This FAQ provides technical and processing information about Greenhouse Notetaker to support your evaluation of the feature.
This is not intended to be legal advice. The answers provided are also not intended to override or contradict the advice of your legal counsel.
Feature overview
What is the Greenhouse Notetaker feature?
Greenhouse Notetaker is an AI-powered feature within Greenhouse Recruiting that automatically records and transcribes video interviews, then generates AI-generated summaries and structured notes mapped to scorecard questions. The feature is designed to reduce interviewer cognitive load during live interviews by capturing the conversation so interviewers can focus on the candidate. After the interview, AI notes and summaries are surfaced within the scorecard to help interviewers submit timely, complete, and accurate feedback. Greenhouse Notetaker is designed as assistive AI – it does not automatically advance or reject candidates, and all hiring decisions remain with recruiters and hiring managers.
What meeting providers does Greenhouse Notetaker support?
Greenhouse Notetaker supports Zoom, Google Meet and Microsoft Teams. In-person and phone-based interviews are not supported.
What technology powers Greenhouse Notetaker?
Greenhouse Notetaker uses Recall.ai as the third-party vendor for meeting bot, recording, and transcription services. AI summaries and structured notes are generated using a large language model (LLM), currently provided by OpenAI.
What outputs does Greenhouse Notetaker generate?
Greenhouse Notetaker generates three outputs from a recorded interview: (1) a video recording of the interview, (2) a searchable text transcript attributed to each speaker, and (3) AI-generated notes — including an overall interview summary and, where scorecard questions are configured, question-level summaries. All three outputs are accessible within the interview kit. Interviewers can append AI notes to their scorecard responses.
Data processing details
What are the roles of the parties under data protection laws?
Greenhouse is a data processor acting on behalf of its customers to process the personal data of candidates and, in certain cases, interviewers. Customers are the data controllers who decide why and how personal data is processed.
What personal data does Greenhouse Notetaker collect and process?
Greenhouse Notetaker processes the following categories of personal data: audio and video content of the interview (including the candidate's voice, image, and any on-screen content); a text transcript of the interview attributed to each speaker; and AI-generated summaries derived from the transcript. Where available, the candidate's name and pronouns from their Greenhouse profile may be used to personalize transcript attribution and summary language.
Does Greenhouse Notetaker process special category data?
Greenhouse Notetaker does not intentionally process special category data (e.g., data relating to racial or ethnic origin, political opinions, religious beliefs, trade union membership, genetic data, biometric data, data concerning health or sex life or sexual orientation). However, because Greenhouse Notetaker records and transcribes interview conversations in full, special category data disclosed incidentally by the candidate or interviewer during the interview may be captured in the recording and transcript. The AI summarization prompt is designed to produce factual, non-evaluative output and is explicitly instructed not to make inferences about candidate personality, intent, or subjective qualities.
Does Greenhouse Notetaker process biometric data?
No. The data used to enable Greenhouse Notetaker does not include biometric data as defined by applicable laws.
Does Greenhouse Notetaker process structured application data?
No. Greenhouse Notetaker generates summaries from the interview recording and transcript only. It does not access or process structured application form data, resume data, or other candidate profile fields beyond the name and pronouns used for transcript attribution.
Who can access the recording, transcript, and AI notes?
Access is controlled by permission settings configured by the customer. Site Admins can access all recordings, transcripts, and AI notes within their organization. Job Admins with the appropriate permission stripe can access materials for interviews on jobs they administer. Interviewers can access their own recordings, transcripts, and AI notes. Access by Greenhouse production engineers is limited to what is necessary to provide the service and is subject to the same access controls applied to other Greenhouse customer production data.
Data retention and storage
How long are recordings retained?
The recording retention period is configurable within Greenhouse data retention settings, subject to any maximum window set by Greenhouse.
How long are transcripts and AI notes retained?
Customers can configure retention rules for transcripts and AI notes within Greenhouse's data retention settings, just as with other data attributes.
What happens to recordings, transcripts, and AI notes when a candidate record is deleted?
When a candidate's personal data is deleted from Greenhouse, all associated recordings, transcripts, and AI notes are also deleted, consistent with the data fields selected at the time of deletion. Customers can include the recording link, transcript, and AI notes in GDPR retention and deletion rules.
What are your data retention obligations as a controller?
Customer Action: Your company must establish its own data retention policies in accordance with applicable laws and your business needs, as recommended by your legal counsel. For Greenhouse Notetaker, consider local employment law requirements, applicable recording consent laws, candidate expectations, and your legitimate business interests in retaining interview recordings and transcripts for quality assurance or future review purposes.
Does Greenhouse use recordings, transcripts, or AI notes for product improvement or model training?
Not currently. If that changes, Greenhouse will not use personal data from interview recordings, transcripts, or AI-generated notes for product improvement or to train machine learning models.
Under our customer agreement, Greenhouse may use anonymized data generated through Greenhouse Notetaker, provided that the data cannot reasonably be used to identify a customer or other individual. Our third-party AI providers are contractually prohibited from using customer data to train, improve, or develop their AI or machine learning models.
Third-party processing
What third parties process candidate data as part of Greenhouse Notetaker?
Two categories of third-party processors are involved: (1) Recall.ai, which provides the recording bot and transcription service; and (2) the LLM provider(s) used to generate AI summaries and notes, currently OpenAI. Both processor relationships are governed by data processing agreements that flow down Greenhouse's data protection obligations.
What data is shared with Recall.ai?
The Greenhouse-hosted meeting bot powered by Recall.ai joins scheduled interviews and captures the audio and video of the interview session. The resulting recording is transmitted to and processed by Recall.ai for transcription. Recall.ai returns the transcript to Greenhouse for storage and use in generating AI summaries.
What data is shared with the LLM provider(s)?
The interview transcript (including candidate name + pronouns, as well as interviewer name) is sent to the LLM provider (OpenAI) for AI summary generation.
What protections exist for Recall.ai data processing and LLM provider data processing?
Greenhouse has executed data processing agreements with each of its subprocessors Recall.ai and OpenAI that flow down all of Greenhouse's data protection obligations, restricts the subprocessors from using personal data for their own purposes, and ensures that data is adequately protected and secured.
Is candidate data used to train third-party AI models?
Greenhouse's data processing agreements prohibit third-party processors from using customer personal data to train or retrain their own models.
International transfers
How are international data transfers handled?
Greenhouse is certified to the EU-U.S. Data Privacy Framework and the UK Extension to the EU-U.S. Data Privacy Framework. Any cross-border transfers of personal data from the EEA or the UK are subject to the EU-U.S. Data Privacy Framework or the UK Extension to the EU-U.S. Data Privacy Framework. Greenhouse secondarily relies on Standard Contractual Clauses (SCCs) and additional safeguards for data transfers.
Greenhouse’s DPAs with both Recall.ai and OpenAI utilize SCCs as the transfer mechanism for cross-border transfers of personal data.
Does Recall.ai process EU candidate data within the EU for EU-silo customers?
No. Data sent to Recall.ai is hosted on Recall.ai’s U.S. servers, just as with any other personal data sent to Greenhouse’s U.S.-based processors.
Security measures
What encryption protects candidate data in Greenhouse Notetaker?
Customer data is encrypted in transit between customers and Greenhouse using Transport Layer Security (TLS) 1.2 or higher. Customer data is encrypted at rest using a minimum Advanced Encryption Standard (AES) 256-bit encryption. Interview recordings are stored in encrypted S3 buckets.
What technical and organizational measures are utilized to protect candidate data?
Greenhouse implements technical and organizational measures as described in Annex II of the Greenhouse DPA. Furthermore, Greenhouse is regularly audited against ISO 27001, ISO 27701, ISO 42001 and SSAE 18 SOC 2 standards by independent third-party auditors. More information about Greenhouse's security program and practices is available at the Greenhouse Trust Portal.
What access controls are in place for recordings, transcripts, and AI notes?
Access to recordings, transcripts, and AI notes is controlled through Greenhouse's permission model. Interviewers access only their own materials. Job Admins require a specific permission stripe to access other interviewers' materials. All org-level configuration changes to Greenhouse Notetaker are tracked in the Greenhouse organization audit log.
What API security measures exist?
The internal services used to power the Greenhouse Notetaker feature is on a private network to prevent internet-facing traffic. All internal services used to power the feature require authentication.
What AI-specific security measures are implemented?
Greenhouse adheres to Generative AI security best practices, including mitigations for vulnerabilities such as prompt injection, output handling, denial of service, and data leakage. The AI summarization prompt is designed with guardrails to prevent the model from making hiring recommendations, speculative inferences, or evaluations of candidate personality or intent.
Individual rights implementation
How can candidates exercise their data subject rights?
Customer Action: You must establish processes for candidates to exercise their rights under applicable data protection laws, as advised by your legal counsel. This should include clear contact information and procedures for handling access, correction, deletion, and objection requests related to recording, transcription, and AI note processing.
How can candidates access their recording, transcript, and AI notes?
If a candidate requests access to their personal data, you can include the recording link, transcript, and AI notes in an exported candidate packet. The recording link will expire after a set period of time, currently 7 days.
How are deletion requests handled for Greenhouse Notetaker data?
Customer Action: As the data controller, you must choose how candidates should contact you to exercise any rights under applicable data protection laws, as advised by your legal counsel. In terms of product functionality, Greenhouse's GDPR data retention and deletion rules include options to delete the recording link, transcript, and AI notes when a candidate's personal data is deleted. If the entire candidate record is deleted, all associated recordings, transcripts, and AI notes are also removed.
What opt-out mechanisms exist for candidates?
Candidates can opt out of being recorded at the time of scheduling via the availability request or self-scheduling flow. A candidate who opts out will not be recorded, and Greenhouse Notetaker will not be activated for their interview. If a candidate opts out during a meeting, the interviewer must remove the recording bot using the controls in the interview kit or, alternatively, by removing the bot using the participant controls in the meeting provider. Candidate opt-in and opt-out status is tracked per application and is visible on the candidate profile.
Customer Action: You should implement processes for candidates to opt out of recording and AI processing if required by applicable law or your privacy policies, as advised by your legal counsel. Consider providing clear notice about recording and AI use, and opt-out options, during the scheduling process.
Is an opt-in consent model available for candidates?
Customer Action: At this time, there is no opt-in consent available within the Greenhouse Notetaker product feature. Customers may choose to add a custom question to the application requesting consent for use of the Greenhouse Notetaker feature, including recording. You should determine, in consultation with your legal counsel, whether opt-out or opt-in consent is appropriate for your candidate population and applicable jurisdictions.
Explicit candidate opt-in (where recording is off unless the candidate actively consents) is planned as a post-GA feature for regions where opt-out may be legally or culturally insufficient. The opt-in model, including the consent flow design, disclosure language, and standalone consent page, has not yet been finalized.
Can candidates correct or challenge AI-generated notes?
Customer Action: As the data controller, you must choose how candidates should contact you to exercise any rights under applicable data protection laws, as advised by your legal counsel. In terms of product functionality, AI notes are surfaced to interviewers as assistive input to the scorecard and are not shared with candidates or used as stand-alone decision records. Interviewers are expected to exercise independent judgment when completing their scorecards.
Automated decision-making
Does Greenhouse Notetaker make automated decisions about candidates?
No. Greenhouse Notetaker does not make automated decisions about candidates. The feature records, transcribes, and summarizes interviews to assist interviewers, but does not produce ratings, scores, hire/no-hire recommendations, or any output that automatically advances or rejects a candidate. The interviewer's scorecard, completed independently using the AI notes as reference material, is the input to hiring decisions. All advancement and rejection decisions require human action by a recruiter or hiring manager.
Therefore, Greenhouse Notetaker is neither a form of automated processing under Article 22 of the GDPR, nor is it an ADMT as defined by applicable law.
What human oversight mechanisms are in place?
Humans are required to make every hiring decision. There is no mechanism by which Greenhouse Notetaker automatically takes action on a candidate application. AI notes are presented within the scorecard edit view as input for the interviewer to review, edit, and supplement; they are not pre-populated as the interviewer's submitted assessment. The AI summarization prompt is explicitly designed to avoid hire/no-hire language, attribute ratings, or evaluative conclusions about the candidate.
How does the system provide transparency to interviewers?
The system provides transparency through: a searchable transcript with speaker attribution so interviewers can verify the source of any AI-generated summary claim; question-level summaries linked to specific scorecard questions; and an overall interview summary. Interviewers are encouraged to refer to the transcript and recording to validate AI note accuracy before submitting their scorecard.
Bias mitigation and fairness
What bias mitigation measures are implemented?
Greenhouse has implemented the following measures: (1) the AI summarization prompt uses inclusive language defaults and, where candidate pronouns are available in the Greenhouse profile, attempts to use them in summaries — defaulting to gender-neutral pronouns otherwise; (2) the prompt is explicitly constrained to avoid speculative, evaluative, or subjective language about candidate personality, intent, or cultural fit.
Is a formal bias audit required for Greenhouse Notetaker?
Based on Greenhouse Legal and AI Ethics Committee review, a formal third-party bias audit is not required for Greenhouse Notetaker because the feature summarizes interview transcripts and does not score, rank, or categorize candidates.
How is the AI summarization prompt evaluated for quality and fairness?
Prior to beta, Greenhouse conducted structured internal testing (Arbor testing) with a diverse group of participants, including a range of accents and speech patterns. Outputs were evaluated using a participant survey and an LLM-as-judge framework assessing factual grounding, non-evaluative tone, coverage, concision, clarity, and correct pronoun/name handling. Results were used to refine the prompt before beta. An evaluation set of 100–200 anonymized interviews will be maintained for ongoing prompt quality monitoring.
Technical specifications
How does Greenhouse Notetaker join a video interview?
A Recall.ai-powered meeting bot joins the scheduled interview automatically at the scheduled start time (within approximately one minute). The bot appears as a named participant ("Greenhouse Interview Notetaker") in the video conference. The interviewer can pause or cancel the recording from the interview kit. If the bot fails to join, the interviewer is responsible for initiating any alternative recording process per your organization's policies.
How is the transcript generated?
After the meeting ends, Recall.ai processes the recording and returns a speaker-attributed text transcript to Greenhouse. Transcription is currently supported in English only. Processing typically completes within a few minutes of interview end; an interim status is shown in the scorecard while processing is in progress.
How are AI notes and summaries generated?
Once the transcript is available, it is sent to the LLM provider for summarization. For interviews with scorecard questions configured in Greenhouse, scorecard question text and question-level summaries are generated by mapping relevant transcript segments to each question. An overall interview summary is also generated.
What happens if transcription or AI summary generation fails?
If transcription or AI summary generation fails, the interviewer will be notified within the scorecard interface. In such cases, the interviewer should complete their scorecard using their own notes and recollection. Feedback mechanisms are available in-app for interviewers to report transcription or summary quality issues.
Legal basis and compliance
What legal basis should I use for processing?
Customer Action: As the data controller, you should make this determination in coordination with your legal counsel. For Greenhouse Notetaker, you must determine your lawful basis for recording and transcribing interviews and generating AI notes about candidates. Consider whether you will rely on legitimate interest (with a balancing test), explicit consent, or contractual necessity. For EU/EEA candidates or customers operating in jurisdictions with recording consent laws, explicit consent may be required. Your choice of legal basis will affect candidate notice requirements, individual rights, and opt-out or opt-in obligations.
What candidate notice must I provide?
Customer Action: As the data controller, you should make this determination in coordination with your legal counsel. You should provide clear notice to candidates before or at the time of scheduling that their interview may be recorded and transcribed, and that AI may be used to generate summaries. This notice should describe the purpose of recording, how data will be used and retained, and how candidates can exercise their rights, including how to opt out. Greenhouse provides a candidate-facing disclosure on the scheduling flow (availability request and self-scheduling pages) to support this requirement.
Are there specific recording consent laws I should be aware of?
Customer Action: Recording consent requirements vary by jurisdiction. In the United States, some states require two-party or all-party consent to record a conversation. Some jurisdictions require that participants be notified audibly or visually that recording is in progress. You should consult your legal counsel to determine which recording consent laws apply to your candidate population and how to satisfy them using the disclosure and notification features available in Greenhouse. The Greenhouse Notetaker feature notifies all meeting participants that a recording bot has joined the call as a visible meeting participant.
What are my obligations regarding data subjects?
Customer Action: As the data controller, you should make this determination in coordination with your legal counsel. You may need to identify your specific data subject categories (job applicants, interviewers) and assess the volume of personal data you expect to process. Consider local employment laws, recording consent requirements in your markets, and any specific requirements for your geographic footprint.
Risk assessment support
What risks should I document in my assessment?
Customer Action: As the data controller, you should make this determination in coordination with your legal counsel. For Greenhouse Notetaker, consider risks specific to your organization including: applicability of recording consent laws in jurisdictions where you hire; the sufficiency of your candidate disclosure and consent model; accuracy and completeness of AI-generated summaries for your interview types; potential for AI notes to introduce or reinforce bias in hiring decisions; data security for video recordings stored in the cloud; compliance with local employment laws and data protection regulations; and the impact on candidate experience and trust. Assess likelihood and severity based on your specific context.
What safeguards should I implement?
Customer Action: As the data controller, you should make this determination in coordination with your legal counsel. Based on your risk assessment, consider implementing: clear pre-interview recording disclosures and consent mechanisms; human oversight requirements so interviewers treat AI notes as input, not conclusions; monitoring for demographic bias in hiring outcomes; candidate opt-out (or opt-in) mechanisms appropriate for your jurisdictions; data retention policies aligned with local laws and your storage cost considerations; and training for interviewers on how to use Greenhouse Notetaker responsibly.
Monitoring and governance
What ongoing monitoring should I implement?
Customer Action: You should make this determination in coordination with your legal counsel. Establish processes to monitor: interviewer compliance with human oversight expectations; candidate feedback and opt-out rates; AI note accuracy for your specific interview formats; any patterns suggesting demographic disparities in hiring outcomes for Greenhouse Notetaker vs. non-Greenhouse Notetaker interviews; and compliance with your established data retention and deletion policies.
What monitoring does Greenhouse perform on Greenhouse Notetaker?
Greenhouse has implemented observability and alerting (including Datadog dashboards) to monitor end-to-end pipeline reliability for recording, transcription, and AI note generation. An LLM-as-judge evaluation framework is maintained to monitor AI summary quality against a fixed rubric covering factual grounding, non-evaluative tone, and pronoun handling. Additionally, Greenhouse will monitor user feedback via in-app ratings and feedback.
How often should I review my assessments?
Customer Action: You should make this determination in coordination with your legal counsel. Schedule regular assessment reviews based on: changes in your hiring volume or geographic scope, updates to applicable data protection or recording consent laws, candidate feedback or regulatory inquiries, and at minimum annually to ensure continued compliance.
Documentation and audit support
What documentation should I maintain?
Customer Action: You should make this determination in coordination with your legal counsel. Records will typically include: your data protection impact assessment (DPIA) or equivalent risk assessment; legal basis determination and balancing test (if applicable); candidate notice and consent mechanisms; interviewer training materials on AI use; records of candidate opt-out requests and how they were handled; data retention and deletion activities; and any bias monitoring results.
Does Greenhouse retain configuration and audit data for Greenhouse Notetaker?
Greenhouse retains organization-level configuration changes for Greenhouse Notetaker in the audit log (including enabling/disabling the feature and office/department-level configuration changes), which customers can access. This log records who made each change and when, to support governance and compliance review.
What documentation does Greenhouse provide to support customer DPIAs?
This FAQ is intended to provide the technical and processing information customers need to conduct their own data protection impact assessments. Additional information about Greenhouse's security program and certifications is available at the Greenhouse Trust Portal and in the Greenhouse DPA. For questions not addressed here, customers should contact their Greenhouse account team.