Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for journalists

Interview Transcript Attribution Agent

A transcript where every attributed quote is supported.

Interview Transcript Attribution Agent: what goes in, what the agent does and what you get

What it does

Edited interview transcripts sometimes put words in the wrong person's mouth, especially when speakers overlap or the editing moved lines around. This agent finds every attribution that might be wrong or ambiguous. It checks each against the audio alignment, the speaker's voice timing and the surrounding context. When the evidence clearly supports one speaker, it corrects the attribution. When evidence conflicts or audio overlaps, the speaker stays marked as uncertain rather than given a plausible name. It prepares a corrected draft with each change explained. The editor or journalist decides before the transcript is used for quotes. Edge case: a quote already used in a published article is flagged separately if its attribution changes.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueApprovedNo 1 STARTS WHEN Transcript edited 2 USES A TOOL Find ambiguous attributions 3 USES A TOOL Check audio alignment and context 4 CHECKS THE RESULT Is the speaker clearly supported? If not: mark the speaker as uncertain. Back to step 3. 5 DOES Flag changed quotes already used elsewhere 6 DOES Prepare the corrected draft 7 YOU APPROVE Journalist approves before quotes are used 8 RESULT Attributed transcript review
Read the steps as a list
  1. Transcript edited
  2. Find ambiguous attributions
  3. Check audio alignment and context
  4. Is the speaker clearly supported?If not: mark the speaker as uncertain. Back to step 3.
  5. Flag changed quotes already used elsewhere
  6. Prepare the corrected draft
  7. Journalist approves before quotes are usedThe agent waits here for your OK.
  8. Attributed transcript review

How it decides

It attributes only with audio or context support.

  • Uncertain speakers stay marked uncertain.
  • Only clear audio evidence assigns a speaker.
  • Changed published quotes are flagged.
  • Every change is explained.

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Audio and transcript sources
  • Confidence needed to assign a speaker (default: clear audio match)
  • How uncertain lines are marked
  • Who approves (default: the journalist or editor)
  • Whether to check published quotes (default yes)

What keeps you in control

It always asks you first

  • Publishing quotes

Hard limits

  • No guessed attribution.

It stops when

  • Done: review ready.

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensA 90-minute panel transcript is edited on August 14. The agent finds 9 ambiguous attributions. Seven are resolved by audio timing. For one sentence, both guests spoke at once, so the check fails and it is marked uncertain. Another quote changes from guest A to guest B and was used in an article draft, so it is flagged. The journalist approves the corrected draft.

More agents for journalists