Amazon enters recruiting.
Amazon Connect Talent has just launched. It can conduct AI voice interviews around the clock, score every candidate, and provide recruiters with a dashboard of results. Here is what it actually means for HR – and what it gets dangerously wrong.
Amazon does not enter markets quietly. When AWS announced Connect Talent in April 2026—an agentic AI platform that conducts structured voice interviews, scores candidates, and delivers hiring recommendations without human involvement – it sent a clear signal: AI is no longer merely assisting recruiters; it is replacing parts of the process entirely.
The headlines were predictably enthusiastic. The product is technically impressive. However, before HR leaders rush to adopt it, there are questions worth asking – about what it actually measures, what it overlooks, whether it is legal to use in Europe, and whether Amazon is truly the hiring role model anyone wants to follow.
What Amazon Connect Talent actually does
Amazon Connect Talent uses AI agents to conduct structured voice interviews, administer science-backed assessments, and score candidates consistently, allowing recruiters to focus on strategic decisions. Candidates can interview 24/7 from any device. The platform includes adaptive questioning, a mobile-first candidate portal, ATS integrations, and a recruiter dashboard with transcripts and evaluations.
Starting with an existing job description, AI agents analyse the role requirements and generate a complete interview plan – identifying key competencies, creating structured questions, and building evaluation criteria. Once approved, the system automatically invites candidates to interview at their convenience.
Informed by decades of Amazon’s hiring science, Amazon Connect Talent provides transparency for every assessment, interview, and candidate score, enabling recruiters to retain control over final hiring decisions. During the preview, the platform supports English, Portuguese, French, Italian, German, and Spanish.
On paper, it sounds compelling, particularly for high-volume hiring environments. AWS is explicit that the target is high-volume hiring: seasonal retail, logistics, healthcare staffing, and hospitality. Amazon hired approximately 250,000 seasonal workers in 2025. The product is, in some sense, Amazon packaging its own peak-season playbook.
What it means for HR
Amazon’s new system is designed to conduct voice interviews, assess responses, and produce hiring recommendations independently, rather than merely assisting with CV screening or interview scheduling as traditional recruiting software does. AI is no longer just acting as a productivity assistant; it is increasingly an operational layer that can run workflows, integrate decisions, and manage candidate interactions on an ongoing basis.
For talent acquisition leaders managing hundreds of open roles, the efficiency argument is compelling. Scheduling alone consumes enormous recruiter bandwidth. A system that conducts interviews around the clock, reduces time-to-hire, and delivers consistent scoring across all candidates addresses genuine pain points.
However, efficiency is not the same as quality. Speed is not the same as fit. The most expensive hiring mistake is not the slow hire – it is the fast hire who leaves after six months.
“It’s very powerful, because if you’re recruiting for a high-turnover role — like a truck driver, warehouse worker, any sort of tactical work — you’re going to get a lot of applicants and need time to schedule interviews. Hiring managers don’t have a lot of time to do it, either. So, it serves as a very good screening tool… it’s a way to take all of the intellectual property that companies use for interviewing and embed it in a much more scalable experience.”
Bersin’s framing is telling: the use case is high-volume, high-turnover hiring. This raises a fundamental question for white-collar HR leaders – is that the problem you are trying to solve?
The dangers nobody is discussing
Amazon’s own track record
The 2014–2018 AI failure
In 2014, Amazon’s engineers started building an internal AI recruiting tool. By 2018, Reuters reported the company had quietly scrapped it after discovering it had taught itself to penalise resumes containing the word “women’s,” downgraded graduates of certain women’s colleges, and favoured verbs statistically more common on male engineers’ resumes. Connect Talent claims to have solved this. The burden of proof is on them.
Amazon’s turnover runs at 150% annually
Amazon’s warehouse turnover runs around 150% annually. Not only that, Fortune reported in July 2025 on a performance review system that employees described as deliberately opaque and designed to move people out.
If the “hiring science” underpinning Connect Talent is Amazon’s own, organisations should ask: optimised for what outcome, exactly?
Candidates are already rejecting it
Fortune documented widespread hostility to AI interviewers in August 2025, with job seekers saying they’d rather risk staying unemployed than go through another AI screening. Connect Talent is launching into a candidate headwind with a voice agent its own team acknowledges needs more work.
Employer brand damage is a real and underestimated cost.
Competency scoring ≠ personality fit
Voice interviews can test structured competency questions. They cannot reliably measure self-efficacy, resilience, optimism, or risk tendencies — the dimensions that decades of research show predict performance and retention far more accurately than any interview, structured or otherwise.
A faster, more consistent bad signal is still a bad signal.
Is it even legal in Europe?
This is where the conversation becomes critical for any European HR leader considering Connect Talent.
The EU AI Act’s most critical compliance deadline for most enterprises is 2 December, 2027 (originally, on August 2, 2026), when requirements for Annex III high-risk AI systems become enforceable. This includes AI used in employment, credit decisions, education, and law enforcement contexts.
The regulation’s extra-territorial reach mirrors the GDPR. Any organisation, regardless of location, must comply if its AI systems are used within the EU or produce outputs that affect EU residents. A US-based company using AI for hiring that serves European customers falls within scope, even if the AI models run on servers outside Europe.
What does this mean in practice? Under the EU AI Act, AI systems used for employment screening are classified as high-risk. They require human oversight mechanisms, transparent documentation, bias auditing, explainability of outputs, and a Fundamental Rights Impact Assessment before deployment.
Non-compliance exposes both the provider and the deploying organisation to administrative fines of up to €35 million or 7% of worldwide annual turnover, whichever is higher.
Amazon signed the GPAI Code of Practice in August 2025, alongside Microsoft, Google, OpenAI, and Anthropic. This is a transparency commitment for general-purpose AI models. It is not equivalent to the EU AI Act conformity assessment required for a high-risk employment screening system. These are distinct obligations, and Connect Talent currently offers no published conformity documentation for European deployment.
Additionally, Amazon Connect Talent includes built-in audit logging and documentation capabilities, which is a start. However, audit logging and EU AI Act compliance are not synonymous. Works councils in Germany, employee representative bodies in France, and data protection authorities across the DACH region will have direct jurisdiction over whether and how this tool can be deployed.
In summary, Connect Talent is currently in preview, built primarily for the US market, and carries significant legal uncertainty for European deployment.
Is this really something we want?
Beyond the legal question lies a question of values, and it is worth considering that seriously.
A hiring process reflects on an organisation. For most people, the candidate experience is their first genuine encounter with your company’s culture. A fully automated voice interview, conducted by AI at 11 pm with no human present, sends a clear message. Whether that message aligns with what your employer brand intends to communicate is a question only you can answer.
For high-volume commodity hiring, the trade-off may be acceptable. For white-collar roles, leadership positions, or teams where culture fit and personality alignment determine whether someone stays and thrives, automating the human element out of the first conversation is a choice with consequences.
The best outcome in hiring is not a faster decision, but a better one.
Why Zortify is a better choice for European HR teams
| Amazon Connect Talent | Zortify | |
|---|---|---|
| What it measures | Competency responses in structured voice interviews | Validated personality profile: Big Five, Entrepreneurial Capital, risk indicators |
| Predicts early attrition? | Indirectly, through competency scoring | ✔Directly — personality dimensions predict first-year retention |
| EU AI Act readiness | ⚠Preview stage, no published conformity documentation for Europe | ✔Built in Luxembourg, GDPR-compliant, designed for European regulatory environment |
| Works council approval | ⚠No documented framework for German/French works council process | ✔Explainable AI outputs, candidate transparency, approved in enterprise DACH deployments |
| Candidate experience | ⚠Fully automated voice interview — no human in the loop | ✔Candidate-facing assessment with own-results access; human decision remains central |
| Research foundation | ⚠Amazon’s internal hiring science (150% annual turnover context) | ✔Peer-reviewed psychometrics since 2018, certified research institution |
| Target use case | High-volume, high-turnover roles: logistics, retail, seasonal | White-collar roles where personality fit predicts performance and retention |
The fundamental difference is not speed or scale, but what is measured and what that measurement actually predicts.
Amazon Connect Talent can tell you, more quickly and consistently, how a candidate performs in a structured competency interview. That is useful. However, it cannot tell you whether that person has the self-efficacy to handle ten rejections in a row, the resilience to recover from a setback, the conscientiousness to follow regulatory requirements without supervision, or the risk profile that could make them a liability in a leadership role.
These are the decisions that determine whether your hires stay, grow, and contribute, or leave within twelve months and cost you 200% of their annual salary to replace.
Zortify was created to address exactly this challenge – not from a generic playbook, but from eight years of certified AI and psychometric research. It is designed for the European regulatory environment and validated specifically for roles where personality fit makes the difference between a good hire and an expensive mistake.
SEE THE DIFFERENCE
Try Zortify for yourself – free
Discover your own personality profile in 30 minutes. Real insights.
Prof. Dr. Florian Feltes
Prof. Dr. Florian Feltes is co-founder and co-CEO of zortify and a forerunner in AI-supported HR innovation. Together with his team, he develops intelligent personality diagnostics and helps companies identify the perfect candidates—without expensive assessments and without bias. His vision: a world in which every company can effortlessly form high-performance teams and create work environments that allow human potential to flourish.
The Sales Gene – does it actually exist?
The mis-hire is not the problem. Not knowing what it costs you is.
