how AI currently describes your company
- Publish salary and benefits as figures (stating the population covered and the period)
- Supply the material on your business and reputation yourself, in concrete terms
- Keep the information consistent across job platforms, your careers site and third-party media
About the survey
| Item | Detail |
|---|---|
| Respondents | 5,000 company employees across Japan (880 who job hunted within the past year were extracted) |
| Period | July 2026 |
| Method | Online survey |
About LLMOcheki
LLMOcheki measures and improves how a brand or site is mentioned and cited within the answers generated by AI such as ChatGPT, Gemini, Perplexity and Google AI Overview.
It measures mention and citation rates daily across eight AI engines, and brings competitor comparison, analysis of citation sources and detection of incorrect information into a single place.
- Launched: 14 July 2026
- Pricing: from ¥29,800 per month (excluding tax), no setup fee
- Free audit: enter a URL to check how a site stands in AI search (no registration required)
LLMOcheki https://llmocheki.com/
Originally published at prtimes.jp