Close checklist management with AI: daily status, bottlenecks and a critical path

When to use it

Use it every day of the close, whether your checklist lives in a close-management tool, a shared spreadsheet or a project tool. The AI isn't doing accounting here; it's summarizing a list. That makes it one of the safest places to start.

Input data

• Checklist export: task, owner, reviewer, due day (e.g. BD+3), status, completion date, dependencies (if tracked).
• The close calendar (business-day mapping to dates).
• Optional: last 3 closes' completed checklists, for the retrospective.

The prompt

Replace the [bracketed] items. Paste it with your files attached. It works in any capable assistant; tool-specific notes are in the steps.

Here is today's close checklist export for [month] close. Today is business day [BD+X] ([date]).

Attached: [checklist file]; close calendar: [file or paste].

1. Status summary: count of tasks by status (not started, in progress, prepared, reviewed) overall and by area.
2. Late or at risk: tasks past their due day, and tasks due today or tomorrow that are not started. Group by owner.
3. Critical path: using the dependencies (or, if none are listed, the standard order: subledgers → reconciliations → accruals/adjustments → flux → reporting), identify the open tasks that are blocking downstream work.
4. Review backlog: tasks prepared but not reviewed for more than [1] business day, grouped by reviewer.
5. Draft a 6–8 line status note for the close channel: facts only, name owners only for late items, no blame language.

[Retrospective only, after close:] Compare the last [3] closes: which tasks most often finished late, which had the longest prepare-to-review gaps, and which day the close actually finished vs the plan.

Steps

1. Export the checklist at the same time each day of the close.
2. Run the prompt and check the late-task list against the tool (two-minute sanity check).
3. The close lead edits and posts the status note.
4. After close, run the retrospective section and bring it to the close debrief.

Control map

What the AI may do, step by step. How ratings work.

Audit trail

• The close tool or checklist stays the record of preparer and reviewer sign-offs. The AI summary is a communication aid, not evidence.
• Keep the retrospective output with your close-improvement notes.

Failure modes

• Stale export gives a wrong "late" list. Include the export timestamp in the note.
• Inferred dependencies that don't match how your team works. Track the real dependencies in the checklist once.
• Tone. Auto-generated nudges naming people can land badly. Keep a human editing what gets posted.

Data handling

• Use a company-approved AI workspace whose terms say your inputs are not used for model training. Check your plan; don't assume it.
• Strip or mask what the task doesn't need: employee names, bank account numbers, customer PII, SSNs or TINs.
• If you are a public company, treat pre-release results as material non-public information and keep them inside approved systems.