Five Prompts Every Operations Leader Should Have Saved

These five prompts cover 80% of the operational decisions that eat your week. Save them somewhere you can find them.

Operations leaders make dozens of high-stakes calls per week: allocation trade-offs, capacity decisions, vendor negotiations, process redesigns. Most of these rely on gut instinct plus a spreadsheet. AI gives you a structured thinking partner for the five recurring decision types that consume the most leadership time. Each prompt below is grounded in a specific framework, which is why it works better than asking a generic question.

1. The Constraint Finder

When to use: Quarterly, or whenever throughput stalls despite adding resources.

Framework: Goldratt's Five Focusing Steps

Here is a description of our current operational flow: [paste process description, metrics, and recent bottleneck symptoms]. Using Goldratt's Theory of Constraints, identify the single binding constraint. Distinguish between physical constraints (equipment, space, labor) and policy constraints (batch rules, scheduling logic, approval workflows). For each candidate constraint, explain why it is or is not the true bottleneck. Then recommend the Exploit step: how can we get more throughput from this constraint without any capital investment?

Example: A $1.2B e-commerce company running 94% fill rates against a 98% target. Warehouse overtime up 18% year over year. The obvious answer is more labor or dock doors. The actual constraint: a wave planning batch size rule set three years ago when order volume was 40% lower. The policy constraint is invisible until you force the analysis.

What to look for: The AI should distinguish policy constraints from physical ones. If it only identifies physical constraints, push back. Policy constraints (batch rules, approval workflows, scheduling logic built for a smaller operation) are where mid-market companies get stuck most often. This is the same diagnostic approach covered in Why Your Warehouse Constraint Is Probably Not What You Think.

Tool note: Claude handles longer process descriptions and context well. Paste the full process, not a summary.

2. The Root Cause Drill

When to use: When a recurring problem keeps getting band-aided.

Framework: First Principles Thinking + Five Whys

We keep experiencing [describe recurring problem with data]. We have tried [list previous fixes]. Using first principles thinking, decompose this problem to its foundational causes. Do not accept our previous assumptions. For each root cause you identify, explain what evidence would confirm or refute it, and what the second-order effects of fixing it would be.

Example: Late shipments spike every Thursday. Previous fixes: added a Thursday shift, moved Thursday orders to Wednesday processing. Neither worked. The root cause turned out to be a Monday ordering policy that creates a predictable demand wave arriving at the warehouse on Wednesday afternoon, overwhelming Thursday processing.

What to look for: The AI should challenge your previous assumptions, not just organize them. If it restates your existing fixes as root causes, it missed the point.

3. The Process Audit

When to use: Before investing in automation or headcount.

Framework: Leverage (Farnam Street) + Subordinate step from TOC

Here is our current process for [describe process, include steps, handoffs, cycle times, and error rates]. Identify the steps with the highest leverage: where a small improvement would create the largest downstream impact. Separately identify steps that exist only to compensate for upstream failures. For any step I am considering automating, tell me whether I am automating a value-adding step or automating a workaround.

Example: An order management process with 14 steps, 4 handoffs, and a 6% error rate at step 9. The error rate is caused by a manual data entry step at step 3. Nobody flagged step 3 because the error does not surface until step 9. Automating step 9 would automate a workaround. Fixing step 3 eliminates the problem.

What to look for: The distinction between value-adding steps and compensating steps. If you are about to invest in automating a workaround, this prompt will catch it.

4. The Trade-off Clarifier

When to use: Before any resource allocation or investment decision.

Framework: Inversion + Second-Order Thinking

I am deciding between [Option A] and [Option B] for [context]. For each option: (1) What is the best realistic outcome? (2) What is the worst realistic outcome? (3) What second-order effects will this create in 6-12 months? (4) What would I have to believe for this option to be the right call? (5) What is the cost of reversing this decision if it is wrong? Present this as a structured comparison, not a recommendation.

Example: Deciding whether to bring fulfillment in-house versus staying with a 3PL. The prompt forces consideration of hidden costs: management overhead, lease commitments, the 18-month ramp to operational parity, and what happens to your 3PL relationship if you partially in-house and need to scale back. These are the second-order effects most planning processes skip.

Tool note: GPT-4o produces clean structured comparisons. Claude tends to add more nuance to the "what would I have to believe" question. Try both and see which output style matches how you present decisions to your leadership team.

5. The Decision Journal Entry

When to use: After making any significant decision.

Framework: Probabilistic Thinking + Map is Not the Territory

I just made this decision: [describe decision and reasoning]. Help me create a decision journal entry. Include: (1) What I knew at the time. (2) What I was uncertain about and my confidence levels. (3) What alternatives I considered and why I rejected them. (4) What would make me revisit this decision in 90 days. (5) What leading indicators I should watch weekly to know if this decision is working.

Example: Deciding to consolidate from 3 regional DCs to 2. The journal entry captures assumptions about transit time impact, carrier rate renegotiation, and customer satisfaction risk. Six months later, when transit times are up 14% and the board asks why, the journal shows exactly what you assumed and what changed.

What to look for: The leading indicators. Most decision reviews focus on lagging outcomes (did revenue go up?). The journal should identify what to watch weekly so you can course-correct before the lagging indicators turn.

What to Do Monday Morning

Pick the decision you are currently stuck on. Find the prompt above that matches the decision type. Paste it into Claude or ChatGPT with as much context as you have. You will either get a sharper answer or a better question. Both are worth the five minutes.

Published by The Clearheaded Operator.