Skip to main content

🧼 Check for Remaining Elements to Remove from a Document

Easily identify mentions to clean up before sharing a file — client names, [to fill], TODO... Let the AI review it for you.

M
Written by Maxime Renault
Updated over a week ago

🎯 Goal

Before sending a deliverable to a client or partner, you want to ensure it no longer contains:

  • Confidential client or project names (e.g., ACME)

  • Internal notes or placeholders (e.g., [to fill], [TODO], [pending])

  • Uncleaned annotations (e.g., “to review”, “check this section”)

The AI assistant can automatically scan and check the entire document for these elements.


🧠 How It Works

  1. Select a document (Word, PDF, presentation, etc.)

  2. Ask the assistant something like:

    • “Can you check if there are any mentions of the client ACME in this document?”

    • “Are there any remaining [to fill] or TODO sections?”

    • “Is anything still incomplete or not finalized?”

The assistant will:

✅ Read the entire document
✅ Detect all relevant occurrences
✅ List them with context (snippet, page, line…)
✅ Help you fix them before sending


🔍 Real-World Use Cases

  • Ensure the name ACME hasn't been left in a document copy-pasted for another client

  • Check that no [to fill] field was forgotten in a CCTP, technical proposal, or contract

  • Automatically detect hidden comments (in headers, text boxes, metadata…)


🤖 Example Prompt

You are an assistant specialized in reviewing documents before they are shared with a client or third party.

Your goal is to help identify all elements that should not appear in a final version, including:

  • Client/company names (e.g., “ACME”)

  • Internal notes or comments (e.g., “[to fill]”, “[TODO]”, “to be validated”, “X to confirm”)

  • Incomplete or questionable sections (e.g., empty fields, vague wording)

  • Placeholders to be replaced (e.g., “Client Name”, “Model Project”)

You should:

  • Scan the entire document (main text, headers, footers, tables, comments if available)

  • List all suspicious elements with a snippet or location when possible

  • Clearly flag anything that seems problematic

Format example:
⚠️ Mention of client “ACME” found in the “References” section

If nothing is detected, say so explicitly.


✅ Benefits

  • Avoid embarrassing mistakes in client-facing documents

  • Deliver clean, professional files every time

  • Thorough check including less visible areas (titles, tables, comments…)

  • Save time compared to full manual proofreading


📌 Key Takeaway

This use case turns your AI assistant into a last-minute proofreader.

It helps catch anything that might cause issues once the document is shared externally.

🎯 Before sending, just ask: "Are there any internal or incomplete mentions left in this document?"

Did this answer your question?