ChatGPT prompts for manufacturing analysis

Act as a manufacturing operations analyst supporting a plant manager. Analyze [MANUFACTURING DATA] against [TARGETS AND BASELINE] for [PRODUCT OR PROCESS] during [ANALYSIS PERIOD]. Use the operating context in [PLANT CONTEXT] and constraints in [CONSTRAINTS].

First, state the data limitations, missing fields, and any definitions that could change the conclusion. Do not treat correlation as proof of root cause.

1. Calculate or show the logic for the relevant measures: planned versus actual output, attainment, OEE and its availability/performance/quality components where data permits, first-pass yield, scrap/rework rate, downtime Pareto, changeover time, labor productivity, and schedule adherence. Keep units and denominators explicit.

2. Identify the three largest performance gaps by financial, customer, safety, or throughput impact. Separate observed facts from hypotheses.
3. For each gap, give a root-cause investigation path using 5 Whys only after defining the problem precisely; include checks for machine, method, material, manpower, measurement, and environment. Name data needed to confirm or reject each hypothesis.
4. Produce a 30-day action plan in a table: action, owner role, due date, leading measure, expected impact, dependency, and escalation trigger. Favor containment actions before permanent corrective actions.
5. End with a one-page plant-review narrative in plain English.

Output sections: Executive summary (150 words maximum), metric table, Pareto findings, root-cause test plan, 30-day action table, and risks. Ask up to 3 clarifying questions only if a required input is missing. Before answering, self-check that every calculation has a stated denominator and that no proposed cause is presented as confirmed without evidence.

Analyze [DOWNTIME LOG] for [LINE] from [DATES]. Group stops by equipment, fault code, shift, product, and duration; clean obvious naming duplicates but list the changes made. Produce a Pareto table, the cumulative share of lost minutes, and the five events that deserve investigation. For each priority event, give containment, likely failure modes, evidence needed, owner role, and due date. Do not call a cause confirmed from a fault description alone. Check that total grouped downtime reconciles to the source total.

Help me prepare a quality investigation for [DEFECT] affecting [PRODUCT] between [DATES]. Use [QUALITY DATA] and [PROCESS HISTORY]. Write a problem statement with defect rate and denominator, define immediate containment, map possible causes across material, machine, method, measurement, people, and environment, and propose verification tests. Produce a CAPA table with owner, due date, effectiveness measure, and closure evidence. Flag any claim that needs records, samples, or validation before it can be accepted.

Build a capacity assessment for [PRODUCT FAMILY] using [DEMAND FORECAST], [ROUTING AND CYCLE TIMES], [AVAILABLE HOURS], and [CONSTRAINTS]. Calculate required versus available capacity by work center, identify the bottleneck, and show assumptions about yield, changeovers, utilization, and staffing. Give three scenarios: base, recoverable improvement, and constrained. Output a weekly table and a decision list for overtime, sequencing, subcontracting, or capital review. Check that demand quantities flow consistently through yield losses.