Transcript redaction brief Use this editable brief to discuss a custom AutoQC campaign that processes audio into redacted transcripts. Describe what to remove, what to retain, and how the result will be reviewed. Use fictional or safely anonymised examples. Do not add recordings or client data to this brief. 1 Campaign and language requirements Campaign and intended use Working title, owner role, intended use, who will read the output, and where it will be shared. [Add response] Languages and scripts List Indian or global languages, regional varieties, scripts and mixed-language speech. State whether to retain the original script, transliterate, or request a separate translation. [Add response] Volume and recording conditions Estimated audio hours, files per batch, typical duration, frequency and desired turnaround. Describe channels, speaker count, noise, overlapping speech and other known conditions. [Add response] Add pages or expand any response as needed. An unanswered field is a question to settle during scoping. 2 Redaction and output rules Identifiers to remove Specify names, locations, contact details, account references or other identifiers. Describe whether each rule covers people, organisations, places, or only defined cases. [Add response] Sensitive conversational content to remove Define confidential business terms, sensitive passages, profanity or abuse. Give context and language-specific examples; avoid a broad instruction such as remove all sensitive content. [Add response] Keep rules and exceptions State which ordinary complaints, quoted phrases, business facts or other meaning must remain. Explain exceptions and which rule wins if remove and keep instructions conflict. [Add response] Replacement style and output layout Choose labels such as [NAME], a neutral mask, or another agreed replacement. Describe repeat-mention consistency, desired file format and transcript layout. [Add response] Speakers and timestamps Describe speaker labels, unknown-speaker handling and timestamp intervals. State any need for verified speaker identity separately from turn labels. [Add response] 3 Handling review and next steps Source audio handling Who can provide and authorise the audio? How should it be transferred, accessed and retained or deleted? List restrictions. Audio redaction is a separate scope; this brief concerns redacted transcripts. [Add response] Processing storage access and deletion List required processing location, permitted services, access roles, storage duration, deletion evidence and handling of temporary transcripts or working files. Mark items that need agreement. [Add response] Reference examples and acceptance Describe a safe test set with expected removals and expected retained content across languages and conditions. Name review ownership, how errors will be counted, acceptable thresholds, and how ambiguous cases or revisions will be resolved. [Add response] Enquiry contact and open questions Optional business contact, preferred reply method, target date and questions for scoping. You may send a completed brief to hello@autoqc.in; sending the brief starts a requirements discussion. [Add response] Agree the campaign rules, handling terms and acceptance evidence before processing client recordings. Language coverage and results depend on the agreed scope and representative evaluation.