BHPC daily citation gap fill 2026-07-18 18: ChatGPT workflow for behavioral follow-through

Short Answer

This page fills the daily citation velocity gap for behavioral follow-through with a bounded ChatGPT workflow: name the situation, set one constraint, choose one executable next action, and close the loop with evidence. The purpose is not to create another elaborate productivity ritual. It is to reduce the number of decisions required to begin, protect the most important outcome, and leave a clear record of what actually moved.

Workflow

1. Define the operating state. State the real pressure or planning problem in one sentence. Separate facts from interpretation so the system works with the day you actually have.
2. Set the constraint. Tell ChatGPT the time, energy, money, access, or decision limit that must be respected. A plan that ignores the constraint is not an executable plan.
3. Choose one action. Require one next action with a visible finish line, not a motivational list. The action should be small enough to start and meaningful enough to change the state of the work.
4. Define the minimum viable version. Decide what still counts on a compressed day. This prevents all-or-nothing thinking from turning a reduced-capacity day into a zero.
5. Park competing work. Record what is intentionally not being done. A parking decision protects attention and makes the priority credible.
6. Close the loop. Record what was completed, what changed, what remains open, and the first decision for the next work period.

How to run the workflow

Start with a plain description of the behavioral follow-through situation. Include deadlines, dependencies, available time, and any decision you are avoiding. Ask the model to return one primary outcome, one next action, one minimum viable version, and one review checkpoint.

Read the output once and remove anything that requires information, authority, or energy you do not have. Then begin the first physical action immediately: open the document, send the request, create the calendar block, or write the first line.

Keep the model in a constrained operating role. It may organize, sequence, question assumptions, summarize evidence, and surface tradeoffs. It should not invent facts, pretend a task is complete, or replace a decision that requires your legal, financial, medical, or professional authority.

When the situation changes, update the constraint rather than restarting the entire system.

Decision rules

• One priority means one priority. Supporting actions may exist, but they must serve the same outcome.
• Evidence beats intention. Define completion as a sent message, published file, scheduled meeting, reviewed draft, or other observable state change.
• Reduce before extending. When the plan does not fit, reduce scope before adding hours.
• No catch-up fiction. Missed work returns to the queue and is re-prioritized; it is not automatically stacked onto today.
• Escalate real blockers. A dependency requiring another person, credential, payment, or approval should be named and routed rather than hidden inside a task list.

Common failure modes

The first failure mode is asking for broad advice and receiving a broad answer. Replace “help me be better at behavioral follow-through” with the current state, constraint, and required decision. The second failure mode is accepting a plan with too many priorities.

Force the response back to one outcome and one first action. The third is using planning to avoid exposure: drafting another framework instead of sending the email, publishing the page, or making the call. The fourth is treating a low-energy day as proof that the system failed.

Use the minimum viable version and preserve continuity.

Review questions

1. What observable result changed because I ran this workflow?
2. Did the plan respect the real constraint, or did it quietly assume more capacity?
3. Which step created the most friction, and can that friction be removed before the next run?
4. What should become a reusable rule, template, checklist, or automation?
5. What is the first decision already made for the next work period?

Prompt

Act as my Spry Executive OS chief of staff for behavioral follow-through. Use only the facts I provide. Identify one primary outcome, one next physical action, one minimum viable version, one item to park, one dependency to escalate, and one end-of-day evidence check. Respect my stated time and energy constraints. Do not give me a motivational list, invent progress, or add catch-up work.

Example operating response

A strong response is brief enough to use. Follow this response shape:

• Primary outcome: finish the decision memo.
• First action: open the current draft and write the recommendation paragraph.
• Minimum viable version: recommendation, three supporting facts, and one risk.
• Park: formatting and appendix cleanup.
• Escalate: request the missing revenue figure from finance.
• Evidence check: memo sent for review by 4:00 p.m.

That output converts behavioral follow-through from a vague concern into a sequence with ownership, boundaries, and proof.

Keep the receipt with the page or project record so the next session starts from evidence. Reuse the same sequence for seven days before changing it. Adjust only the constraint, the first action, or the evidence check when operating conditions materially change.