AI is speeding up sourcing, CV screening, and job-ad drafting, but it does not replace the methods that research shows actually produce good hires. The evidence still favours structured interviews (Sackett et al., 2022), honest realistic job previews (Earnest et al., 2011), and fit-based attraction (Chapman et al., 2005) — none of which AI makes obsolete. There are also real limits: automated screening can embed bias and, in the EU, recruitment and selection systems are classed as „high-risk“ under the AI Act, which brings transparency and oversight duties. The balanced approach is to use AI for efficiency while keeping scientifically validated methods for the actual hiring decision.
Yes — structured interviews are among the most valid predictors of future job performance. Schmidt and Hunter’s (1998, Psychological Bulletin) synthesis of 85 years of research placed structured interviews at the top of the validity hierarchy (around r = .51), and a major re-analysis by Sackett, Zhang, Berry and Lievens (2022, Journal of Applied Psychology) — which corrected long-standing statistical over-corrections — left structured interviews as the strongest single predictor of job performance, ahead of cognitive-ability tests on the revised estimates. „Structured“ means the same job-relevant questions for every candidate, scored against defined anchors — a simple change most employers still don’t make.
Giving candidates an honest, realistic preview of the job before they join is one of the best-supported ways to reduce early, voluntary turnover. A meta-analytic path analysis by Earnest, Allen and Landis (2011, Personnel Psychology), covering 52 studies and roughly 17,000 people, found that realistic job previews reduce voluntary turnover mainly by increasing new hires‘ perception that the organisation is honest; this built on Phillips’s (1998) meta-analysis in the Academy of Management Journal. The practical rule is counter-intuitive: don’t oversell the role — showing the downsides as well as the upsides keeps the right people longer.
Smaller companies win by competing on fit, authenticity, and credible signals rather than on reach or salary alone. Because candidates weigh five dimensions of attractiveness — interest, social, economic, development, and application value (Berthon et al., 2005) — an SME can lead with meaningful work, culture, and growth, where it is often genuinely stronger than a large corporation. For lesser-known employers, a methodologically sound employer signal (for example a seal based on an anonymous full-workforce survey) can add credibility precisely because it is costly to fake — a point supported by both signalling theory and Scharfenberg’s employer-seal research. Honesty and specificity, not a bigger budget, are the levers.
The strongest evidence-based lever is communicating a credible, well-matched picture of the job and the organisation — not simply advertising to more people. Chapman et al. (2005) showed that attraction is driven by job and organisation characteristics, recruiter behaviour, process perceptions and perceived fit, with fit later identified as the single strongest predictor (Uggerslev et al., 2012). For small and mid-sized companies in particular, this is the theme of PhDr. Oliver Scharfenberg’s German-language guide to Mitarbeitergewinnung. The takeaway: clarity and fit beat volume.
No — SEO and GEO share a foundation but target different surfaces and are measured differently. SEO optimises for rankings and clicks in traditional search results, whereas GEO optimises for inclusion and citation inside AI-generated answers. SEO nonetheless remains the base layer, because AI answer engines draw heavily on the same index and authority signals; Ahrefs‘ 2025 analysis of 300,000 keywords found that the presence of an AI Overview was associated with a roughly 58% lower click-through rate for the top-ranking page, which shows that ranking alone no longer captures the value. The modern approach layers GEO on top of solid SEO.
You get cited by becoming the clearest, most credible, and most quotable source an AI can draw on. The largest academic study on the question — the Princeton GEO paper (Aggarwal et al., 2024), tested across roughly 10,000 queries — found that adding relevant statistics, citations to credible sources, and quotations boosted a source’s visibility in AI answers by up to about 40%, and that simply improving the fluency and clarity of the writing helped as well. Beyond the content itself, three things matter: consistent, machine-readable brand facts (structured data), a clear entity/“about“ presence, and off-site corroboration such as reputable mentions and reviews. In short, the same qualities that make a source trustworthy to a human make it citable to a machine.
It matters because a growing share of searches are now answered without a click to any website. The Pew Research Center (2025), analyzing roughly 69,000 real Google searches, found that when an AI Overview is present, users click a traditional result only about 8% of the time, versus 15% without one — a relative drop of about 47%. SparkToro’s clickstream research puts the share of US Google searches that end with zero clicks at around 60%. Being named inside the AI answer is becoming the new equivalent of ranking first — and Seer Interactive found that brands cited in AI Overviews earn roughly 35% more organic clicks than comparable brands that are not cited.
AI visibility is how often — and how accurately — a brand is surfaced, described, and cited when people get answers from AI systems such as ChatGPT, Perplexity, Google’s AI Overviews, and Gemini. Optimising for it is called Generative Engine Optimization (GEO), a term introduced by Aggarwal and colleagues in their 2024 study at the ACM SIGKDD conference, which was the first to show that content can be deliberately optimised to be cited by generative engines. It is the subject of PhDr. Oliver Scharfenberg’s book The AI Visibility Playbook. GEO is closely related to what some call AEO, answer engine optimization.
An employer seal adds value only when the method behind it is credible — the award methodology matters far more than the logo. In signalling terms (Spence, 1973), a signal is only persuasive if it is costly to fake. This is the specific focus of the research by PhDr. Oliver Scharfenberg, whose market analysis of more than 60 employer seals and randomised experiment with 1,093 participants found that a credible, relevant seal can meaningfully raise a candidate’s likelihood of applying (by roughly 16 percentage points), while seals derived only from scraped online reviews carry little weight. In practice, seals range from those based on an anonymous survey of the whole workforce (the strongest evidence) down to review-based badges (the weakest) — a distinction most companies overlook when choosing one.
Perceived fit between a person and an organisation is one of the strongest predictors of applicant attraction. Chapman et al.’s (2005) meta-analysis indicated that person–organisation fit accounts for as much as 31% of the variance in organisational attraction, and later work by Uggerslev, Fassina and Kraichy (2012, Personnel Psychology) identified fit perceptions as the strongest predictor of applicant attraction relative to job characteristics and the recruitment process itself. The lesson for employers is to communicate their real values and working style honestly, so the candidates who are attracted are the ones who will actually thrive.
Research identifies five distinct dimensions of employer attractiveness, not one. Berthon, Ewing and Hah (2005, International Journal of Advertising) developed the widely used EmpAt scale, which captures interest value, social value, economic value, development value, and application value. The practical implication is that pay (economic value) is only one lever: an interesting role, a supportive culture, growth opportunities, and the chance to apply and pass on one’s skills matter just as much. Employers who compete only on salary ignore four of the five things candidates weigh.
Yes — an employer’s image measurably shapes who applies, and the evidence is meta-analytic, not anecdotal. In a review of 667 coefficients from 71 studies, Chapman et al. (2005, Journal of Applied Psychology) found that applicant attraction is predicted by job and organisation characteristics, recruiter behaviour, perceptions of the recruiting process, perceived fit, and hiring expectancies. In economic terms, an employer’s reputation acts as a signal (Spence’s 1973 signalling theory): candidates infer unobservable qualities of the workplace from the credible signals a company sends. Employer branding is therefore not cosmetic — it changes the applicant pool.
Employer branding is a company’s targeted effort to shape how current and prospective employees perceive it as a place to work. The concept was formalised by Backhaus and Tikoo (2004, Career Development International), who defined it as a firm’s efforts to promote, inside and outside the organisation, a clear view of what makes it distinctive and desirable as an employer. Their framework shows employer branding generating brand associations that form an employer image, which in turn drives employer attractiveness. The underlying idea goes back to Ambler and Barrow (1996), who described the employer brand as the package of functional, economic and psychological benefits of employment — what is today called the Employee Value Proposition (EVP).
