What you will make

A CSV action table and a short reconciliation note for the exact section reviewed.

Project teams need to find actionable statements quickly. An extraction becomes misleading if it merges unrelated actions, invents deadlines, or assigns work to someone the source never names. A transparent table supports review and discussion.

• Define the section and record set before extracting actions.
• Preserve source wording and mark unspecified responsibility or timing.
• Reconcile extracted rows with the selected source statements.

Work through the method

1. Set the section boundary

   Select a small public section and record headings, pages, date, and URL. Define included statements, such as numbered recommendations, before listing their source identifiers.

   Check your work: The inclusion rule and section boundary are explicit.

2. Design the table

   Use Source_ID, Action_Text, Page_or_Section, Person_or_Organization_Stated, Timing_Stated, and Review_Note. Keep original identifiers and use not stated where responsibility or timing is absent.

   Check your work: Source fields and interpretation notes remain distinct.

3. Extract conservatively

   Populate the table from the selected section, using AI or manual entry. Preserve action wording; flag compound statements instead of turning every goal into a task.

   Check your work: Each row matches a specific source statement.

4. Check wording and status

   Check wording and whether each item is a recommendation, adopted action, or example. Remove inferred responsibility, dates, or acceptance of work.

   Check your work: The table retains the source's stated commitment.

5. Reconcile the section

   Compare source identifiers with output identifiers. Check missing, duplicate, and extra items. Explain any split rows sharing a source identifier; row count alone is insufficient.

   Check your work: Every included source identifier has a recorded disposition.

6. Save for review

   Export and reopen the CSV. Check headers, punctuation, long text, and missing fields. Add scope, unresolved questions, and an AI-assistance note where applicable. Save only the CSV block as .csv; keep the reconciliation note in a separate text file or in your combined practice PDF.

   Check your work: Another student can audit the reopened table.

Prompts to adapt

Replace the bracketed details with your own question and authorized source material. Check the output rather than treating it as evidence.

Extract only explicitly stated actions from [authorized public document, named section, pages]. Output CSV columns Source_ID, Action_Text, Page_or_Section, Person_or_Organization_Stated, Timing_Stated, Review_Note. Preserve wording and source identifiers. Use not stated for missing responsibility or timing. Do not infer an award, approval, commitment, deadline, or implementation status. Flag compound or ambiguous statements instead of resolving them. Return the CSV as one separate output containing only the header and rows. In a second, separate text output, list the source identifiers examined and any unreadable or excluded material.

Create a bounded action register

1. Select one public plan section and list its source action identifiers.
2. Extract text, source location, and only explicitly stated responsibility and timing.
3. Check every row, then reconcile missing, duplicate, and extra identifiers.
4. Export and reopen a CSV with a scope and ambiguity note.

Bring the method into a product

• CSV action table with preserved source identifiers
• Section reconciliation and review note

Suggested formats: CSV, PDF, TXT.

Reflect: Which part of your table would become misleading if an inferred responsibility looked like a source statement?

Avoid these mistakes

• A goal, example, and implementation action are not automatically interchangeable.
• Assigning a contact or date without source evidence creates a false commitment.
• More rows do not prove more complete extraction; reconcile source identifiers.