Greenhouse Voice AI

Voice AI doesn’t replace human interviews. It provides a structured first conversation that helps teams identify qualified candidates more reliably than CVs. By filtering out low-fit applicants, Voice AI allows recruiters to focus their limited time on the candidates most likely to succeed in final rounds.

Candidate ratings average 4.5 out of 5, with 85–90% of candidates rating the experience 4 or 5 stars, and opt-out rates are just 10%. Candidates tend to value the flexibility of interviewing on their own schedule, the ability to ask candid questions early and the consistent screening experience.

Greenhouse Voice AI conducts a two-way voice conversation with dynamic follow-up questions and approved answers to candidate questions. It does not ask candidates to submit a one-way recording or rely on a fixed chatbot script. Your hiring manager defines the rubric before interviews begin, and your team can review the evidence behind every score.

“Conversational AI” is a broad category – chatbots, text agents, scripted voice products. Most are built to automate contact, not evaluate for hiring.

Greenhouse Voice AI is purpose-built for the interview: a two-way voice conversation scored against your team’s own rubric.

Three things set it apart:

• Rubric-anchored follow-ups. Voice AI probes on the criteria your team defined, not a fixed script.
• Sub-two-second response latency. The conversation flows naturally, not like a lag-and-respond bot.
• Monthly public audits. Warden AI audits Voice AI’s scoring every month and publishes the results.

‍

Unlike thin CV data, Voice AI provides a structured interview that reveals how candidates think and apply their experience to the role. It surfaces richer insights while filtering out fraud, low-intent applicants and fake profiles before they reach your recruiters.

Because it gives them a better chance to stand out – and to signal how serious they are about the role. Instead of being reduced to keywords on a page, candidates get to explain their experience, answer structured questions and ask their own questions about the role and company. For jobseekers who actually want this job (not just any job), Voice AI is the first place in the process where that intent counts.

Scores are generated against the customer’s own questions, rubric, and examples of strong, acceptable and weak answers. Recruiters and hiring managers retain full control: They can review the transcript and rationale and override scores if they disagree. Hiring decisions remain human-controlled throughout.

Three layers of protection:

• Monthly independent bias audits. We publish the results of a monthly independent bias audit here, which exceeds the annual cadence required by many regulations. Combined with mandatory human review, that’s the core of our fairness posture.
• Human-owned decisions. Voice AI never rejects or advances a candidate, and your team makes every hiring decision. Every score is overridable, and every interview leaves a full transcript and audit trail.
• Safe and secure. Greenhouse is ISO 42001 certified, an international AI governance standard that affirms our commitment to safe, ethical and transparent AI. Voice AI is in scope for our next audit cycle in Q4. Our SOC 2 Type II audit reports are also available on our trust portal.

No. Greenhouse does not use personal customer data to train AI models. Teams control the questions, scoring criteria and approved role information used during candidate conversations.

Voice AI is sold as an annual subscription that scales with your organisation’s size and hiring needs, not as a per-seat add-on. It’s priced as a standalone product based on company or scoped team headcount, with a generous fair-use allowance of interviews sized to your expected volume. Customers typically start with a 12-month+ contract, with optional pilot periods structured as an opt-out window on that annual agreement.