Home/Blog/Ashby Candidate Notes: Visibility and AI Write Controls
Recruiting AI

Ashby candidate notes: visibility checks before an AI agent writes

Understand Ashby candidate note visibility, private notes and permissions. Build a reviewed AI note workflow without exposing role-specific context.

Remqa · November 14, 2026 · 5 min read

Short answer

Ashby notes belong to the candidate profile, not a specific job consideration. Before an AI integration adds a note, check its intended audience, the actual note permissions and whether private handling is required. Treat AI preparation and the final write as separate steps with recruiter review.

  • Candidate notes are profile-level, even when a job consideration is selected
  • Followers from previous or closed considerations may see notes and receive notifications
  • Private notes have a different audience from ordinary candidate notes
  • An AI write needs an explicit audience decision, approved text and a verified result

A recruiter opens a candidate under one role and asks an AI assistant to save a handoff note. The text includes a concern from another interview process. The recruiter expects the current hiring team to see it. That expectation needs checking before anything is written. In Ashby, a note belongs to the candidate profile rather than a specific job consideration. The job selected on screen does not make the note private to that role.

This guide explains the native note behavior and a recommended design for custom AI preparation. The approval record, audience checks and write journal below are engineering controls to build around an integration. They are not claims that Ashby automatically supplies those controls for every note.

Start with the profile-level audience

Ashby's candidate feed tracks events, notes, feedback, emails and other activity. The feed generally shows information relevant to the selected job consideration. Notes have an important separate rule. Ashby explicitly says they are tied to the candidate profile and do not currently support role-specific visibility and notification limits.

Users following the candidate may include hiring team members from earlier or closed considerations. Ashby warns that those users may be able to see notes and receive notifications. A custom assistant should therefore ask which information belongs on the shared profile, rather than assuming every observation from the current role belongs there.

The selected job is context, not a note audience control. If a note is only appropriate for a particular hiring team, do not publish it as an ordinary profile note on the assumption that the job selector limits access.

Separate viewing, adding and private access

Ashby's access-role settings distinguish Can See Notes from Can Add Notes. Adding notes requires the viewing permission. The add permission is separate from Can Add/Edit Candidates. That distinction matters when designing a recruiter workflow that should allow an operational handoff without granting broad candidate editing.

The candidate feed guide also provides an Is note private? option. It describes private notes as visible to Organization Admins and Elevated Access users with Admin Private access to the job, department or location. Private does not mean visible only to the author or to one selected recruiter. Check the effective roles for the actual people involved.

QuestionNative distinctionIntegration decision
Can the person read ordinary notes?Can See NotesCheck the reviewer's effective access
Can the person add a note?Can Add Notes depends on Can See NotesKeep note creation separate from profile editing
Is the text sensitive?Private notes use a restricted audienceConfirm that audience before saving
Is an API key writing?candidate.createNote requires candidatesWriteReview API scope separately from human roles

Build a narrow source packet

Before generating text, define the purpose of the note. A sourcing handoff, an interview follow-up and an internal operational reminder have different evidence needs. Include the candidate identity, the relevant job context, approved source excerpts and the intended audience. Leave unrelated history out of the model input unless it is authorized and necessary.

Mark each excerpt with its origin and date. Keep a recruiter's statement distinct from an interviewer's observation and from an AI inference. If the assistant cannot support a sentence from the packet, remove it or flag it for review. A plausible summary should not become a new factual claim on the candidate profile.

Cross-role context deserves an explicit decision. Suppose a candidate was considered for two teams with different requirements. A gap against one role's requirement should not become an unqualified statement about the person. Preserve the role and source, or leave the statement out of a general handoff.

Review the text and the destination together

A reviewer should see the exact candidate, proposed note, source references and visibility choice in one place. Show the job consideration as context while clearly labeling the write as profile-level. A generic approval button that hides the destination is not enough for this workflow.

  1. Confirm the candidate identity against a stable identifier, not a name alone.
  2. Read the evidence supporting each material sentence.
  3. Check whether the text is appropriate for the proposed audience.
  4. Decide whether mentions or tasks are necessary.
  5. Approve the final text and intended write destination together.

Make edits invalidate the previous approval. If the assistant regenerates the note, changes its audience or switches candidate, ask the reviewer to examine the new proposal. This is a recommended integration rule. It prevents an earlier approval from silently authorizing different content.

Handle mentions and tasks deliberately

Ashby's manual note flow supports mentioning users. The recruiter can select Create a task for the tagged user and set a date. That task appears on the tagged user's Home page after the note is added. A summary integration should not assume that mentioning someone is merely decorative text.

For an initial rollout, keep generated handoff notes free of mentions and tasks unless the reviewer explicitly selects them. Test the final native behavior separately. Do not infer that the note API supports every option in the manual interface, or that pasted mention text creates the same task behavior.

Verify the write before retrying

Keep a custom journal of the approved request, candidate identifier, proposed content and observed outcome. After submission, verify the resulting note using an authorized read or the native interface. If the response is lost, record an uncertain result and reconcile it before sending again. A timeout does not prove that no note was created.

Do not assume createNote provides a duplicate-prevention key unless the current contract explicitly establishes it. The same principle applies to edits. Ashby's feed guide says a user cannot edit a note on behalf of another user. Design correction ownership before rollout rather than assuming a recruiter can rewrite every integration-created note.

Test visibility with real access roles

Use test candidate data and representative accounts before enabling production writes. Include an ordinary reviewer, a user without note access and a user with the relevant private access. Where possible, include a follower from a closed consideration. Check both the note view and notification behavior without exposing real sensitive content.

Common mistakes are treating a job filter as a security boundary, assuming private means author-only, copying restricted evidence into ordinary text and retrying uncertain writes blindly. Also test your own review screen, exports and logs. Restricting the native note does not automatically restrict copies made by the custom application.

How Remqa approaches Ashby note workflows

Remqa builds AI agents for recruiting teams with explicit source, review and write boundaries. A useful starting scope is one note type with a known audience and an operator who owns corrections. Expand only after the team can verify the result and explain who can see it.

For the integration side, review our Ashby API permissions guide. If your team wants to remove repetitive preparation without delegating candidate judgments, bring a sample workflow to a recruiting workflow discussion. Start with the audience decision, then build the note writer.

FAQ

Questions people ask.

Are Ashby candidate notes tied to a job consideration?

No. Ashby documents notes as belonging to the candidate profile. Selecting a job consideration does not create a role-specific note visibility boundary.

Who can see private notes in Ashby?

The candidate feed guide identifies Organization Admins and Elevated Access users with Admin Private access to the job, department or location. Test the effective access in your configuration before using private notes for sensitive material.

Can someone add notes without editing candidate profiles?

Ashby access roles separate Can Add Notes from Can Add/Edit Candidates. Can Add Notes requires Can See Notes. Inspect the effective role rather than assuming note creation requires general candidate editing.

Can an integration create Ashby candidate notes?

The candidate.createNote API endpoint creates a candidate note and requires candidatesWrite. That API permission is different from the human access-role setting. Review the current API contract before choosing visibility or attribution behavior.

Does selecting a job stop former hiring team members seeing a note?

No. Ashby warns that candidate followers, including team members from previous or closed considerations, may be able to view notes and receive notifications. Do not treat the selected job as an audience restriction.

Talk to an engineer

Want an engineer to look at your setup?

Book a free 30-minute call. We look at your tools and your team's week and tell you what we would do first.

Book a free 30-minute call
30 minutesFree, with an engineer
Your stackWe look at what you run today
Straight answerIncluding when we're not the fit