> ## Documentation Index > Fetch the complete documentation index at: https://conveo.ai/docs/llms.txt > Use this file to discover all available pages before exploring further. # Review Personal Information in Transcripts > Configure detection, review flagged text, redact transcript content, and understand recording limitations. Conveo can detect personal information in interview transcripts and remove or flag text for review. Use the review workflow before sharing sensitive research findings. Transcript redaction does not alter audio or video recordings. Detection can miss information or flag harmless text. A scan failure does not stop analysis, so completed analysis is not proof that detection succeeded or that all personal information was removed. ## Enable detection An organization admin controls these settings in **Settings → Organization → Organization profile**: * **Detect personal information in transcripts** enables automatic scanning. * **Also remove the participant's name** extends detection when the main setting is enabled. The scan runs after normal interview completion, abandonment, or uploaded-interview transcript creation. When detection is disabled, the scan is skipped and analysis can continue. Enabling detection should not be treated as proof that every historical interview has been rescanned. High-confidence findings in supported categories can be removed automatically. These include addresses, phone numbers, payment-card information, identification numbers, and third-party names. The participant's own name is included when **Also remove the participant's name** is enabled. Lower-confidence findings in these categories are flagged for review. Brand or product names and public-figure references are not automatically removed or flagged by this workflow. Unsupported or unclassified findings may also be skipped. The review queue therefore does not contain every potentially sensitive reference. ## Review a transcript In the interview grid, use the responses menu's **Show … with unreviewed flags** option to focus on interviews needing review. Choose **Show all interviews** to leave that view. Open an interview transcript. When findings exist, the **Personal information** banner reports removed items and items flagged for review. Select **Review** where available to locate flagged content. Open a highlighted finding to inspect its category, explanation, and available action. Reviewing, removing, or restoring findings requires study editing access; the interface may describe this as study-manager access. | Action | Result | | ------------------------ | ------------------------------------------------------------------------------- | | Remove | Redacts the flagged text and records the review. | | Not personal information | Dismisses the flag without removing that text. | | Restore | Returns previously redacted text when the retained original is still available. | A confidence value helps explain the automated finding; it does not guarantee that the classification is correct. Read the surrounding response before making the decision. ## Remove text manually When personal information was not detected, open the participant transcript line's editing controls, choose **Redact this line**, and confirm **Redact**. Manual line redaction removes the whole selected line of text, rather than an arbitrary character selection. The original and translated transcript layers are distinct; review both when a translation exists. Redaction updates the affected transcript and related text or caption surfaces. When a caption track cannot be safely updated, Conveo can suppress that track rather than continue serving its old text. Do not infer that the spoken information has disappeared from a recording. ## Restore and check downstream results Open a removed-text marker and choose **Restore** when offered. Confirm the action and review the transcript afterward. Restoration is unavailable once the retained original has expired or been removed, and later text changes can prevent automatic restoration from safely locating the original position. If redaction affected both original and translated text, those findings may need to be restored separately. Subsequent analysis and dependent content can take time to update. Review reports and shared outputs after making a correction, and do not assume a previously downloaded file changes with the transcript. ## Limitations * Automatic detection is a review aid, not a guarantee of anonymization. * Flagged text remains visible until removed; dismissing a flag keeps the text. * A failed or skipped scan can still be followed by analysis. * Original audio and video remain unchanged, including when captions are redacted. * Redaction and restoration do not revoke previously distributed files or copies. ## When participant data has been erased An interview can show **Participant data removed on request** or **Interview data removed by retention policy** after its content has been scrubbed. These notices explain deliberate removal, rather than a recording or transcription failure. The removed recording, transcript, and answers cannot be restored through transcript restoration controls. A record can remain for recruitment and payment reconciliation. Contact Conveo to handle a participant erasure request and confirm which participant records and studies it covers. Erasure is not the same as hiding an interview or redacting one line. A person can have several participant records, and removing shared participant identity information can affect how that person appears in other studies even when the interview-content scope is narrower. Review previously shared outputs and downloaded copies separately. ## Related guides * [Interview grid and transcripts](/docs/interview-data/interview-grid-and-transcripts) * [Organization settings](/docs/settings-and-management/manage-organization-members) * [Study and output sharing](/docs/settings-and-management/permissions-and-sharing)