AI Interviewer FAQ: 16 Questions from Recruiters, Answered (April 2026)

An AI interviewer is software that conducts all or part of a job interview via text, voice, or video without a live human interviewer present. There are two main categories: async video with AI scoring (platforms like HireVue and myInterview record a candidate's video responses and score them using natural language processing and vocal-feature analysis) and conversational chatbot platforms (like Sapia.ai and Paradox Olivia, which conduct text or SMS conversations). AI interviewers typically operate at the top of the hiring funnel, screening and ranking candidates before a human recruiter reviews the shortlist. They do not replace the entire interview process; they automate the most repetitive screening layer.

An AI interviewer (HireVue, Sapia, Paradox) replaces the human interviewer at the screening stage: candidates interact directly with the AI, and the AI produces a score or ranking. An interviewer-assistant (Metaview is the leading example) assists a human interviewer during a live interview: it records, transcribes, and summarises the conversation, generating structured notes for the recruiter. The interviewer-assistant category is important for legal compliance reasons: Metaview likely falls outside the NYC AEDT definition (it does not produce a hiring-decision score) and may have a lower EU AI Act risk classification. Buyers evaluating both categories should request legal confirmation of the classification for each tool in their jurisdiction.

SMB and lower-mid-market entry points with transparent pricing: Willo from £89/month, Spark Hire from $149/month, myInterview from £149/month. These platforms serve companies with 50-500 hires per year without requiring enterprise procurement processes. Enterprise platforms (HireVue, Sapia, Paradox) are custom-quote only. Based on available RFP data and user reports, typical bands are: Sapia $15,000-$60,000 per year, HireVue $35,000-$100,000+ per year, Paradox $25,000-$150,000+ per year. The bands are wide because pricing is volume-based and negotiable. Always request a scoped quote with your actual hire volume, ATS requirements, and jurisdiction list before any pricing negotiation.

AI interviewing is legal in the US, but subject to growing regulation that varies by jurisdiction. NYC Local Law 144 (AEDT, in force since July 2023) requires annual independent bias audits, published audit summaries, and 10-business-day candidate notice for any AI hiring tool used for NYC positions. Illinois AIVIA requires notice, consent, and data destruction on request for AI video interview tools used for Illinois-based positions. The Colorado AI Act (effective February 2026) adds disclosure and impact-assessment obligations. At the federal level, Title VII and EEOC guidance hold employers liable for disparate-impact outcomes from AI tools, regardless of whether the tool was vendor-built. If you hire in any of these jurisdictions, treat legal review as mandatory before deployment.

NYC Local Law 144, effective July 5, 2023, requires any employer or employment agency using an Automated Employment Decision Tool (AEDT) for any New York City employment action to: (1) conduct an annual independent bias audit of the AEDT; (2) publish a summary of the audit on their public website; (3) give candidates at least 10 business days' notice before using the AEDT. If your company uses an AI interviewer to screen candidates for any position in New York City, you are likely subject to this law. The law applies to the employer, not the vendor; vendor audit documentation may not satisfy the "independent" requirement. Penalties: $500 first violation, $500-$1,500 per subsequent violation per day. Consult employment counsel for your specific use case.

No. The EU AI Act (Regulation 2024/1689) classifies AI used in recruitment, selection, and employment decisions as high-risk (Annex III, Point 4). High-risk classification means a compliance framework applies, not a ban. The framework requires conformity assessment by the vendor, CE marking, registration in the EU database of high-risk AI systems, human oversight by the deployer, candidate transparency (Article 26(11)), and 6-month log retention. High-risk AI obligations are phased: fully in effect from August 2026 for systems under harmonised standards, August 2027 for others. If you are deploying an AI interviewer for EU-resident candidates, start compliance documentation now. Penalties for high-risk system obligation breaches are up to EUR 15 million or 3% of global turnover, whichever is higher.

For Greenhouse: HireVue, Sapia, myInterview, VidCruiter, and Spark Hire all have native Greenhouse integration with bidirectional scorecard sync. Greenhouse has the broadest partner ecosystem in the AI interviewer category (40+ partners). For Lever: HireVue, Sapia, myInterview, and Spark Hire all have native Lever integration. For Workday: HireVue, Paradox, Sapia, and Talview have established native Workday integrations. Workday is the most demanding ATS for integration certification, so mid-market platforms are often API-only rather than native. For all three: verify current integration depth in the ATS's official partner directory before the vendor demo. Integration quality changes with ATS and AI platform releases.

The right alternative depends on why HireVue is not fitting your needs. If bias-defensibility is the priority: Sapia.ai, which publishes an annual FairChoice bias audit and has candidate NPS in the mid-70s. If you want transparent pricing: myInterview (from £149/month) or Spark Hire (from $149/month). If you need structured-interview depth with 20+ ATS integrations: VidCruiter. If you hire primarily hourly roles: Paradox (Olivia), which is SMS-native and the category leader for QSR and retail. If you want an interviewer-assistant rather than a full AI screener: Metaview. Each alternative has trade-offs documented in the full comparison at /platforms-compared.

For some buyers, yes. For others, no. Sapia.ai is better than HireVue if: bias-audit defensibility is a formal procurement requirement; candidate experience is a strategic priority (Sapia publishes NPS in the mid-70s, HireVue does not); text-only format works for your role types; you are a UK, EU, or ANZ buyer; or your budget is mid-market ($15k-$60k) rather than enterprise ($35k-$100k+). HireVue is better than Sapia if: you need deep Oracle HCM or SAP SuccessFactors integration; you need video (not just text) for the role; you are a Fortune 500 with 1,000+ annual hires; or your Workday instance requires a vendor with an established native integration. Neither is universally better; they serve different buyer segments.

No, based on 2026 evidence. AI interviewers automate the screening and structured-assessment layer at the top of the hiring funnel, which is the most repetitive and time-intensive part of a recruiter's workflow. Strategic sourcing, candidate relationship management, final hiring decisions, offer negotiation, and the personal judgment required at later interview stages remain human. The displacement pattern from current AI interviewer deployments is toil-reduction rather than role-elimination. Recruiters who adopt AI interviewers well describe spending less time on scheduling and phone screens, and more time on candidate relationships and closing. The risk of AI fully automating a recruiter role in the near term is low; the risk of a recruiter who uses AI tools outcompeting a recruiter who doesn't is real.

Vendor-claimed predictive validity varies and independent peer-reviewed validation is limited. Sapia publishes correlation coefficients of 0.5+ between their scoring and job performance in their FairChoice research, which is a strong result if the methodology is sound. HireVue has published internal validation studies. The academic literature on AI hiring tool validity is more cautious: facial-analysis-based scoring (discontinued by HireVue in 2021) had weaker validity evidence; text-and-vocal-feature scoring has somewhat stronger evidence but still limited by the quality of job-performance criteria used in training. For most buyers, the practical comparison is not AI-vs-human accuracy but AI-screening-at-scale vs phone-screen-at-scale. Phone screens conducted by time-pressured recruiters doing 50 per week are not high-validity instruments either.

Results are mixed and vary significantly by demographic, role type, and platform. Sapia publishes candidate NPS in the mid-70s for their text-based interview format, which is unusually high for automated hiring tools. Async video interview completion rates (which are a proxy for candidate willingness) vary from 60-85% depending on the company brand and the communication quality of the invitation. Younger candidates report higher comfort with AI interview formats; senior candidates and candidates for senior roles report more resistance. The practical guidance: AI at the top of the funnel (screening) with human at later stages, combined with clear and transparent candidate communication before the interview, produces the most positive candidate experience outcomes.

Depends on hire volume, baseline cost-per-hire, and change-management quality. Using SHRM's $4,700 average cost-per-hire and 300 hires per year as a baseline: if 60% of hires go through AI screening with a 20% cost reduction on in-scope hires, gross savings are $169,200 per year. Against a $40,000 mid-market annual contract, payback is approximately 3 months. The model assumes good change management and recruiter adoption. Without those, realistic reduction is 5-10%, not 20%, and the business case deteriorates. Our ROI calculator lets you model your specific numbers without an email gate.

Ten questions that separate vendors with solid compliance posture from those without: (1) Who is the independent auditor for your AEDT bias audit, and when was the last audit? (2) Can we see the EU AI Act conformity assessment documentation? (3) What is the data-deletion timeline when a candidate requests deletion? (4) What are the data portability terms when our contract ends? (5) What ATS integrations are native vs API vs Zapier? (6) What is the SLA for integration maintenance when our ATS releases an update? (7) Can you provide a candidate-experience NPS or satisfaction score? (8) What does human oversight look like in your architecture? (9) Are there any pending regulatory actions or enforcement notices against your platform? (10) What are the contract termination rights if the product quality deteriorates?

From contract signature to first live candidates: typically 6-10 weeks for a well-resourced mid-market rollout. Breakdown: legal review and contract finalisation (2 weeks), ATS integration and testing in staging (3-4 weeks, can overlap), recruiter training (1 week), candidate notice template and legal review of communication (1 week), ATS workflow configuration (1 week), pilot launch. The most common cause of delay is legal review of the candidate notice and AEDT compliance obligations. Starting the legal review before contract signature rather than after reduces this significantly. See the full implementation playbook at /implementation-guide.

The consequences depend on jurisdiction and how the bias is discovered. If discovered through your own AEDT audit: you are in a better position; you can remediate proactively, adjust the deployment, and document your response. If discovered through an EEOC complaint or class action: employer liability under Title VII applies regardless of vendor responsibility. The EEOC's 2023 guidance is explicit: the employer, not the vendor, is the liable party. If discovered in NYC through DCWP enforcement: fines from $500 to $1,500 per violation per day, plus reputational exposure. Immediate steps: halt the AI interview for the affected role type, preserve all data, engage employment counsel, and contact the vendor for remediation options. The key preventive measure is annual independent bias auditing, which gives you evidence of good-faith compliance even if adverse-impact patterns exist.