The Candidate Insights Assistant lets you ask natural-language questions about a candidate and get instant answers drawn from that candidate's Greenhouse data — application history, scorecard feedback, stage timing, offer and salary details, and more. Instead of clicking through tabs to assemble a picture of where a candidate stands, you can ask directly and get a synthesized answer in seconds, ready to pass along to a hiring manager or drop into your own notes.
The Candidate Insights Assistant is assistive AI, not automated decision-making. It does not advance, reject, or take any action on candidates, and it will not make hiring recommendations. When asked to weigh in on a decision, the Assistant returns factual status and data only — recruiters and hiring managers remain responsible for all hiring decisions.
The Assistant keeps a history of your chats. Each conversation is saved and can be renamed, and you can start a fresh conversation at any time with New chat.
Turn on the Candidate Insights Assistant for your organization
During the Beta, the Candidate Insights Assistant is available on request. To enable it for your organization, contact your Greenhouse Customer Success Manager (CSM). There is no self-service setup — once your CSM turns the feature on, it becomes available to your users automatically.
Once your CSM has enabled Beta access, you can turn the Assistant on or off from the AI Features menu in Greenhouse Recruiting. This is a per-feature toggle, so you can control the Assistant independently of other AI features.
Note: The Candidate Insights Assistant does not surface any information to candidates and has no candidate-facing component.
Availability and permissions
Once enabled for your organization, the Candidate Insights Assistant is available account-wide. Every user who can access a candidate can use the Assistant on that candidate.
Access to information is governed entirely by your existing permission structure. The Assistant can only draw on data a given user is already permitted to see. If a user cannot view a candidate's scorecards, offer details, or other records, the Assistant will not surface that information to them. In other words, the Assistant introduces no new data access — it respects the permissions you already have in place.
Beyond the on/off toggle, there is no user-level permission specific to the Assistant. To manage who can use it, manage who has access to candidates and their data through your normal Greenhouse permissions.
Use the Candidate Insights Assistant on a candidate
Open the Assistant on a candidate
The Candidate Insights Assistant is available from any candidate's profile.
- Open the candidate you want to ask about.
- In the vertical icon rail on the right side of the profile, click the Assistant icon.
The Assistant panel opens to the right of the candidate's profile. When you first open it, you'll see an Ask about this candidate empty state with a short set of suggested starter questions to get you going.
The Assistant is scoped to the individual candidate whose profile you're viewing. It can reason across everything that person has with your company — every application and prospect record tied to them — not just the application you happen to have open. All answers remain bound by your existing permissions.
Ask a question
To ask the Assistant about a candidate, do either of the following:
- Type your question in the Ask about this candidate… field and send it, or
- Click one of the suggested questions.
The Assistant briefly shows a Thinking… state while it works, then returns an answer in the panel. Answers are specific and grounded in the candidate's Greenhouse data. For example, you might ask about how long a candidate has been in their current stage, their full application and prospect history with your company, a summary of scorecard feedback across interviews, or their offer and salary details — and get a concise, ready-to-share answer back.
Depending on your question, the Assistant may offer a few follow-up questions beneath its answer. These are situational, they won't appear every time, and give you a quick way to go deeper without retyping.
Understand what the Assistant can and can't answer
The Candidate Insights Assistant is designed to help you understand a single candidate quickly — not to make decisions for you. It's useful to know where its boundaries are.
It won't make hiring recommendations or advise on decisions. If you ask the Assistant whether to advance, reject, or take an action on a candidate, it will not give you a recommendation. Instead, it returns the relevant factual status and data and leaves the decision to you. For example, if you ask whether to skip a step and send an offer, the Assistant will decline to advise and instead tell you the factual status of that step.
It answers about one candidate at a time. The Assistant is scoped to the individual candidate whose profile you're viewing. It can reason across all of that person's applications and records with your company, but it cannot answer team-wide, aggregate, or cross-candidate questions. For example, if you ask what your team currently pays for a given role, the Assistant can share that candidate's own offer details but cannot report on other employees' or candidates' compensation.
It's bound by your permissions. The Assistant can only draw on data you already have access to. If you can't see something on the candidate's profile, the Assistant won't surface it either.
Manage your chats
The Assistant saves your conversations so you can return to them later. Chats are saved per candidate profile, per job, and are tied to your user account — another user won't see your chats, and chats you start on one candidate stay with that candidate.
Start a new chat. Click New chat to clear the current conversation and start fresh. Your previous chats are preserved in your history.
Chats are auto-titled. Each chat is automatically given a title based on your conversation. You can rename it if you'd like a clearer label (see below).
View your chat history. Click the chevron next to the chat title, then click View Full History to see your saved chats for this candidate.
Rename a chat.
- Click the chevron next to the chat title, then click View Full History.
- Highlight the chat you want to rename.
- Click the ... icon, then click Rename.
Delete a chat.
- Click the chevron next to the chat title, then click View Full History.
- Highlight the chat you want to delete.
- Click the ... icon, then click Delete.
Data retention and GDPR
The Candidate Insights Assistant does not change how candidate data is retained. It reads the candidate data already in your account and does not create a separate retention regime. Chat content generated by the Assistant is stored as message records associated with the relevant application, so it is governed by the same candidate data retention framework as your other candidate data.
Retention of candidate personal data works the same way it always has in Greenhouse:
- You control retention as the data controller. You set your own data retention rules in Greenhouse Recruiting — per office, and separately for rejected and (if enabled) hired candidates. You choose how long to retain personal data after a candidate is hired or rejected, and which fields are anonymized when that period ends. If you never configure a rule, data remains in your account indefinitely, subject to your own actions and applicable law.
- Greenhouse notifies you when the timer passes. Once a retention period elapses, Greenhouse notifies designated users so they can anonymize or delete candidate personal data, individually or in bulk.
- After your contract ends, all personal data in your account is automatically deleted 90 days after expiration or termination, subject to a short backup period.
Because the Assistant reflects data that already exists on the candidate, anonymizing or deleting a candidate's underlying data through your retention rules removes the source that the Assistant draws on.
Note: You can also delete individual saved chats manually at any time from the Assistant panel (see Manage your chats above).
Frequently asked questions
Greenhouse is built around structured hiring — consistent, well-documented processes that help teams make fair, informed decisions. The Candidate Insights Assistant supports that approach by helping you understand what's already in a candidate's record more quickly, so you spend less time assembling context and more time on the conversation and the decision. It is assistive AI: it retrieves and summarizes information you already have access to, and it does not evaluate, score, rank, or make decisions about candidates.
How does the Candidate Insights Assistant work?
You ask a question about a candidate in plain language, and the Assistant answers using that candidate's Greenhouse data — their applications and prospect records, scorecard feedback, stage timing, offer and salary details, and similar information. It reads only the data you already have permission to see and returns a synthesized, plain-language answer. You can ask follow-up questions, and the Assistant keeps the context of your conversation.
Is AI making hiring decisions in the Candidate Insights Assistant?
No. The Assistant does not make, recommend, or influence hiring decisions. It does not advance, reject, score, rank, or evaluate candidates. Its only job is to retrieve and summarize information you already have access to. If you ask it whether to take an action on a candidate, it will decline to advise and instead return the relevant factual status, leaving the decision to you. Recruiters and hiring managers remain responsible for all hiring decisions.
Does the Candidate Insights Assistant auto-reject or auto-advance candidates?
No. There is no mechanism in the Assistant to take any action on a candidate. It cannot move candidates between stages, reject them, or trigger any workflow. It only answers questions.
How does Greenhouse audit the Assistant for bias or fairness?
The Candidate Insights Assistant does not evaluate, score, or assess candidates, so it does not produce the kind of output that a disparate-impact or fairness audit measures. Because the feature makes no decisions and no evaluations, there is nothing in its output that ranks or rates one candidate against another. It surfaces information that already exists in your account, subject to your existing permissions. Decisions — and responsibility for them — remain entirely with your team.
How does the Candidate Insights Assistant comply with AI-in-hiring regulations?
Many AI-in-hiring regulations focus on tools that evaluate, score, or make automated decisions about candidates. The Candidate Insights Assistant does none of these things — it retrieves and summarizes information you already have access to, makes no decisions, and performs no candidate evaluations. Because it is assistive by design and keeps every decision with a human, it sits outside the category of automated decision-making tools that these regulations typically govern. For technical detail on how candidate data is processed, see the Candidate Insights Assistant - Data Processing FAQ.
As with any tool, your organization is responsible for using the Assistant in line with your own policies and applicable law. We recommend aligning with your Legal team on how your team uses AI-assisted candidate insights.
Can I trust the Assistant's answers?
The Assistant is designed to answer accurately from the candidate data in your account, but like any AI-generated summary, its answers can occasionally be incomplete or misinterpret a question. The Assistant states its answers directly and does not link back to the underlying records. When an answer will inform a decision, confirm the specifics on the candidate's profile — the scorecard, offer, or activity feed — rather than relying on the summary alone. Treat the Assistant as a fast way to orient yourself, not as a system of record.
Additional resources
For technical detail on data sources, personal-data categories, retention, sub-processors, and security for data controllers, see the Candidate Insights Assistant - Data Processing FAQ.
To roll out the Assistant, we recommend aligning with your Talent Acquisition, Legal, and Security stakeholders on how your team will use AI-assisted candidate insights.